TOPIC Retrofitting of Buildings Page

Transcription

1 TOPIC Retrofitting of Buildings Page Full papers - NSB 2014 page 1171

2 Full paper no: 146 Holistic retrofit and follow-up through monitoring: Case Virkakatu, Oulu, Finland Yrsa Cronhjort, M.Sc. Architecture 1 Simon le Roux, M.Sc. Architecture 1 1 Aalto University School of Arts, Design and Architecture, Finland KEYWORDS: E2ReBuild, monitoring, retrofit, residential, TES-system, user interface SUMMARY: Case Virkakatu, Oulu, exemplifies an extensive retrofit of an apartment building representing the prefabricated concrete element BES system. The original house was built according to Finnish building standards of the early 1980 s. Building works included a complete refurbishment of the interiors, a renewal of the building envelope including floor slabs, facades, windows, doors and the roof. New building service systems were installed. The apartments were equipped with separate air ventilation units with efficient heat recovery, and the target was for passive house level according to the local suggestion by VTT. This site is the second building in Finland in which the facades have been retrofitted using timber based elements, the TES-system. The demonstration is one of seven pilots realized within the EU Fp7 funded project E2ReBuild. The project has developed a monitoring plan that was adopted and extended in case Virkakatu. The building envelope is monitored for moisture and temperature, the indoor air quality is surveyed through following ventilation operation, carbon dioxide levels, and room temperature. Energy usage including heating, hot water and electricity is measured. The monitoring is completed with an own weather station on site. The user is involved with apartment wise user interfaces. The monitoring scheme has proved a useful means to prove retrofit results, follow energy use, and secure the functionality of building service systems. 1. Introduction Case Virkakatu, Oulu exemplifies an extensive energy retrofit of an apartment building representing the concrete element BES-system developed in the 1960 s and typical Finnish housing production. The system is based on the use of standardized prefabricated elements and joint details. The load bearing structure of the demonstration building is a bookshelf frame with a precast concrete sandwich facade. The building in Oulu is two storeys high, and contains 8 student apartments. The original gross floor area was 744 m 2 and the heating was based on district heating. The building was constructed according to Finnish building standards of the early 1980 s. Before building works the condition of the building and its structures was analysed: The original floor slab consisted of a 70 mm reinforced concrete slab isolated with 50 mm expanded polystyrene and the original facades of an mm thick external layer of brickwork, mm of thermal wool insulation and an inner layer of concrete. The facade structure was additionally analysed using GPR survey, as an experiment, as to verify the surface profile of the inner concrete layer. The building is part of a housing cooperative with five apartment buildings and a service building with communal facilities. As part of the EU FP7 funded project E2ReBuild one of the buildings was refurbished. The other buildings will be renovated at a later stage in a conventional manner with less demanding targets for energy efficiency. These will be monitored for comparison. Full papers - NSB 2014 page 1172

3 FIG 1. The structural core of the case building in Finland represents the BES system with a load bearing concrete frame and precast concrete sandwich façade elements. Image Simon le Roux. 2. Refurbishment Case Virkakatu Case Virkakatu is one of seven demonstrations within the EU FP7 funded project E2ReBuild. The project applies, demonstrates, evaluates and follows up the results of various cost effective and advanced strategies for industrialized retrofitting of residential buildings. The common targets for the demonstrations are to improve the overall energy efficiency as to fulfil at least national limit values for the energy use in new buildings and to reduce the space heating demand by at least 75%. (E2ReBuild 2013) The objective for the refurbishment in Oulu was to reach passive house level of energy efficiency according to the local suggestion by VTT (Nieminen Lylykangas 2009), a level above current demands for new building. The chosen strategy was a holistic retrofit including the application of TES Energy Facade for improving the thermal insulation and air tightness of the building envelope. Building works included a complete refurbishment of the interiors and the building envelope including floor slabs, facades, windows, doors and the roof. New building service systems were installed. The aim was to create student family apartments responding to modern living requirements. Building works on site started with the removal of the original in-situ ground floor slabs and the external layer of concrete and brickwork from the precast facade elements. The old thermal insulation layer was stripped away. New facades were manufactured from prefabricated, timber based elements (TES, FIG 2-4). The facade elements consisted of a load bearing timber frame and glass mineral wool thermal insulation, gypsum wind barrier boarding on the outside and plywood on the inside. A thin thermal insulation layer was added to the elements on site as an adjustment layer between the elements and the uneven existing concrete surface. External cladding and windows were assembled on site. The total thickness of new thermal insulation in the completed facade is 300 mm with a U-value of 0.11 W/m 2 K (FIG 5). A 200 mm thick layer of graphite-enhanced EPS thermal insulation was installed beneath the new ground floor slab and 550 mm of blown loose fill mineral wool was installed in the roof, with U-values of 0.11 W/m 2 K and 0.08 W/m 2 K respectively (FIG 6). Inward opening wood aluminium passive house casement windows were installed, with an average U-value of 0,8 W/m2K. Special attention was paid to improving the airtightness of the building envelope, with an original measured n50 value of 3.3 l/h and 0.8 l/h after renovation works (Puotiniemi 2012, Puotiniemi 2013). Corresponding measured q50 values were 3.1 m3/h*m2 and 1.2 m3/h*m2 (Puotiniemi 2012, Puotiniemi 2013). Full papers - NSB 2014 page 1173

4 This site is the second building in Finland in which the facades have been retrofitted using timber based elements, the TES system, and the most northern demonstration in project E2ReBuild. The EnerPhiT Passive House Certificate for old buildings in accordance with EnerPhiT criteria for Residential-Use Refurbished Buildings (Feist 2010) is to be considered. FIG 2-4. The façade retrofit started with the removal of redundant structures and an uneven inner concrete surface was revealed. A thin layer of thermal insulation was added as an adjustment layer to the prefabricated timber based façade elements prior to assembly. TES-elements were finally assembled directly onto the existing building, thus forming a new façade. The cladding was installed on site. Image to the left: Simon le Roux. Images 3 and 4: Jaakko Kallio-Koski M3 Architects. The old roof of the building was also demolished and replaced by a new roof and thermal insulation. The roof structure was prefabricated in four sections on site and lifted onto the building as elements. Air-tightness of the building was improved by several measures. The original concrete frame was repaired before adding the new facades and ground floor slab as it had significant air leaks which were filled with cement grout. The TES-elements were encapsulated with plywood and a wind barrier. The air-tightness of window and door fittings, ceilings and duct penetrations was ensured on site during installation works. Windows were sealed to the inner concrete frame with a polyurethane based elastomeric joint sealant. The remaining air leaks after refurbishment were determined to derive from shrinkage of the new in-situ concrete floor slab during the drying of the concrete (Puotiniemi 2013). The internal works included a remodelling of the student flats into family apartments with renewed floor plans, new kitchens and bathrooms including saunas and new surfaces. The lighting was replaced with energy efficient LED fixtures. The apartments were additionally upgraded with separate air ventilation units with rotary heat exchangers of an annual heat recovery efficiency rated 75.7%. FIG 5-6. Façade and roof structure after refurbishment. Outer façade layers were removed leaving the inner concrete layer in place. A new façade was retrofitted with prefabricated timber based elements. The U-value of the finished wall is 0.11 W/m2K. The old roof was completely replaced, adding 550 mm thermal insulation, resulting in a U-value of 0.08 W/m2K. The external wall structure from the outside: cladding, mm air gap, 9mm wind barrier GU 9, 50 mm thermal insulation layer, 200 mm thermal insulation layer, 9 mm plywood board, 50 mm thermal insulation (adjustment layer), old inner concrete layer. The thermal insulation is glass mineral wool. Images: Simon le Roux. Full papers - NSB 2014 page 1174

5 3. Monitoring Scheme Project E2ReBuild includes one Work Package devoted to the monitoring and follow up of the implemented retrofit strategies; WP5 Innovation in Operation and Use. Within this Work Package a monitoring scheme was designed to be applied in all demonstrations of the project. This monitoring scheme was implemented and extended in case Virkakatu to also include monitoring the building physics performance in addition to measuring the attributes directly related to energy use. The building facades, ground floor slab and roof are monitored for moisture and temperature. Indoor air quality is surveyed by following the ventilation operation, carbon dioxide levels, and room temperatures in selected apartments. Energy usage including space heating, domestic hot water and electricity is separately measured for the building as a whole and for individual apartments. The monitoring is completed with an onsite weather station. The user is provided with a digital display of real-time information on private electricity and water usage in each apartment. In this paper our focus is on the follow-up of the energy use in the building and the effects of the monitoring scheme. See Table 1 below. TABLE 1. Oulu demonstration monitoring scheme Main parameters for comparison Comments User interface Purchased energy Shared by 5 houses Space heating Domestic hot water Building electricity Household/tenant electricity District heating Flow and temperature x (volume) x (kwh) Additional parameters Indoor dwelling temperatures and RH Indoor CO 2 Outdoor temperature, RH and irradiation Airtightness and thermal imaging Ventilation rates, electricity Building envelope performance 2 apartments 1 apartment Operation x (temperature) The monitoring is followed up through an online interface with access to all data. The interface includes an alarm function, allowing for early intervention in case of malfunction of i.e. an air ventilation unit. Airtightness was measured before and after refurbishment using a blower door test and thermal imaging (Puotiniemi 2012, Puotiniemi 2013). Reference data on purchased energy, space heating and building electricity was collected also prior to building works. The monitoring is realized using a digital building automation system, which collects measurement input to a programmable substation (FIDELIX FX-2025a Digital Controller), from which data is transferred via virtual private network to outside monitoring. The controller interface has input/output modules to directly connect the network of different monitoring points around the building (36 channel combination module, 16 channel digital input and 8 channel analogue input modules). The substation communicates with the input modules using Modbus communication protocol (Modbus RTU RS-485), and also collects monitoring input from all eight apartment ventilation units (ENERVENT PINGVIN eco ECE ventilation unit Multi Web-ModBus). Each apartment has a FIDELIX Multi-LCD room panel with 3,5 colour LCD touch screen, programmed to display hot and cold water volume, electricity meter and outdoor temperature, daily, weekly and monthly use. The building automation controls the supply of heating water from the district heating heat exchanger according to the current outdoor temperature using a valve actuator (HRYD24-SR). Outdoor Full papers - NSB 2014 page 1175

6 conditions are monitored with a weather station (DAVIS Vantage Pro2 Plus) and outdoor illumination and temperature sensors (PRODUAL LUX 11 + NTC 10). An ultrasonic compact energy meter (SHARKY 775) is installed to measure and calculate the total energy demand in district heating for space heating and domestic hot water. Immersion temperature sensors (PRODUAL TEAT NTC10) and the flow meter together measure the overall space heating delivered to the radiators and the domestic hot water supplied to each apartment. Two apartments are monitored for indoor temperature/humidity with transmitters in 4 rooms (PRODUAL KLH100), and with C0 2 transmitters in air intake and exhaust ducts (PRODUAL HDK). Structure humidity and temperature sensors are installed in four TES elements which face north, south, east and west, and in the ground floor slab (HONEYWELL HIH-4010 moisture sensors and PRODUAL TE NTC10 temperature sensors). The roof insulation is monitored with PRODUAL KLU- 100 outdoor humidity and temperature transmitters. (Palosaari 2013) 4. Experiences and Results The building works on site were finished in February 2013 with the first tenants moving in 1 st of March. Hence the monitoring is in its early phases covering a period of 9 months and excluding the winter period, but nevertheless, shows the effects of the refurbishment. As to illustrate the results of the energy retrofit the energy demand for space heating and domestic hot water before, as simulated and after were compared (FIG 7). The energy demand before is the monthly average based on the last eight year purchased district heating for the whole building complex. The simulation prior to renovation was done using the IDA Indoor Climate and Energy dynamic simulation tool for studying thermal indoor climate and energy consumption of the building (Equa). Assumptions were made in the simulation of the average domestic hot water demand based on Finnish regulations. As seen in the graph, the simulated heating demand after refurbishment is higher in summer than the previous average purchased energy for the same months of June to August, which suggests that the simulation has a large safety margin for worst cases, or that students use less water in summer than expected. Estimates were made in the simulation to include the influence of additional heating coming from tenants use of under floor electrical comfort heating. A calculation was also made of the average heat losses in the transfer of space heating and hot water in underground heating pipes from the centralised district heating heat exchanger of the building complex. The actual monitored heating use excludes the heat losses in underground pipes, which are outside the case study building. However the heat losses are estimated to be at worst 10 15%, and even so the current monitored energy use is less that the simulated energy demand. The energy demand in the graph (FIG 7) is calculated according to the Finnish standard calculation methodology for the total gross floor area of the building. The comparison of energy performance is complicated by the variety of definitions for floor and volume. Care had to be taken to check for inconsistencies in figures used by different designers, energy calculations, planners and authorities. The size of the building grew slightly after refurbishment, since the volume was filled in places, which increased the net floor area. The thickness of the external walls, ground slab and roof insulation was increased, but the leasable floor area remained approximately the same. The volume calculation of the building was particularly complex, since the external surface is multi-layered and irregular. (TABLE 2) Full papers - NSB 2014 page 1176

7 FIG 7. First 9 months monitoring results of energy use for space heating and hot water case Virkakatu, Oulu, Finland. Graphs from the top: Average purchased district heating energy for space heating and hot water before, IDA-ICE simulated energy demand for space heating and hot water after, Monitored energy demand for space heating and hot water after, Monitored energy demand for space heating only after refurbishment. Numbers are net energy without normalization. The demand is calculated for the gross floor area of the building. The image shows a dramatic drop in energy demand, averaging a 60% decrease in energy consumption for space heating and hot water after refurbishment. Image Simon le Roux. TABLE 2. Case Virkakatu, Oulu, Finland details of the building Parameters Before After Total gross building volume 2196 m m3 Total heated/cooled net volume Air volume Total gross area Leasable area 1802 m m3 744 m2 572 m m m3 791 m2 578 m2 Apartments and stairwells Total heated/cooled net floor area (internal) 633 m2 652 m2 661 m2 680 m2 The savings in energy demand for space heating and hot water benefits in this case the building owner, PSOAS. In Finland, the rent for an apartment includes space heating and hot water, sometimes paid for separately as a lump sum based on the number of dwellers in the apartment. In addition to the rent, the tenant pays separately for his own electricity directly to the electricity distributor. As to allow for user control of their own, personal energy consumption for electricity, the apartments were equipped with user interfaces monitoring the apartment s use of hot water and electricity. Over the period of the first 8 months the results show a great variety in usage with less consumption during the summer months, typical vacation time in Finland. However, the tenants of one Full papers - NSB 2014 page 1177

8 apartment reported on following their electricity use from the interface and having saved 25% in electricity by changing their own behaviour accordingly. (FIG 8) FIG 8. First 8 months monitoring results of energy use for hot water, ventilation and electricity in single apartments case Virkakatu, Oulu, Finland. Graphs from the top: Hot water use 8 apartments, ventilation 8 apartments, and electricity use 8 apartments. Measurement units in kwh. After the first months the energy demand for ventilation units has stabilized as well as the average demand for hot water. The user electricity shows the biggest fluctuation. This fluctuation may be caused by vacations in summer time and the increased need for artificial interior light towards the dark winter months. Only one apartment shows a fairly constant level of electricity use after the first two months. The graph shows full measures for the months from March through October Image Simon le Roux. The monitoring scheme has proved a useful means to prove retrofit results, follow energy use and secure the functionality of building service systems. The first two months showed unreliable measures for the use of hot water, but since then the monitoring results have stabilized at average levels of 30,2 kwh energy used for ventilation and 98,5 kwh energy used for hot water per apartment and month. With an average leasable floor area of 71,5 m2/apartment this equals, on average, 0,42 kwh/m2 leasable area of net energy used for ventilation per month or 5 kwh/m2/annum, with 1,38 kwh/m2 leasable area of net energy used for hot water per month or 16,5 kwh/m2/annum. These results indicate that by using water saving showerheads and dual flush toilets in the apartments, the average water usage after refurbishment is 55% less than assumed in building regulations. The building owner had already installed water saving showerheads in the apartments in The user interface has proven useful for affecting user behaviour in one of the eight case apartments. Full papers - NSB 2014 page 1178

9 5. Conclusions The paper presents a comparison of energy use before an extensive building refurbishment, simulated energy use and monitored energy use after refurbishment. The case building is a two storey high apartment building situated in Oulu, Finland. Coordinates of the location are 65 01' N, 25 28' E (Oulu). The aim for the refurbishment was passive house level of energy efficiency according to local suggestion by VTT. A comprehensive monitoring scheme was applied to the project as to verify and follow up achieved results for energy efficiency, functionality of applied timber structures and achieved interior air quality. The results for the first eight months indicate a high level of energy efficiency including 55% savings in the use of hot water as compared to simulation. Additionally, the results indicate a positive effect on tenant s use of electricity and hot water as a result of their access to real-time data through apartment wise user interfaces. Overall the monitoring scheme has proven an efficient and easy approach to verifying the results of an ambitious refurbishment project. 6. Acknowledgements The paper is based on research and academic support for the passive house demonstration in Oulu, Finland, realized within the EU FP7 funded project E2ReBuild Industrialised energy efficient retrofitting of resident buildings in cold climates. The project is coordinated by Christina Claeson- Jonsson, NCC AB. The project started in January 2011 and ends in June Partners from Finland include the Aalto University, NCC Rakennus Oy and Pohjois-Suomen Opiskelija-Asuntosäätiö PSOAS. The monitoring scheme for the Oulu demonstration project was developed in collaboration with Carl Magnus-Capener, SP Sveriges Tekniska Forskningsinstitut AB, leader of Work Package 5 Innovation in Operation and Use. The research leading to these results has received funding from the European Union's Seventh Framework Programme (FP7/ ) under grant agreement n References Feist W EnerPHit Certification as Quality-Approved Energy Retrofit with Passive House Components Criteria for Residential-Use Refurbished Buildings. 16 p. Nieminen J. & Lylykangas K Passiivitalon määritelmä. ohjeita passiivitalon arkkitehtisuunnitteluun. p 9. Palosaari M Rakennemittausten liittäminen Fidelix-automaatiojärjestelmään. Opinnäytetyö. Automaatiotekniikan koulutusohjelma. Oulun seudun ammattikorkeakoulu. Oulu, Finland. 42 p. Puotiniemi J Ilmatiiveysmittausraportti Psoas Virkakatu Oulu CRAMO. Oulu, Finland. p 4. Puotiniemi J Tiiviysmittausraportti Psoas Virkakatu Oulu. CRAMO. Oulu, Finland. p 1. IDA ICE software, Equa Solutions: accessed Fenestra FW-Passive window U-values: accessed Project E2ReBuild accessed The City of Oulu accessed Full papers - NSB 2014 page 1179

10 Full paper no: 147 Energy retrofitting of an old multi-storey building with heritage value. A case study in Copenhagen with full-scale measurements Maria Harrestrup, M.Sc. 1 Svend Svendsen, Professor 1 Agis M. Papadopoulos, Professor 2 1 Technical University of Denmark, Denmark 2 Aristotle University of Thessaloniki, Greece KEYWORDS: Energy renovation, Energy savings, Heritage Building, Internal Insulation. SUMMARY: Europe has a vision of reducing energy consumption significantly and Denmark has set up an even more ambitious goal aiming at being completely fossil-free by But already in 2035 the energysupply mix for buildings are aimed to be free of fossil fuels. Energy retrofitting of buildings is an important solution of securing energy reductions. But challenges occur when it comes to retrofitting of heritage buildings. A case study on a typical multi-storey building with heritage value in Copenhagen has been carried out. Theoretical investigations are validated with full-scale measurements on energy consumptions before and after the renovation. Energy-saving measures that pay regard to the heritage values of the building are in focus. This implies solutions such as internal insulation. When insulating the facade from the inside, the facade will become cold and condensation in the interface can occur. Furthermore the beam construction will not be insulated, which create a large thermal bridge and the moisture and temperature conditions in the wooden beams will change and attention needs to be given to risk of degradation of the wood. The investigation showed that the actual energy consumption was reduced by 42% whereas the calculated was reduced by 58%. The deviation might be due to altered occupant behaviour. Furthermore relative humidity and temperature measurements in the beam-ends showed no risk of wood decay. 1. Introduction Europe has a vision of reducing greenhouse gas emissions and energy consumptions by 20 % by 2020 and 80% in 2050 compared to 1990-levels (European Commission 2010) and Denmark has set an even more ambitious goal, aiming at being completely fossil-fuel-free by Even more, the energysupply mix for buildings is aimed to be free of fossil fuels as soon as 2035 (Danish Minister of Climate 2011; Danish Energy Agency 2010). Since the building sector is responsible for approximately 40% of the total energy consumption in EU today (Lechtenböhmer and Schüring, 2011) and less than 1% is replaced with new low-energy buildings (Hartless 2003), focus needs to be given to the old inefficient building stock. Different legislative frameworks have been introduced within the area of energy efficiency and buildings with the main legislative instrument in Europe to be the Energy Performance in Buildings Directive (EPBD) implemented in 2002 (EU 2002). In 2010 a recast of the directive was introduced stating that new and retrofitted buildings should consume nearly zero energy (EU, 2010). Energy retrofitting of buildings is an important solution of securing energy reductions, but challenges occur when it comes to retrofitting of heritage buildings where the facade cannot be modified due to the architectural value of the building. The study (Morelli et al. 2011) investigated an energy retrofit of a multi-storey building with heritage value from 1930 constructed with brick facades and found that it was possible to save 70% of the energy consumption. In such cases internal insulation is the only solution for insulating the facade. When insulating the facade from the inside, the facade will become cold and the drying potential of the wall will be reduced. Condensation in the interface between the insulation and the brick wall can occur, which can lead to mould growth (Christensen and Bunch-Nielsen 2009; Munch-Andersen 2008). Furthermore the Full papers - NSB 2014 page 1180

11 horizontal division separating the floors will not be insulated, which will lead to the occurrence of a thermal bridge, where the load bearing beam construction is placed. Many buildings with heritage value are constructed with wooden beams as load bearing structural elements (Engelmark 1983). The moisture and temperature conditions in the wood will change and attention needs to be given to the risk of the wood s degradation, which can, in extremis, lead to a risk of fatal structural damage. Multiple studies have investigated the effect on the wooden beam constructions when applying internal insulation, including Krebs and Collet (1981), Chistensen and Bunch-Nielsen (2009), Munch- Andersen (2008) and Rasmussen (2010). These studies, however, do not include detailed investigation of the impact of Wind Driven Rain (WDR). Morelli & Svendsen (2012) investigated different intensities of WDR on the facade and concluded that the WRD has a great impact on the performance and durability of the wooden beam ends. Kehl et al. (2013a) provides a literature review that also concludes that WDR has an important influence on the behaviour of moisture content and the risk of decay of the beam ends. Other studies that investigate the impact of internal insulation are Ruisinger (2013) and Kehl et al (2013b). This paper presents results from an energy retrofit of a typical multi-storey building in Copenhagen with heritage value. The aim of the project was to demonstrate a method of how to energy renovate a heritage building focusing on energy savings and to demonstrate the efficiency of different technologies and products. Theoretical investigations are validated with full-scale measurements of energy consumption before and after the renovation. Since the building is protected, energy-saving measures complying with its heritage value have been considered. Economic aspects are not discussed in this paper, since they had already been evaluated before the renovation took place in order to determine the most feasible energy saving measures. The results from that phase (the planning phase of this study) are described by Morelli et al (2012). 2. Methodological approach 2.1 Approach Step 1: The existing energy consumption is calculated with the building simulations software IDA ICE 4.5 and validated with the measured energy consumption. The measured energy consumption consists of space heating and domestic hot water consumption and is an average consumption from the period Step 2: The energy consumption of the renovated building is calculated with the retrofit measures implemented in the IDA ICE model and the reduction in energy consumption is evaluated based on the results. Step 3: The actual energy consumption of the renovated building is measured including heating consumption, domestic hot water consumption and electricity consumption for mechanical ventilation. The heating consumption was measured for the entire building including space heating and domestic hot water provided by HOFOR A/S. Three different mechanical ventilation systems were implemented in the building. A comparison between the electricity usages for each system was carried out. Step 4: Since the building is of heritage value, the use of internal insulation was used. The relative humidity and temperature conditions are measured in the wooden beam end embedded in the masonry wall. The measurements are carried out in the apartment that is more exposed to wind, rain, and sun. The measurements are used to evaluate the risk of moisture and degradation problems in the beam that may occur when implementing internal insulation. 2.2 Description of the existing building The building is located in Copenhagen, Denmark, and was built in It is a multi-storey building with 6 floors with a heated area of 2717 m 2 and an unheated basement. The building consists of 30 apartments located in three apartment blocks (Block A, B and C). The plan view of the building together with a 3D-view of the model build in IDA ICE 4.5 can be seen in Figure 1. The load-bearing construction is made of wooden beams and brick walls. The thickness of the brick wall varies from mm. There is no insulation in the building envelope and the windows are old 1-layer inefficient windows. The building has a heritage value corresponding to Class 4 from the SAVE Full papers - NSB 2014 page 1181

12 classification system in Denmark (SAVE 2011), which imply that the building envelope cannot be changed. The U-values for the existing building can be seen in Table 1. The building is heated with district heating and fresh air is provided by natural ventilation from opening windows and leaks. In the toilet/bathrooms and kitchens an exhaust ventilation system was installed but in most apartments the ducts were blocked and did not work. Many of the apartments did not have a shower. Block B Block C Block A FIG 1. Left: 3D-view of IDA ICE model. Right: Plan drawing of the apartments. The red circle indicates the apartment for moisture and temperature measurements for evaluating the use of internal insulation. TABLE 1. U-values for the existing building. U- values W/(m! K) Average brick wall 1.40 Roof 0.52 Floor to basement 1.50 Windows 1 pane Retrofitting approach General improvements The building went through a deep retrofitting where energy saving measures were in focus, but also the interior comfort was improved by installing bathrooms in all the apartments, which were lacking previously. Furthermore permission was given from the municipality to install penthouse flats with roof terraces and solar cells, increasing the total value of the building Building envelope Before choosing which energy saving measures that should be used for the renovation, different energy saving measures were tested in a test apartment on the first floor in Block C. Different window-solutions and different internal insulation products were installed and tested with regards to costs, energy performance, and moisture and temperature conditions. The different solutions are described in Morelli et al. (2012). Since the building has heritage value, the facade cannot be changed. However, the municipality accepted that the windows were changed to new windows since it was proven from the test apartment that they are the most cost- and energy-optimal choice. The windows for the deep retrofitting were from Frovin Windows and Doors A/S and reconstructed as the old Full papers - NSB 2014 page 1182

13 windows. The frame is made of wood and the U-value for the window is U = 0.89 W/m 2 K. Due to the heritage value of the building internal insulation was used except from the north-east facade (Block C). This facade is facing a narrow passage and 250mm external mineral wool from Rockwool was applied (U-value = 0.39 W/m 2 K). The remaining facades were insulated with internal insulation. The apartments in Block A and B were insulated with Aerorock (a mix of Rockwool and Aerogel) with a U-value of 0.19 W/m 2 K and in the apartments in Block C Kingspan (PUR) were applied with a U- value of 0.20 W/m 2 K. The installation of internal insulation can create a risk of mould growth at the interface between the insulation material and the brick-wall since the temperature at the interface will decrease when applying internal insulation. Experience from the test apartment showed that it is crucial to clean the brick wall from organic material before applying the internal insulation in order to minimise the risk Moisture and temperature measurements In order to evaluate the implementation of internal insulation and the effect on the beam ends, temperature and relative humidity measurements are carried out in the beam ends. In the test apartment a similar investigation was done and showed no risk. However, the test apartment was orientated north-east facing a small passage and was located on the 1 st floor, which implies a minimised amount of wind-driving rain. The measurements for this study are therefore implemented in the apartment that is exposed to the most extreme weather conditions (wind driven rain, sun, wind etc.). The apartment is facing south-west and is on the 5 th floor where no shadow or shelter is present. The measuring points and a drawing of how the internal insulation was installed are seen in Figure 2. Due to wall socket and the cables for electricity the insulation was stopped 200 mm above the floor (see Figure 2). This solution is supported by investigations carried out by Morelli and Svendsen (2012) who concluded that if the insulation is stopped 200 mm above the floor, the risk of wood decay is decreased significantly. Gypsum board Insulation Brick wall FIG 2. Left: Measurement points for temperature and relative humidity in the beam-ends. Right: Drawing of how the internal insulation was mounted Mechanical ventilation Vapour barrier Panel Wooden beam Sensor point Three different mechanical ventilation systems with heat recovery were installed in the three apartment blocks respectively. Block A: A traditional central mechanical ventilation system is installed. The air handling unit (AHU) is placed in the basement and the existing chimneys are used for the exhaust from the bathrooms and kitchens. The supply air to the living rooms is at a constant rate of 140 m 3 /h. The exhaust air from the bathrooms is constant at 20 m 3 /h when the relative humidity is below 55% and 54 m 3 /h when it exceeds 55%. In the kitchen the exhaust rate is 72 m 3 /h, but increases to 144 m 3 /h when the cooker hood is activated. Block B: The ventilation system in Block B is a central ventilation system with the AHU placed in the basement. The ventilation system is Full papers - NSB 2014 page 1183

14 operated as demand controlled based on CO 2, relative humidity and temperature. Furthermore there is a user panel to regulate specific needs. During unoccupied hours a dispensation was given from the municipality to ventilate only with an air exchange rate of 22 m 3 /h. Block C: The ventilation systems in Block C are decentralized systems on apartment level. This implies that each apartment has an AHU. The control for the ventilation system is demand controlled as for the system in Block B. 3. Results 3.1 Energy consumption The annual energy consumption calculated and measured is shown in Table 2. The measured energy consumption for the existing building is an average from and includes space heating and domestic hot water. The deviation between the measured and calculated annual heating and hot water consumption for the existing building is 2.5 % which is considered acceptable. The calculated energy consumption for the renovated building is 63.0 kwh/m 2 /yr, which corresponds to a reduction in the energy use of 58% compared to the calculated energy use before the renovation. The biggest reduction is found in Block C (65 %) since the north/east facades with no sunlight were insulated with 250 mm external insulation reducing the heat loss dramatically. Morelli et al. (2012) calculated the energy consumption to be reduced by 68% with the building energy simulation software BE10 that is based on Danish standards. The software BE10 is however a simplified software and is working as a onezone-model whereas IDA ICE software is a multi-zonal software providing more details. The calculated electricity consumption for ventilation is seen to be less for the demand controlled systems followed by the traditional central ventilation with constant airflow, but the calculated energy consumptions are very similar in all three cases. TABLE 2. Calculated and measured energy consumption for existing building. Calculated energy consumption for the renovated building. Existing building Renovated building [kwh/m 2 /yr] Measured Calculated Calculated Space heating Space heating Space heating Mechanical Total and hot water and hot water and hot water ventilation Block A Block B Block C Total weighted Measured energy consumption in the renovated building Heating consumption The energy consumption was measured after the renovation was carried out. The total heating consumption for the entire building including space heating and domestic hot water was measured and is shown in Figure 3. Since the measured values include space heating and domestic hot water, the monthly hot water consumption is estimated based on an average value of the heat consumption from June, July and August. It is expected that there is no space heating consumption during these month and even though variations in the hot water consumption occurs over the year, it gives a reasonable estimate. The month November and December are estimated values in order to calculate the annual heating consumption per meter square. The total heated floor area after the renovation is 2892 m 2. This gives a total annual heat consumption of 88.8 kwh/m 2 (see Table 3). Considering weather data it was found that the monthly average temperature in 2013 was 1.5 C higher than that of the design reference year used in the simulation model. Full papers - NSB 2014 page 1184

15 Heat consumption [MWh] Total heat consumption Estimated Estimated FIG 3. Total heat consumption for the entire building Electricity consumption for ventilation The electricity consumption from the three different ventilation systems was measured and presented in Figure 4. Based on the measurements in Figure 4 the annual consumption per meter square was estimated for each system in order to compare them. As seen from Table 3 the ventilation system in Block B (Centralised demand controlled ventilation) is using the most electricity followed by Block C (Decentralised demand controlled ventilation) and A (Centralised constant ventilation). Table 3 shows that the measured annual energy consumption is 90.7 kwh/m 2, which is 28 % more than the calculated energy consumption. Daily electricity consumption for the ventilation systems Electricity consumption [kwh] Aug 6. Aug 11. Aug 16. Aug 21. Aug 26. Aug 31. Aug 5. Sep 10. Sep 15. Sep 20. Sep Ap.Block A Ap. Block B Ap. Block C FIG 4. Daily measure electricity consumption for the ventilation systems TABLE 3. Estimated annual energy consumption per meter square based on measurements [kwh/m 2 /year] Energy consumption Total Heating 87.8 Hot water 27.0 Space heating 60.8 Ventilation (weighted) 2.9 Block A 1.9 Block B 3.9 Block C 2.9 Total energy consumption Moisture and temperature measurements Figure 5 shows the measured temperature and relative humidity in the beam ends. According to Viitanen (1997) there is no risk of degradation of the beam when the RH is less than 75%. As seen Full papers - NSB 2014 page 1185

16 from Figure 5 the RH is less than 75% except from point 3 that has RH>75% during the first month. Point 3 has generally higher RH than the other points. The reason for that may be due to the orientation of this beam end, which is facing straight west and therefore exposed to less sun compared to point 4 and 5. Point 1 and 2 are not mounted in the facade and therefore these points are not as critical with regards to high relative humidity. This is also indicated in Figure 5 where the temperature in point 1 and 2 is higher in October/November. Temperature [ C] Temperature in beam-ends RH [%] Relative humidity in beam-ends Point 1 Point 2 Point 3 Point 4 Point 5 FIG 5. Left: Temperature in the beam ends. Right: Relative humidity in the beam ends. 4. Discussion While the measured and calculated energy consumption before the renovation deviated with only 2.5%, the measured energy consumption of the renovated building was 28% higher than the calculated energy consumption for the renovated building. One reason for this can be the occupant behaviour. It is often seen that when a building undergo a renovation, the occupant feels an increased comfort and a tendency of an increased room temperature is often seen (Harrestrup and Svendsen 2013). Therefore the actual energy consumption is often higher than what has been calculated. The study showed that when the room set point temperature is increased with 2 C, the space heating consumption increases by approximately 30 %. Since the monthly average temperature in 2013 in average is 1.5 C higher than the average temperature from the DRY used in the simulation, the increased consumption cannot be explained based on colder weather conditions in 2013 and therefore increased energy consumption. Another reason could be that when insulating from the inside the thermal capacity of the building is decreased, and therefore less heat might be stored during the day and released during night, which might increase the heating consumption. These reasons are to be investigated but are not within the scope of this paper. The measured and calculated electricity consumption for ventilation is seen to deviate with the actual consumptions being approximately 40-70% less than the calculated consumption. Furthermore the calculated results show that the demand controlled systems are more energy efficient than the traditional central system with constant airflow, whereas the actual consumptions show that the traditional central ventilation system with constant air flow is the most energy efficient system. 5. Conclusions Point 1 Point 2 Point 3 Point 4 Point 5 This paper presents a method on how to energy-renovate an old heritage multi-storey building in Copenhagen and document the energy use before and after the renovation. The calculated energy consumption was found to be reduced by 58% but the actual energy consumption was estimated to be reduced by 42% compared to the measured energy consumption before the renovation. A reason for this deviation can be due to altered occupant behaviour. Three different ventilation systems were installed in the building, one in each apartment block. The results from the simulations showed that Full papers - NSB 2014 page 1186

17 the decentralised systems were slightly more energy efficient compared to the traditional centralised system with constant airflow, whereas the measurements on the actual energy consumption showed the opposite. When applying internal insulation, as was done in the studied building, the risk of moisture problems and wood decay can be present. The temperature and relative humidity measurements carried out in the wooden beam ends showed no risk even in the facade of the apartment exposed to more wind driven rain. In this case study the internal insulation was stopped 200mm above the floor, which can be the reason that no risk of moisture problems seems to be present. This solution can therefore be a durable solution for insulating from the inside and still save a considerable amount of energy. 6. Acknowledgements The results provided in this paper were financed by the Danish Energy Agency under the project: Development and 1:1 demonstration of concepts for renovation of older multi-storey building to low energy class 1. The project team consist of DTU Civil Engineering, COWI, Rönby.dk, and Ecolab. References Danish Energy Agency Grøn Energi vejen mod et dansk energisystem uden fossile brændsler. Copenhagen: Danish Energy Agency (in Danish). Danish Ministry of Climate, Energy and Buildings Vores Energi. Copenhagen: Danish Energy Agency (in Danish). Christensen G. & Bunch-Nielsen T Indvendig Efterisolering Af Ældre Ydermure (BYG-ERFA erfaringsblad (31) ). Ballerup: BYG-ERFA. European Commission, Roadmap for a low carbon economy by Directorate-General Climate Action, European Commission, Bruxelles EU Directive 2002/91/EC of the European Parliament and of the Council of 16 December 2002 on the energy performance in buildings, Official Journal of the European Communities, 4/1/2003. EU Directive 2010/31/EU of the European Parliament and of the Council of 19 May 2010 on the energy performance in buildings (recast), Official Journal of the European Union, 18/06/2010. Engelmark J Københavns Etageboligbyggeri En Byggeteknisk Undersøgelse. Hørsholm: Danish Building Research Institute (in Danish). Harrestrup M., Svendsen S., Changes in heat load profile of typical Danish multi-storey buildings when energy-renovated and supplied with low-temperature district heating. International Journal of Sustainable Energy, DOI: / Hartless R Application of Energy Performance Regulations to Existing Buildings, Final Report of the Task B4, ENPER TEBUC, SAVE /C/00-018/2000. Watford, UK: Building Research Establishment. Kehl D., Ruisinger U., Plagge R., Grunewald J. 2013a. Wooden Beam Ends in Masonry with Interior Insulation A Literature Review and Simulation on Causes and Assessment of Decay, 2nd Central European Symposium on Building Physics, Vienna. Kehl D., Ruisinger U., Plagge R., Grunewald J. 2013b. Holzbalkenköpfe bei innengedämmten Mauerwerk Ursachen der Holzzerstörung und Beurteilung von Holz zerstörenden Pilzen, 2nd International Congress on interior Insulation, Dresden (in German)" Krebs HJ and Collet PF (1981) Indvendig Efterisolering: Indmurede Bjælkeenders Fugt 2 Og Temperaturforhold i Etagekryds. 1st ed. Tåstrup: Danish Technological Institute. Full papers - NSB 2014 page 1187

18 Lechtenböhmer S. & Schüring A The Potential for Large-Scale Savings from Insulating Residential Buildings in the EU. Energy Efficiency 4 (2): Morelli M., Rönby L., Mikkelsen S.E., Christensen M.G., Kildemoes T. & Tommerup H Energy retrofitting of a typical old Danish multi-family building to a nearly-zero energy building based on experiences from a test apartment. Energy and Buildings 54: Morelli M. & Svendsen S Investigation of interior post-insulated masonry walls with wooden beam ends. Journal of Building Physics 36(3): Morelli M., Tommerup H., Tafdrup M.K. & Svendsen S Holistic energy retrofitting of multistorey building to low energy level. Proceedings of the 9th Nordic Symposium on Building physics, Tampere, Finland: Tampere University of Technology, Vol. 3, Session C10, Munch-Andersen J SBi-Anvisning 221. Efterisolering Af Etageboliger. Hørsholm: Danish Building Research Institute, Aalborg University (in Danish). Rasmussen T.V Post-insulation of existing buildings constructed between 1850 and1920. In: Thermal performance of the exterior envelopes of whole buildings XI, Clearwater Beach, FL, 5 9 December. Atlanta, GA: ASHRAE. Ruisinger U Long-term measurements and simulations of five internal insulation systems and their impact on wooden beam heads, 2nd Central European Symposium on Building Physics, Vienna. SAVE SAVE Kortlægning og registrering af bymiljøers og bygningers bevaringsværdi. Copenhagen: Ministry of Culture, Cultural heritage board. Viitanen H.A Modelling the time factor in the development of mould fungi-the effect of critical humidity and temperature conditions on pine and spruce sapwood, Holzforschung- International Journal of the Biology, Chemistry, Physics and Technology of Wood. 51 (1997) Full papers - NSB 2014 page 1188

19 Full paper no: 148 Low Energy Retrofitting in Sweden Thirty-one feasibility studies Katarina Högdal, M.Sc. 1 Emma Karlsson, B.Sc. 1 1 WSP Environmental, Sweden KEYWORDS: Energy efficiency, Retrofitting, Feasibility studies SUMMARY: To accelerate the energy efficiency rate in the residential sector, the Swedish Energy Agencies network for energy efficient apartment buildings has implemented the campaign Halvera Mera (Halve More). The campaign aimed to produce feasibility studies that would identify the measures needed to halve the energy consumption in apartment buildings. A total of 31 studies where completed within Halvera Mera. The studies show that there are significant differences among property owners regarding which measures they are interested in, how they carry out their energy and profitability calculations and which measures they plan to invest in. An analysis of the reports show that municipal property owners, property owners who have their building in a larger city and property owners with buildings that have a high energy usage seem to be more likely to invest in energy efficiency measures. The analysis also shows that the calculated energy savings in the feasibility studies are widely spread. The estimated energy savings range from 30 percent up to 82 percent and the new estimated energy performance range from 20 kwh/m 2 and year to100 kwh/m 2 and year. The estimations, assumptions and results presented in the reports indicate that the property owners knowledge lever differ greatly, as well as the knowledge level of the consultants. To ensure that energy measures are being adequately implemented, and that the targets for energy efficiency will be reached, a general knowledge lift is required. 1. Introduction 1.1 Background In 2007, the European Council decided on energy and climate goals for 2020, the so-called ' targets. These state that by year 2020 greenhouse gas emissions shall be reduced by 20 percent, the share of renewable energy shall increase by 20 percent and energy efficiency must improve by 20 percent, relative to In 2011, a new energy strategy was adopted where energy efficiency was highlighted as a key factor for the long-term energy and climate goals. (SEK(2010)1349) Sweden has set their own climate targets, based on the same foundations as the targets. The Swedish environmental goals for built environment set specific targets for energy efficiency in the housing stock. These targets are that the energy consumption in the residential sector shall decrease by 20 percent by 2020, and by 50 percent by 2050, relative to (2008/09:NU25) The Swedish Energy Agency has a mission to accelerate the energy efficiency rate in the residential sector, and a method that has proven to be effective in Sweden is demonstration projects. The Energy Agencies network for energy efficient apartment buildings, BeBo, has developed the concept Full papers - NSB 2014 page 1189

20 Rekorderlig Renovering (Smart Renovation). The purpose of Rekorderlig Renovering is to ensure that demonstration projects for low energy retrofitting, which will improve the energy performance by at least 50 percent, is carried out. The concept involves a combination of energy efficiency measures, customized to each individual project, which in addition to halve the energy usage also shall result in an increased living comfort, reduced environmental impact and economic benefits for the property owner. To increase the number of demonstration projects in accordance with Rekorderlig Renovering the campaign Halvera Mera was implemented. 1.2 Goal The goal of the campaign Halvera Mera was to produce feasibility studies, in accordance with Rekorderlig Renovering, which identifies the measures and actions needed to halve the energy consumption in apartment buildings while retrofitting. 2. Project Description Halvera Mera 2.1 Scope The goal of Halvera Mera was to produce 25 feasibility studies in accordance with the concept Rekorderlig Renovering. The campaigns target group was property owners in the private and community sectors, as well as housing associations. As the interest for the campaign turned out to be larger than expected, it was decided to expand the number of studies to 35. Despite this, there were still property owners on the waiting list. To encourage these owners to implement energy-saving measures, they were offered an energy inspection. The feasibility study reports were written according to a given template, containing information such as the condition of the building, descriptions of the investigated energy measures, reports of calculated energy savings and LCC analysis for selected measures. A grant of EUR was paid to the property owners who presented a preliminary report to the specified criteria. A total of 31 studies and 17 energy inspections were carried out. After the reports from the feasibility studies were compiled the results were summarized and analysed. 2.2 Implementation Halvera Mera was coordinated by WSP Environmental and led by a project team consisting of: Fred Nordstrom, NENET, NENET Maria Malmkvist (project manager), WSP Environmental (later replaced by Saga Ekelin, WSP Environmental) Bengt Linné, Bengt Dahlgren AB Christina Andersson, WSP Environmental (later replaced by Emma Karlsson, WSP Environmental) Katarina Högdal, WSP Environmental The project group also included Göran Werner, coordinator of BeBo, and Mats Björs, the then president of Byggherrarna AB Feasibility Studies The feasibility studies were conducted by the property owners, with the support from the project team. Some property owners conducted the feasibility study within the company, but many took help of external consultants. BeBo offered support for designing a survey of inhabitants, information about moist problems and in some cases financing for additional measurements. Full papers - NSB 2014 page 1190

21 3. Results 3.1 Feasibility Studies A total of 36 property owners entered the campaign Halvera Mera. During the project some property owners left the project, due to various reasons. At the end of the project 31 reports were submitted and approved by BeBo board. The goal for the property owners was to identify measures that would halve their energy consumption. The average calculated energy saving in the reports is 78 kwh/m 2 and year, corresponding to 54 percent. However, the results are widely spread. The estimated energy saving ranges from 30 kwh/ m 2 and year to 150 kwh/ m 2 and year, equivalent to a reduction of energy use ranging between 30 percent and 82 percent. The new, calculated energy performances are also widely spread. In the reports values between 20 kwh/ m 2 and year and 100 kwh/ m 2 and year are presented, with a mean of 63 kwh/m 2 and year. Figure 1 show which energy measures that have been investigated. The most common measures to examine are the exhaust air heat pump, façade insulation, attic insulation, energy-efficient windows and commissioning of the heating system. Installing energy-efficient windows was the most popular measure. This was examined in 75 percent of the feasibility studies, and 55 percent plan to implement it. About 10 percent opted out the action, usually because it was too costly in relation to the energy saving, and 10 percent had not yet made a decision. When replacing or upgrading with an energy-efficient window a U-value between 0.7 and 1.8 were chosen in this project. The average U-value was 1.1. Commissioning the heating system was not investigated as often as the energy-efficient windows was, but almost as many property owners planned to carry out the action. The implementation of these two measures should reasonably be linked to each other, since an adjustment of the heating system is recommended for greater measures in the building envelope. Facade insulation, both external and internal, was examined in about 70 percent of the feasibility studies, but only 20 percent planned to implement the measure and nearly 30 percent opted it out. This was mainly due to the fact that façade insulation often requires large and expensive actions that can only be motivated if the façade also needs to be renovated. Some property owners chose to go ahead with facade insulation, even though it was unprofitable as an energy measure, to achieve other positive effects such as improved indoor climate. Full papers - NSB 2014 page 1191

22 FIG 1. Energy measures examined in the feasibility studies. The bars show the percentage of the property owners who have studied each measure. The lighter portion of the bars show the fraction that plan to carry out the action and the black portion of the bars shows the proportion that has opted out the action. The grey areas represent the property owner who have investigated the action but has not yet made a decision about implementation. Full papers - NSB 2014 page 1192

23 4. Analysis 4.1 Energy Calculations In Halvera Mera, there were no requirements on how the energy calculations should be performed, or which method or tool that was to be used. This resulted in a differentiation in the calculated energy savings. Table 1 shows the span of the estimated energy savings for some measures. Low estimated energy savings can in some reports be explained by the fact that the energy measure was already implemented, and only needed to be updated or replaced. High estimated energy savings can in some reports be due to that additional measures are implemented at the same time, but are not fully described in the report. However, in most cases it has been difficult to find a good explanation for the large variations. This indicates that there are significant differences in how property owners and consultants perform their energy calculations, and that these differences will affect the result of the energy calculation. TABLE 1. Calculated energy savings for chosen action. Figures presented in kwh/m 2 and year Water supply system Heating system Exhaust Air Heat Pump FTX When the potential energy savings have been calculated, assumptions have been made, as to e.g. the efficiency of heat pumps and the potential of individual measure and billing of water. The large differences in energy reduction suggest that these assumptions also differ. An examination of the energy calculations suggests that the assumptions are often too optimistic, which leads to results that presents unrealistically high energy savings. 4.2 Investment cost Investment costs for the different measurements have been estimated from experiences from other projects or theoretically calculated costs. How the investment costs have been presented differ in the reports. In some reports only the total cost is presented, but in other reports the marginal cost is Full papers - NSB 2014 page 1193

24 described as well. In some reports, it has been difficult to determine whether the given cost intends the total cost or the marginal cost. Figure 2 shows the investment cost for various measures against their potential energy saving. The cost is presented as EUR per square meter A temp (floorarea heated to above 10 C). The energy saving for the various measures is presented in kwh per square meter during the actions presumed lifetime. For installation measures a lifetime of 15 years has been assumed and for construction measures measures a lifetime of 40 years, according to BeBo guidelines. The points represent averages, calculated from the reports. The average investment cost has been calculated from the marginal cost, when presented. The measures found in the lower left corner has a relatively low investment cost, but also gives a rather small energy saving. The measures that are found in the upper right corner provide a greater energy saving, but also have a higher investment cost. In the lower right corner are the measures that could be considered the most cost-effective, that give a fairly large energy saving but still has a low investment cost. FIG 2. Investment vs. potential energy saving. The y-axis presents the cost as EUR per square meter A temp. The x-axis present the energy savings for the various measures is presented in kwh per square meter based on the actions presumed lifetime. For installations a lifetime of 15 years has been assumed and for construction measures a lifetime of 40 years. 4.3 Profitability A requirement in the feasibility studies was to perform both an LCC analysis for each measure as well as a profitability analysis. A summary of the property owners calculations show that the cost of capital rate is set to 5.2 percent at average, while the rate of return is generally a bit lower with an average of 4.8 percent. Profitability requirements and calculations have not always been presented in a similar way, why it has been difficult to get a general picture of the economic circumstances. Since so few property owners have indicated their profitability requirements, it has not been possible to make a deeper analysis on the subject. Full papers - NSB 2014 page 1194

25 In the profitability calculations it has not always been clear whether the measures have been considered as energy measures or as maintenance. Nor has it always been clarified whether the costs that are presented include both maintenance and investment, or only investment cost. However, there seems to be a difference in how property owners carry out their profitability calculations, and if they consider their measures to be investments or as maintenance. When investing in energy-saving measures, and when performing profitability calculations for energy measures, it is important to first determine the buildings need for maintenance, in order to make a correct profitability analysis. For example, if the windows need to be replaced, it is not correct to start from a zero-cost position in your profitability calculations for energy efficient windows. Instead, you should estimate the cost for a conventional window replacement, and use the margin cost for investing in energy efficient window in the profitability calculation. This way, the energy savings that the energy efficient windows gives only have to cover the cost of the actual investment, to be consider profitable. A measure that is performed with the sole purpose to reduce energy usage should be considered as an energy measure, and its cost should be counted as an investment. If a measure is necessary due to needs of restoration, it should instead be considered as maintenance or a raise of standards, even if it results in a reduction in energy use as well. However, this distinction is sometimes difficult to make, since some renovations could include both maintenance and energy reducing measures. In this project some property owners have performed their profitability calculations using the total cost for a measure, even if the measure also met a maintenance need. This made it difficult to find energy investments that were profitable. 5. Conclusions The feasibility studies in Halvera Mera show that there are large differences among property owners in what measures they are interested in, how they carry out their energy calculations and how they implement their profitability calculations. A summary of the results of the feasibility studies show that measures on the heating and ventilation systems and measures in the building envelope are the keys to significant energy savings. The analysis indicates that housing associations, property owners with buildings in small towns and property owners whose buildings have a low energy usage seems to have difficulties to make decisions about these measures. The analysis of the feasibility studies shows three circumstances that make property owners more likely to invest in energy measures. 1. Municipal ownership. Municipal property owners often plan for more and larger investments in energy measures than private owners and housing associations do. 2. Larger cities. Property owners who have their building in a city with more than 140,000 residents have greater ambitions for implementing energy measures than property owners in smaller cities. This is probably linked to better market conditions. 3. High energy usage. Property owners with buildings that have an energy usage greater than 150 kwh/m 2 and year more frequently plans to implement their action than the property owners whose buildings have a lower energy usage. This is probably due to the greater incentive to implement energy measures that comes with a high energy usage. The reports also show that the calculated energy savings in the feasibility studies are widely spread. The estimated energy savings range from 30 percent up to 82 percent and the new estimated energy performance range from 20 kwh/m 2 and year to100 kwh/m 2 and year. The estimations, assumptions and results presented in the reports indicate that the property owners knowledge lever differ greatly, as Full papers - NSB 2014 page 1195

26 well as the knowledge level of the consultants. To ensure that energy measures are being adequately implemented, and that the national targets for energy efficiency will be reached, a general knowledge lift is necessary. 6. Acknowledgements We wish to acknowledge Arne Elmroth, Prof Em at Lund University, for his valuable audit of the calculations in the reports and his advice on how to best carry out the analysis. We would also like to acknowledge Göran Werner and Saga Ekelin at WSP Environmental for their feedback and inputs on the project summary and the analysis. References SEK(2010)1349. Energi 2020 En strategi för hållbar och trygg energiförsörjning på en konkurrenssatt marknad. The European Commission KOM (2010) /09:NU25T. Riktlinjer för energipolitiken. The Swedish Commission for Industry and Trade Full papers - NSB 2014 page 1196

27 Full paper no: 149 Methodology to ensure good indoor environment in energy efficient retrofitting Tuomo Ojanen, M.Sc. 1 Riikka Holopainen, Ph.D. 1 Hannu Viitanen, Ph.D. 1 Jorma Lehtovaara, M.Sc. 2 Paavo Kero, M.Sc. 3 Juha Vinha, Professor 3 1 VTT Technical Research Centre of Finland 2 Aalto University, Finland 3 Tampere University of Technology, Finland KEYWORDS: Building renovation, methodology, indoor environment, moisture, energy efficiency, SUMMARY: The main challenge in renovation is how to focus typically limited resources in the right renovation actions. The target of the renovation is a building with healthy and comfortable indoor conditions, safe moisture performance of structures and improved energy efficiency. Prioritization is needed and it should be based on reliable information and a concept of the performance of the whole building. This paper presents a methodology developed to assist the decision-makers in the renovation process. The objective was to form a practical and simple methodology that is aimed especially for the assessment of the present state of the building and to set priorities for different needs in a renovation project. A three level risk and condition classification is used to form a survey of the indoor environment, structure and system performance and the energy efficiency aspects. The user is guided through the survey using the depth suitable for the building, ranging from check-list to finer analysis. Depending on the performance classification and risk analysis of the building, suitable tools and check-lists are presented to help in the analysis of the different elements. The technical present value of the building before and after renovation can be used to study the costeffectiveness of the renovation actions. A school building renovation project will be presented as a practical application of the methodology. 1. Introduction The motivation for building renovation is typically in problems with the perceived indoor environment, deterioration or aging of structures or other building components or moisture damages. Energy efficient renovation of old building stock is in major role when trying to meet with the regulations aimed to reduce the emission of green-house gases. The challenge in renovation is how to choose the right action to reach the required level of performance: Improved energy efficiency allowing healthy and comfortable indoor conditions with safe moisture performance of structures. There is typically lots of information of the maintenance processes, moisture problems and accidents, possible conditions surveys, energy consumption and earlier renovations through the history of the building. Proper analysis of the present state condition of the building is the requirement for proper design of the renovation so that the typically limited resources can be used in an efficient way. Full papers - NSB 2014 page 1197

28 A methodology and tool has been developed to help in the decision making and design of a renovation. The objective was to develop user-friendly, practical, simple and clear methodology to help the stakeholders in the planning, design and implementation phases of a building renovation process. The aim was to create a concept that could guide the user to through the process. The focus of the methodology is in the present state analysis revealing the renovation needs and in the evaluation of the risk levels of different building systems. The risks are studied from different perspectives: indoor safety and comfort (thermal comfort, air quality, illumination), building physical performance and energy efficiency aspects for different structures and systems. Also a cost study is integrated in the tool. 2. Objectives for renovation methodology The objective of the ENERSIS concept is to form a clear survey of the present state of the building to be renovated. The decisions made in the planning phase of the project are the most important for the success of the renovation. Right renovation action has to be chosen to ensure good results. 2.1 Aims of the methodology The aim was to form a simple and user friendly methodology that could be used to clarify and illustratively present the present state analysis of the building. The methodology is meant only to help and guide in the design phase of the renovation process. The user of the methodology makes decisions and sets the priorities using the results of the methodology when found suitable. The methodology only clarifies and justifies the actions according to the information given by the user (Ojanen & al. 2013). 2.2 Evaluation principles The ENERSIS concept was developed to guide the user of the methodology to take into account not only the primary reason for renovation, but also the indoor environment performance and the overall energy efficiency of the building. This approach includes the building physical performance, thermal comfort, perceived indoor air quality and lighting conditions The methodology advises which aspects should be considered to achieve the set performance goals and how to evaluate the interactions, possible risks and energy saving potentials of different building components. Suitable tools and check-lists are presented to help in the analysis of the different elements. The methodology reveals the risks linked to a straight -forward renovation which does not consider the overall performance of the building or the risks and potentials that are included. The costeffectiveness of the renovation can be evaluated by solving the apparent technical value of the building before and after different renovation actions. 3. Structure of the methodology tool The methodology was formed in an Excel format tool. The structure of the methodology and the development principles of the tool are presented in the following. Figure 1 presents the decision making process of a building renovation and Figure 2 the structure of the tool to help in the methodology application. 4. Renovation tool application in a case study The use of the renovation tool was studied in a renovation case study. The case study was carried out for the renovation process of a school building built in This school building had reported to have poor indoor air (IAQ) and environment (IEQ) quality and modest energy efficiency. Full papers - NSB 2014 page 1198

29 The renovation process was followed during the research project and the methodology could be applied in this case study by using the information from the site. The conditions and results of the application of the methodology are presented in the following. Present state of the building Risk and competence classification Need for more information Renovation needs Critical factors Additional surveys Concept user : Budget, costs and payback, property development principles Setting the objectives Possible renovation solutions Analysis of the effects of different renovation solutions Selection of renovation method, scale and solutions FIG 1.Decision making process of a building renovation. Building identification Risk and competence classes Indoor environment Structures and systems Energy efficiency Value, before / after Risks of microbial growth Technical life expectancy Instructions for the user FIG 2. Structure of the Enersis-tool for the evaluation of renovation needs. 4.1 Indoor environment The symptoms on bad IAQ and thermal comfort were evaluated based on the building user inquiries. The feedback from the users of the buildings is the first step when evaluating the present state of the building and possibly needed additional surveys and focused conditions studies. Indoor air quality in many times related to insufficient ventilation and or also moisture damages causing bio-deterioration of structures. Several different organisms can be involved in the moisture Full papers - NSB 2014 page 1199

30 problems and damage of buildings, but their effect on IAQ are often indirect: particles, smell and Microbial Volatile Organic Compounds (MVOC:s). Other aspect of indoor climate is the thermal comfort. This more technical orientated performance depends on the temperature, air flow velocity and partly also humidity conditions that can be studied by measurements. In the site case the indoor conditions evaluation was based on user questionnaires and further microbiological studies and the evaluation results are presented in FIG 3. Class 2,5-3 1,5 2,5 1 1,5 1 < 1 Definitions Essential need for renovation, severe risks included Renovation needed, increased risks Performance lower than required, low risks Fulfills current requirements Better than required FIG 3. Classification categories for risks and renovation needs (left) and the presentation of the evaluation results. Continuous line for average levels, dashed for maximum values of each sub area. 4.2 Structures and systems The critical structures and systems for heat and moisture performance were evaluated. The design and implementation requirements vary by structures and the analysis was done for different parts of the building envelope and ground supported structures. The moisture performance of the building envelope parts can be evaluated. One of the most suitable criterias for the moisture performance analysis is the risk of bio-deterioration that can be studied numerically using mould and decay models incorporated with the hygroscopic simulations (Viitanen 1997a, 1997 b, Viitanen et al 2000, 2010, Ojanen et al 2010). Mould development depends on factors like ambient temperature, exposure time and the type and surface conditions of building material. Mould typically affects the quality of the adjacent air space with volatile compounds and spores. Degradation in one structure or system increases the risk for damage to other structures. Therefore it is important to pay attention also to the starting degradation. Other evaluation criterion in this section is the performance of design solutions, their functionality and fault tolerance. For example wood-frame walls can be more sensitive to the excess of moisture than massive brick walls, and generally ventilated structures are more tolerant against additional moisture than unventilated structures. In addition to structures, also the present state of the building systems (ventilation, heating, plumbing) were evaluated. The evaluation results for the school building are presented in Figure 4 (left). 4.3 Microbial condition of building and materials The role of microbes is important for the performance of buildings and for indoor air quality (IAQ). Visual inspection of structure and analysis of the eventual damage of the building are the first step to Full papers - NSB 2014 page 1200

31 evaluate the condition of the materials and structure. Especially the details where moisture can be accumulated during the service live of the building are important to notice. In the next step, more detailed analysis of the critical parts of the structures should to be performed, when also samples from materials will be taken. Based on the analysis, an evaluation of the condition of the building will be performed using a systematic process. Different methods can be used. The direct microscope analyses are also very useful especially for damage analytic. Figure 4 (right) presents the results of the evaluation of the school building. The evaluation is based on different microbial analysis of the structures of the school building. Three different levels of classification were used: 1) no microbial problems, 2) local problems or microbial growth, 3) severe or large microbial problems or damage. The most attacked structures were the inside parts of outdoor walls caused by water penetration through the brick façade. Also the outer part of constructions in ground were wet and attacked by micro-organisms. A part of the floor structure (crawl base) was partially attacked by decay. FIG 4. Evaluation results of the structures and building systems (left) and microbial condition of a school building (right). Continuous line for average levels and dashed line for maximum values of each sub area. 4.4 Lighting The main goal of the lighting renovation process is to improve lighting quality and energy efficiency and to achieve light levels mentioned in latest recommendations. User acceptance and visual comfort is improved with advanced daylight use, modern controls and attractive luminaire outlook. Renovation of lighting system is normally a part of large building renovation process of the building. The lighting evaluation part of the tool helps the experts to find the biggest risks of the old lighting systems and to make decisions for the future lighting plans. The check list of lighting tool concern age of installation, quality and quantity of lighting, types of light source and luminaires, lacks and faults of luminaires and parts, controls and sensors, use of daylight and overall energy efficiency. Again, the three level risk classification is used to present the current and future risks of the lighting system. The lighting evaluation is presented in Figure 5 (left). The first and even sufficient reason to renew the old lighting installation is the high age of fixtures. Expected life period of the lighting installation is typically 25 years. During this period all the components, especially plastic lamp holders and light diffusing materials, electrical parts like ballasts are normally aged. Energy efficiency of old luminaires may be poor due to missing or darkened reflectors and diffusers or glare shields. Lamp types may be out of date even, in some cases like incandescent and mercury lamps, already denied by the European directives. Full papers - NSB 2014 page 1201

32 With LED technology it is already possible to gain better energy efficiency than with the present T5 tubular fluorescent technology. The main disadvantages of led general lights are higher luminaire price, so far limited selection of systems for general indoor lighting, and only moderate glare reduction due to the lacks in intensity distributions. 4.5 Energy efficiency The normalized specific heating energy use of the school building was 48 kwh/m 3 and the specific electricity use 8.3 kwh/m3 (year 2009). The original energy class of the school building was class D (Finnish energy classification of buildings) representing a quite average energy use. This reasonable high energy class is due to the insufficient ventilation rate. Besides being under-dimensioned, the existing mechanical outlet ventilation system also had no heat recovery. The renovation will increase the energy-efficiency of the school building by means of ventilation heat recovery, increased insulation levels of wall and roof structures and a new energy-efficient lighting system. The evaluation of the energy efficiency of the original school building is presented in Figure 5 (right). FIG 5. Evaluation results for lighting quality (left) and energy efficiency (right) of the school building. The new mechanical inlet and outlet ventilation system with heat recovery will increase the electricity use of the building. Some of this increase can be compensated using on-demand ventilation, which was estimated to reduce the average ventilation airflow by %. Also the more efficient fans will reduce the electricity use of the ventilation system: the specific fan power (SPF) of the original outlet air fans was estimated to be 2.5 kw/(m3/s), whereas that of the new inlet and outlet ventilation fans was estimated to be 1.9 kw/(m3/s). With these assumptions the annual electricity use of the ventilation system would increase 9 MWh. Including lighting improvements the total effect of the renovation on the annual electricity use is -10 MWh. The estimated decrease of the annual heating energy use of the building is 400 MWh, which represents the annual saving of with the average heating energy price of 54 /MWh. The renovation will increase the energy class of the school building from class D to class A and the total annual energy saving due to the renovation is Besides energy savings the new ventilation system will also improve the thermal comfort of the occupants and improve the indoor air quality, which have been linked to better learning results of the pupils. Full papers - NSB 2014 page 1202

33 4.6 Renovation cost evaluation Renovation actions affect both book value and technical value of a real estate. Book value is usually calculated subtracting the annual depreciation from the value of the new building and adding the cost of repairs. Renovation actions are not always properly aligned, damaged structures may remain unrepaired and undamaged structures may be repaired. Such renovation does not raise directly the book value as calculation model assumes. Therefore, book value and technical value may differ significantly. New calculation model to estimate technical value of real estate was developed in research project. In the model building is divided in seven parts: Roof, external wall, ground-supported and bathroom constructions, and plumbing, heating and ventilation systems. Firstly, each part is defined as the percentage of its share in the total value of the building by its current condition. Secondly, each part is evaluated by the risk for moisture problems by its design solution. A combination of condition and risk evaluation gives a comprehensive approximation of the value of building. Model can be used to evaluate technical value before and after renovation. This gives valuable information when deciding the extent of renovation or prioritizing renovations. The technical value of case building before and after renovation is shown in Table 1. TABLE 1. The technical value of case building before and after renovation. Area brm 2 Price of new building 2400 /brm 2 Before renovation After renovation Structure Cond. % Risk % Cond. % Risk % Roof constructions 2 9 % 2 5 % 1 13 % 1 1 % External wall constructions 3 5 % 3 6 % 1 14 % 2 4 % Ground-supported constructions 2 7 % 2 5 % 2 7 % 2 5 % Bathroom constructions 2 6 % 3 7 % 1 9 % 1 1 % Plumbing systems 2 7 % 3 9 % 1 12 % 2 7 % Heating systems 2 5 % 2 6 % 1 10 % 1 2 % Ventilation systems 3 6 % 2 4 % 1 16 % 1 1 % Total 45 % 42 % 81 % 21 % 26 % 64 % Price Price Discussion and conclusions The renovation methodology and tool developed in a project are described in a concise way. The methodology guides the user through the renovation process, emphasizing the present state analysis and the evaluation of the renovation needs. Different levels of analysis depths can be applied varying from a check-list approach to detailed performance analysis. The concept methodology tool only guides the user to take into account the needed actions and gives information about the performance requirements. The final decisions remain always to the user who is responsible for the aims and allowed costs of the renovation actions. A case study was carried out using the information of a school building. The case study showed that using this approach the risks related to different parts and systems of the building can be better taken into account than without a systematic tool. The simple three-level presentation of the risks and Full papers - NSB 2014 page 1203

34 renovation needs of building structures and systems reveals the actual problem areas and helps to prioritize the renovation actions. Also the interactions and reciprocal risks of different renovation solutions are easier to comprehend. The improvement of the indoor conditions were evaluated by numerical thermal comfort analysis of different renovation solutions. Survey after the renovation is ready will show the effects on indoor air conditions and energy efficiency. The renovation cost evaluation shows the best ways to increase the value of the building. This approach takes into account the present and renovated conditions and risk evaluations of different parts of the building and gives a comprehensive approximation of the value before and after selected renovation actions. Using this methodology the renovation process will lead to solutions where the existing problems of the old building will be sorted and mended and the renovation will be directed to good energy efficient solutions with healthy and comfortable indoor environment. Due to the simple structure of the tool it is easy to use and update according the new requirements, research results and user needs. 6. Acknowledgements Project ENERSIS (A Concept Ensuring High Indoor Environment Quality and Structure Moisture Performance in Energy Efficiency Renovations) was financed by TEKES (The Finnish Funding Agency for Technology and Innovation), VTT (Technical Research Centre of Finland), City of Helsinki and industry, and it was carried out at VTT Technical Research Centre of Finland, Aalto University and Tampere University of Technology during References Viitanen, H. 1997a. Modelling the time factor in the development of brown rot decay in pine and spruce sapwood - the effect of critical humidity and temperature conditions. Holzforschung, vol. 51, 1, pp Viitanen, H. 1997b. Critical time of different humidity and temperature conditions for the development of brown rot decay in pine and spruce. Holzforschung, vol. 51, 2, pp Viitanen, H; Hanhijärvi, A; Hukka, A; Koskela, K. Modelling mould growth and decay damages. Healthy Buildings 2000: Design and Operation of HVAC Proceedings. Espoo, 6-10 August Vol. 3. SIY Indoor Air Information, pp Viitanen, H; Vinha,J.; Salminen, K.; Ojanen, T.; Peuhkuri, R; Paajanen, L.; Lähdesmäki, K Moisture and bio-deterioration risk of building materials and structures. Journal of Building Physics, vol. 33, 3, pp Ojanen, T., Viitanen, H., Peuhkuri,R., Lähdesmäki, K., VinhaJ., Salminen K. Mould growth modeling of building structures using sensitivity classes of materials. Thermal Performance of the Exterior Envelopes of Whole Buildings XI. Clearwater Beach, Dec CD. ASHRAE, DOE, ORNL. Atlanta, USA (2010). Ojanen, Tuomo; Holopainen, Riikka; Viitanen, Hannu; Lehtovaara, J.; Vinha, J.; Kero, P. Methodology to integrate energy efficiency, safe moisture performance and indoor environment quality in building renovation projects. 2nd Central European Symposium on Building Physics, CESBP 2013, 9-11 September 2013, Vienna, Austria. Proceedings. WienUniversity of Technology (2013), pp Full papers - NSB 2014 page 1204

35 Full paper no: 150 Individual energy savings for individual flats in blocks of flats Anker Nielsen, Professor 1 Jørgen Rose, Ph.D. 1 1 Danish Building Research Institute, Aalborg University, Denmark KEYWORDS: Energy savings, individual flats, lower temperature, variation SUMMARY: It is well known that similar flats in a block do not have the same energy demand. Part of the explanation for this is the location of the flat in the building, e.g. on the top floor, at the house end or in the middle of the building. It is possible to take this into account when the heating bill is distributed on the individual flats. Today, most blocks of flats have individual heat meters to save energy and to ensure a fair distribution of the cost. If all flats have the same indoor temperature, the distribution is correct. In practice, the inhabitants of the different flats maintain different indoor temperatures. The result is that heat flows between individual flats. This decreases the energy consumption in the flat where the owner maintains a lower temperature. The neighbouring flats will have higher energy consumption. Calculations were performed for Danish blocks of flats from 1920, 1940, 1960 and Normally, we expect the reduction in energy consumption to be around 20% for a 2 C lower temperature, but for an inner flat the reduction can be up to 71%. The owners of the adjoining flats get an increase in energy demand of 10 to 20% each. They will not be able to figure out whether this is because the neighbour maintains a low temperature or the fact that they maintain a higher temperature. The best solution is to keep your own indoor temperature low. We can also turn the problem around: if you maintain a higher temperature than your neighbours, then you will pay part of their heating bill. 1. Introduction It is a well-known fact that energy consumption in dwellings varies significantly depending on the number of inhabitants and their individual behaviour. This has been documented in several previous publications, e.g. (Hiller 2003) showing measurements of variations in energy consumption for 38 individual single-family houses over a period of 10 years, (Pettersen 1997) presenting the mean value and standard deviation of energy consumption in more than 900 flats spread over 9 different blocks and (Mørck 2011) showing the variations in heating energy consumption for 64 individual housing units. For blocks of flats, it is interesting to know not only the energy demand of the entire building, but also the energy demand of each flat. It is typical that flats immediately below the roof, above an unheated basement and at the building ends have a higher energy demand. This is the effect of different heat losses, but another factor also has an impact. That is the indoor temperature. A lower indoor temperature results in energy savings that are very economic as it does not cost any money. If we calculate with Danish climate conditions, a lowering of the temperature by 2 C will result in a 20% energy saving or 10% per C. That is the case if you live in a single-family house, where you can control the temperature yourself. If we lower the temperature by 2 C in all flats in the block, we get the 20% energy saving. But that is not the case when we consider lowering the temperature in a individual flat. Here, a lowering of the temperature gives a much higher energy saving as you receive heat from your neighbours if they do not lower their temperature. It is important to be aware of this effect if you make individual Full papers - NSB 2014 page 1205

36 measurements of the heating demand and perform calculations of the expected heating bill. It is normal in Denmark to have a central heating system and each flat pays part of the total heating bill of the building. Typically this is based on measuring the indoor temperature or the heating consumption of each flat. 2. Energy calculation The calculations of the energy demand presented in this paper is done with the model described in Nielsen 1980 based on a monthly energy balance. The model can calculate multi-zones, where each flat is a zone and staircases, basement etc. are other zones. The zones can be heated to a fixed constant temperature or the temperature in each zone can be dependent on the heat balance with other zones. In each zone, the heat loss and ventilation loss to the outdoor air and the heat gain from solar radiation, persons and equipment is taken into account. The heat flow between individual flats, i.e. transmission heat flow, is also taken into account in the model. The outdoor climate is the Danish standard climate data. The outdoor temperature is given as a monthly mean value. Solar radiation through a typical pane is given as a monthly sum depending on the orientation of the window. The calculation method also takes into account that the heat gain cannot always be fully utilized and could instead give rise to overheating. This is determined by the energy balance calculated month by month. For each heated zone (as flats), a heating demand is calculated for a constant indoor temperature of 21 C. The heat flows between zones can be positive or negative depending on the temperature difference. This method is similar to that of Be10 (Aggerholm and Grau 2005), which is used for most small buildings in Denmark today. But Be10 only considers the building as a single zone. For multi-zone models, it is normal to use much more complex models that calculate the energy balances hour by hour, but the effect of lowering the temperature in individual flats the model by (Nielsen 1980) should be sufficient. 3. Selected buildings In a thesis work Rasmussen 1980 calculated energy savings and energy economy in blocks of flats from different time periods. The buildings were selected as typical for the period, but it was important to have drawings and descriptions as a basis. All U-values and areas were calculated and the method by Nielsen 1980 was used in the energy calculations. The thesis work presented different calculated energy savings for individual energy-saving measures such as new windows or extra wall insulation and their economy. The selected buildings were: Struenseegade, Nørrebro, København, 1920, built from bricks, with wooden floors and single glazing. Bispeparken, Bispebjerg, København, 1940, built from bricks, with floors of concrete and single glazing. Hedeparken, Ballerup, 1960, industrialised construction built from concrete, with wooden façade elements and double glazing. Tinggården, Herfølge, 1980, low-dense buildings with good insulation. Eremitageparken, Lyngby, 1972, modern industrialised buildings. This paper concentrates on Struenseegade from 1920 and Hedeparken from 1960, but the effect of later changes in the building methods, insulation levels etc. are discussed. 4. The case: Struenseegade The buildings in Struenseegade were built in They are 5 storeys high and consist of two- and three-room flats. Calculations were done based on drawings and descriptions of the flats, and in order Full papers - NSB 2014 page 1206

37 to represent all types and locations in the buildings, 40 flats were included in the analysis. The overall insulation level of the buildings was extremely poor with a solid brick wall thickness of cm with a U-value ranging from 1.0 to 1.5 W/m 2 K. Heat flows between individual flats depend on the nature of the partition walls and floors. In 1920, Denmark did not have any particular requirements concerning thermal insulation of buildings, and there were no requirements concerning sound insulation either. In this particular building, the partition walls consisted of brick and the floors were wooden joists with clay deposits. The U-value of the partition walls were 2.2 W/m 2 K and the U-value of the floors were 0.7 W/m 2 K and therefore differences in indoor temperature in individual flats have had a huge impact on the energy consumption in the surrounding flats. The block of flats was grouped with other buildings meaning that it did not have gable flats. Figure 1 shows the individual energy demand of flats around 4 staircases in MWh per year calculated at an indoor temperature of 21 C FIG 1. Energy demand of flats around 4 staircases in MWh per year Figure 1 shows the energy demand of the flats around four staircases, if the indoor temperature was 21 C in all flats. Flats nos. 1 to 5 were located at the building end but since the building was grouped with other buildings, they had no extra heat loss. The four staircases, nos. 6+7, 18+19, and as well as the basement 49 were unheated. Flats nos were three-room flats while the rest were two-room flats. The result was that the flats on the ground floor like nos. 1, 8, 13, 20, 25, 32, 37 and 44 had a higher energy demand than more centrally located flats. A similar effect was seen on the top floor with flats nos. 5, 12, 17, 24, 29, 36, 41 and 48 which had an even higher heat demand due to the extra heat loss through the roof. The calculated energy demand of the flats varied from 7.0 MWh/year for a two-room flat placed in the centre of the building to 28.8 MWh/year for a three-room flat at the top of the building. Full papers - NSB 2014 page 1207

38 Temperature lowered by 2 C Entryway 6 and 7 Left Flat Above Flat Right Flat Left Flat Below Entryway 6 and 7 Right Lowering everywhere -20 FIG 2. Energy demand of flats nos. 1-5 and 8-17 if the temperature is lowered by 2 C in one flat. Small numbers are flat numbers, large numbers are changes in energy demand in % compared with normal. Negative numbers are energy savings and positive numbers increased energy consumption Temperature lowered by 2 C Entryway 18 and 19 Left Flat Above Flat Right Flat Left Flat Below Entryway 18 and 19 Right Lowering everywhere -20 FIG 3. Energy demand of flats nos and if temperature is lowered by 2 C in one flat. Small numbers are flat numbers, large numbers are changes in energy demand in % compared with normal. Negative numbers are energy savings and positive numbers increased energy consumption Full papers - NSB 2014 page 1208

39 Figures 2 and 3, had the same layout as the flats located in the building. For example, if we look at flat (3) in Figure 2; if the temperature in this flat is lowered by 2 C, the savings will be 34 %. In turn, this increased the energy consumption of the neighbouring flat (10 ) by 1% for the upstairs neighbour (4) 6% and the downstairs neighbour (2) 7%. Results for the other flats can be read in a similar manner. For a single-family house you would expect savings corresponding to approximately 20% if the temperature was lowered by 2 C, but here the results showed significantly higher savings. This was due to the transmission of heat to the flat from adjacent flats. The savings of two-room flats were: Top floor (on the roof) approx. 30% Between floors approx. 45% Lower floor (against basement) approx. 30% As a result of these savings, the energy consumption of the adjoining flats increases by up to 11%. The exact figures are shown in Figs. 2 and 3. The building's total energy consumption is almost unaffected by individual flats lowering temperature. 5. The case: Hedeparken The buildings in Hedeparken were built in 1960 as one of the large industrialised buildings consisting of four-storey blocks of flats. All flats in this block are of the same size three-roomed. The calculation was performed based on the drawings and descriptions of the flats. The calculation was performed for three staircases in a block of flats. The thermal insulation of 75 mm mineral wool was typical for the period around The windows had double glazing with a U-value of 2.50 W/m 2 K. The walls were a light wooden prefabricated solution with a U-value of 0.44 W/m 2 K. The floor between the cellar and the flats had a U-value of 0.60 W/m 2 K. The roof had a U-value of 0.48 W/m 2 K. The heat flows between the individual flats depend on the constructions in the building. The Danish Building Regulations from 1960 specify rules for the sound insulation between the flats but there are no regulations concerning heat flow. To achieve good sound insulation, it is important to use heavy constructions, e.g. concrete. This, however, results in a high U-value. The partition walls between flats were 150 mm concrete with a U-value of 2.8 W/m 2 K. For floors between flats the U-value was 1.35 W/m 2 K FIG 4. Energy demand of flats around three staircases in MWh per year The drawing in Figure 4 shows the energy demand for the flats around three staircases, if the indoor temperature was 21 C in all flats. Flats nos. 1 to 4 are located at the gable with extra heat loss. The three staircases nos. 5, 14 and 23 were calculated as unheated. The only heat comes from the adjoining flats and solar radiation. Below the block of flats was a basement no. 28, which was also unheated. The result was that the flats nos. 1, 6, 10, 15, 19 and 24 on the ground floor had a higher energy demand than most inner flats. The calculated energy demand of the flats varied from 5.9 MWh per year in an inner flat to 11.2 MWh per year for the flats at the top floor at the gable. The average Full papers - NSB 2014 page 1209

40 temperature in the staircases had a variation from 17.0 C in December to 21.2 C in July. The variation in energy demand gives a variation on the energy bill of the flats, also if we had the same indoor temperature and free heat from solar radiation, electricity and persons. In real life, the energy bill depends on the inhabitants use of the flats, i.e. some have a higher temperature and some could have more ventilation and also an effect of the variation in the number of inhabitants in each flat. Temperature lowered by 2 C Flat Above Flat Right Flat Left Flat Below Entryway 5 Left Entryway 5 Right Lowering everywhere -20 FIG 5. Energy demand of flats nos. 1-9 if temperature is lowered by 2 C. Small numbers are flat numbers, large numbers are changes in energy demand in % compared with normal. Negative numbers are energy savings and positive numbers increased energy consumption Figure 5 shows the result of a calculation on the energy demand of flats nos. 1-9 if the temperature was lowered by 2 C from 21 C to 19 C. This is not unrealistic if you want to save energy and reduce your energy bill. For flats in the middle of the building, like nos. 7 and 8 the energy demand for heating and ventilation would be reduced by 71%. The result was of course that flats above, below and next door to the left and right had a higher energy demand and a higher bill. The increase of these flats was 7-18%. If we take an example flat no. 8, the saving were 71%. For the flat above no. 9, the increase was 11%. For the flat below no. 7 the increase was 18%. For the neighbours to the right, no. 12, the increase was 16%. For the neighbour to the left, no. 3, the increase was 7%. For the top floor, the saving was 47% and for the first floor above the basement 38%. The energy saving was less if you lived in a flat at the gable. The savings were: Top floor at the end and under the roof: 35% Floors 2 and 3 at the end: 54% First floor at the end and above basement: 38% All the savings and extra heat demand of the neighbours are given in Figure 5. These savings correspond to the obtained 20% saving if all flats reduced the indoor temperature by 2 C. All calculation were performed with a lower temperature but we can reverse it so a 2 C higher temperature for a flat in the inner part of the building gets a 71% higher energy demand. If you have a Full papers - NSB 2014 page 1210

41 higher temperature than your neighbours, then you get a high energy bill as you also pay for your neighbours heating. They would save 7-18% on their individual energy bills. 6. Discussion Calculations for the other buildings show similar effects of variations in energy demand between the flats and that lowering the temperature in an individual flat gives much more saving than the 20% for 2 C. The numerical values can be less but the effect of the heat flow between the flats is very important for the savings. Is this still relevant today many years after the calculation? The answer is yes. Many of the buildings have been extra insulated or had new windows so the energy demand of the flats has been reduced. But you cannot do anything with the heat flow between flats. So there is still a heat flow between the flats if we lower the temperature and the saving from a lower temperature is the same or slightly higher. New buildings are typically built with constructions as described in Rasmussen 2011 in order to obtain the necessary sound insulation. There are no rules regarding thermal insulation between flats. Looking at the cases in the report, we can find the best insulated constructions. For the floor, this is 25 mm insulation and for the wall it is 50 mm insulation material, if the construction is made as cavity wall with a 60 mm space with 50 mm mineral wool. In reality, this wall construction is seldom used as it is quite complicated to build. We can now look at Hedegården and look at the changes that will occur if extra insulation is added. For walls between flats, the extra 50 mm insulation decreases the U-value from 2.80 W/m 2 K to 0.62 W/m 2 K. For floors between flats, the extra 25 mm insulation decreased the U-value from 1.35 W/m 2 K to 0.73 W/m 2 K. If we again look at flat no. 7, Table 1 shows the effect of a change in the constructions. TABLE 1. Savings with original wall/floor or better insulated wall/floor between flats Original New constructions Realistic Flat 7-71% -40% -58% Flat 8 above 18% 10% 14% Flat 6 below 13% 7% 10% Flat 2 left 7% 2% 5% Flat 11 right 16% 4% 13% The new constructions reduces the heat flow between the flats as flat 8 goes from 18% to 10% increase of energy demand. The savings in flat 8 is now reduced to 40%, but still a very large effect. The new constructions are a theoretical case as there are some practical problems. If we look at the solution with 25 mm insulation; this is only possible under a wooden floor and not under inner walls in the flat and probably not under bathrooms and toilets. So there will be thermal bridges and some areas that do not have 25 mm insulation. We must also remember that some of the sound insulation solutions do not have thermal insulation. A best estimate is given in the last column. For the wall between flats it is, as mentioned, more complicated to build the wall as two separate walls with thermal insulation between, so only few houses will have this solution. A more realistic estimate is therefore shown in the right column of Table Conclusions The calculations presented in this paper show that it is much easier to achieve energy savings by lowering the temperature in a flat in the central part of a block than for flats closer to the top, bottom Full papers - NSB 2014 page 1211

42 or end of the building since it will receive heat from the flats around it. This effect is still relevant even in new blocks of flats as the floors and walls between individual flats typically do not have specific thermal insulation. The most efficient way of saving energy is that all inhabitants know that you increase savings if the neighbours do not lower the temperature. If all flats lower the temperature, we will achieve the highest total energy savings. Is it possible to determine whether your neighbours are stealing your heat? This is probably not possible as there will always be variations in the behaviour and number of people in the flats. So you have to take the risk, but we could also look at it, so that if your neighbour maintain a high temperature you will achieve an energy saving. References Aggerholm, S. and Grau, K. 2005; Bygningers energibehov - Pc-program og beregningsvejledning. (Building energy demand PC program and user guide) SBi-Anvisning 213. Statens Byggeforskningsinstitut (SBi), Hørsholm, Denmark Hiller, C. 2003; Sustainable energy use in 40 houses A study of changes over a ten-year period, Report TVBH-3044, Department of Building Physics, Lund Institute of Technology, Lund University, Sweden. Mørck, O., Thomsen, K. E. and Rose, J. 2012; The EU CONCERTO project Class 1 - Demonstrating cost-effective low-energy buildings - Recent results with special focus on comparison of calculated and measured energy performance of Danish buildings, Applied Energy, Vol. 97, , pp Nielsen, A 1980: Beregning af ruminddelte bygningers energiforbrug. Metoderne EFB2 og EFB3 (Calculation of energy demand for building divided in rooms), meddelelse nr. 103, Thermal Insulation Laboratory, Technical University of Denmark Pettersen, T. D. 1997; Uncertainty analysis of energy consumption in dwellings, NTNU, Trondheim, Doktor ingeniøravhandling 1997:122, Høgskolen i Narvik, Norway. Rasmussen, N.H. 1980: Energibesparelser og energiøkonomi i etageboliger, (Energy savings and energy economy for blocks of flats), Thesis work, Thermal Insulation Laboratory, Technical University of Denmark Rasmussen, B Lydisolering mellem boliger nybyggeri (sound insulation between flats) SBianvisning 237, Statens Byggeforskningsinstitut (SBi), Hørsholm, Denmark Full papers - NSB 2014 page 1212

43 Full paper no: 151 Renovation vs. demolition of an old apartment building: energy use, economic and environmental analysis Kalle Kuusk, M.Sc. Simo Ilomets, M.Sc. Targo Kalamees, Professor Sten Tuudak, M.Sc. Andre Luman, M.Sc. Tallinn University of Technology, Estonia KEYWORDS: renovation scenarios; energy performance; cost effectiveness; embodied energy; apartment buildings. SUMMARY: The paper analyses four renovation scenarios for one concrete element building type. These scenarios were: major renovation, low-energy renovation, low-energy renovation with extensions, demolition of the original building and construction of a new building. Results reveal that in the current case an existing building can be renovated to meet the same energy-efficiency levels as a new building. Demolition of an existing building and construction of a new one raises the global cost at least fourfold. Analyses of embodied energy via CO 2 emissions from the materials and energy production for a building during 20 years show that a new building has higher environmental impact than low-energy renovation. Therefore, the condition and low energy efficiency of an old concrete element apartment building are not the reasons to consider its demolition. 1. Introduction Increasing energy prices and energy saving policies have shifted attention to the energy performance of dwellings. EU has made a commitment to reduce the emission of greenhouse gases by 20% by the year 2020 compared to the level of Estonia has set the goal of maintaining the final energy consumption at the same level as in However, this will require a decrease in energy use and an increase in energy efficiency. Energy economy and heat retention were included as basic requirements for construction works (Construction Production Regulation 2011) earlier but sustainable use of natural resources is a recent addition. This means that focus should also be on the durability and low long-term environmental impact of the construction process and exploitation (often expressed via carbon dioxide CO 2 emission). The design of renovation raises the question of the extent and economic viability of renovation. Frequent discussions also address the demolition of an existing building and the construction of a new building. The agenda of Tallinn Vision Council contains a target to demolish 103 of the oldest prefabricated concrete large panel element apartment buildings (hereinafter: concrete element buildings) in Tallinn (Sarv 2013). The concept targeted to demolishing existing buildings introduces new economic and environmental challenges. Previous studies have pointed out that such areas as exact embodied energy values, the costs and applicability of refurbishment, direct energy impact of demolition and its wider environmental impact still remain unclear. However, both broad arguments and concrete evidence support maintaining a focus on renovation rather than on large-scale demolition (Power 2008). Full papers - NSB 2014 Page 1213

44 Results of research covering the current technical condition of Estonian old concrete element housing stock refer to satisfactory condition in terms of load-bearing but to insufficient energy performance, indoor climate and hygrothermal performance of the building envelope (Kalamees 2011). Also, durability of concrete façades has been found to be problematic regarding to corrosion and frost damage (Ilomets 2011). This study analyses different renovation scenarios for one concrete element building type. The aim was to find out how renovation, renovation with extensions, and construction of a new building affect the energy efficiency, economic viability, and embodied energy. 2. Methods 2.1 Studied building The study object was a five-storey apartment building with prefabricated concrete large panel elements (TABLE 1, FIG 1) constructed in That type of construction was typical in Estonia and in other countries in Eastern Europe during the period In total, there are almost 3500 old concrete element buildings in Estonia (National Register of Construction Works). During renovation works in 2011, additional insulation was placed to the building envelope, old windows were replaced, a new two-pipe heating system and a ventilation system with heat recovery were installed. TABLE 1. Characterisation of the studied building Net area, m Heated area, m Number of apartments 60 Compactness: Building envelope, m 2 / volume, m FIG 1. Picture of the studied building before (left) and after major renovation (right). 2.2 Simulations and calculations Indoor climate and energy simulations were made for different stages of the building: original building without any renovation measures major renovation renovation on a low-energy building level renovation on a low-energy building with extensions of the building demolition of the original building and construction of a new building Full papers - NSB 2014 Page 1214

45 As occupant behaviour related to energy usage is variable and unrelated to the building type, the energy calculations were made at standard indoor climate and by a unified calculation methodology. The methodology is specified in local regulations (Estonian Government s Ordinance No. 68 and Ministry of Economic Affairs and Communications Ordinance No ). Energy performance of buildings was calculated with dynamic simulations using the IDA Indoor Climate and Energy 4.5 simulation program. During the first phase of our calculations, simulation models for the pre- and the post-renovation stage were validated using the measured indoor climate and energy consumption data from the building. In the second step, validated models were calculated at standard usage and energy efficiency packages for a low-energy building, for a low-energy building with extensions, and for a new building were composed (TABLE 2). TABLE 2. Variables of the simulation model Variables Without renovation Major renovation Lowenergy Low-energy (extensions) New building Thermal transmittance, W/(m 2 K) : walls U wall, roof U roof basement ceiling U basement windows U window doors U door Additional insulation, mm: external wall roof basement ceiling Air leakage rate q 50, m 3 /(h m 2 ) HVAC systems: heating 1-pipe 2-pipe, 2-pipe, 2-pipe, 2-pipe, thermostats thermostats thermostats thermostats ventilation natural exhaust air apartment apartment apartment ventilation heat pump based AHU based AHU based AHU renewable energy solar collectors Extensions were attached to kitchens and staircases in the simulations of low-energy buildings with extensions. Additional space was used to accommodate the ventilation air handling units and increase the small floor area of the existing kitchen. Solar collectors were installed for heating domestic hot water (DHW) to compensate the increased heat loss caused by the additional constructions. The construction of the new building scenario followed the principle that its energy efficiency would be higher than minimum requirements and would comply with the low-energy requirements for energy efficiency. Estonian Test Reference Year (Kalamees and Kurnitski 2006) was used to simulate outdoor climate conditions. Full papers - NSB 2014 Page 1215

46 Primary energy (PE) was used as the indicator for the energy efficiency. The requirements for the apartment buildings were as follows: major renovation: PE 180 kwh/(m 2 a) new buildings: PE 150 kwh/(m 2 a) low-energy buildings: PE 120 kwh/(m 2 a) According to the energy source, the use of the primary energy and the environmental impact were taken into account with the weighting factors: district heating 0.9 electricity Economic analysis The global cost (EN 15459, Eq. (1)) calculation was used to assess the cost effectiveness of different renovation strategies. 20 Ci C i ai j C ( ( ) 1 g ) Afloor Rd ( i)) ( (1) where: C g (τ) global incremental cost ( /m 2 ) C i initial investment cost ( ) C a,i (j) annual cost year i for component j (energy cost) ( ) R d (i) discount factor for year i A floor net floor area ( m 2 ) A period of 20 years was selected because the maximum period for renovation loans for apartment owners associations in Estonia is 20 years. Global cost was calculated at the interest rate 4%. To show sensitivity to the escalation rate, five escalation rate scenarios were considered: 1% escalation, 3% escalation, 5% escalation, 7% escalation, and 9% escalation. Construction costs (TABLE 3) were taken from a database of apartment owners associations containing reports of their real renovation costs and from the estimations of construction companies. The energy price levels used were 0.14 /kwh for electricity and /kwh for district heating (mainly based on gas). TABLE 3. Construction costs Scenario Cost, Cost, /net m 2 Major renovation Low-energy Low-energy + extensions New building Environmental analysis Environmental impact of the five scenarios was analysed via the emission of CO 2 during the renovation/construction and 20 years of exploitation. An existing building before the renovation was chosen as the reference point, which means that CO 2 emissions of the construction materials used in 1966 and CO 2 emissions of energy carriers until the major renovation in 2011 were excluded from the analysis. Embodied energy of the renovation scenario was calculated for the quantity of each material from CO 2 emissions of the construction materials based on the literature. Also, thermal energy and electricity consumption were calculated and multiplied with CO 2 emissions to produce that energy. Full papers - NSB 2014 Page 1216

47 To simplify the analysis, the following aspects were not taken into account: transportation energy and water demand at the building site workmanship HVAC systems installed into the building and the solar collector at low-energy with an extensions scenario small details and fixing (glue mortar, fastening, sealing foam and tapes etc.) thin layers (floor covering, rendering, filler, colour) transmission loss of district heating and electricity from the plant to the building site It was also assumed that the impact of recycling the materials from a demolished building is negligible when replacing the existing building with a new one. The reason was that no dangerous waste (asbestos etc.) originates from the existing building and the majority of the remaining materials can be reused to a landfill construction site nearby. A minority of materials unsuitable for reuse can be sorted and handed over to the licenced company of construction waste management. CO 2 emissions of the construction materials and the energy carriers used in the analysis are presented in TABLE 4. TABLE 4. CO 2 emissions of the construction materials and the energy carriers (Kurnitski 2011 and Hegger 2008) Material CO 2 emission, Energy carrier CO 2 emission, kg/co 2 eq/kg t/mwh Expanded polystyrene 3.4 District heating Stone wool 0.99 Bitumen polymer sheeting 1.21 Glass 0.66 Precast concrete element Steel 0.73 PVC 2.28 Gypsum board Results (mainly based on gas) Electricity (mainly based on oil shale) Energy usage Delivered energy calculation results showed that the use of the delivered space heating energy can be decreased from 153 kwh/(m 2 a) to 15 kwh/(m 2 a) (FIG 2 left). The low-energy renovation scenario with extensions has a higher space heating energy need (32 kwh/(m 2 a)) than the low-energy scenario with the current building body shape (19 kwh/(m 2 a)) due to decreased compactness and additional linear thermal bridges. Solar collectors are used for heating DHW to compensate the increased heat loss through the building envelope. Use of the primary energy in the standard usage is shown in FIG 2 right. Electricity accounts for the largest share of the primary energy consumption in different renovation scenarios. For further reduction of the primary energy, it is necessary to reduce the electricity demand. Comparison of the energy use for low-energy renovation and for a new building shows no substantial differences. Thus, existing buildings can be renovated to meet the same energyefficiency levels as new buildings. Full papers - NSB 2014 Page 1217

48 Delivered energy, kwh/(m 2 a) Without renovation 131 Major renovation Low-energy Low-energy (extensions) New building Heat pump electricity Auxiliary electricity Household electricity DHW Space heating Primary energy, kwh/(m 2 a) Without renovation 175 Major renovation Low-energy Low-energy (extensions) New building Electricity DHW Space heating FIG 2. Delivered energy usage (left) and primary energy usage (right) of different renovation strategies. 3.2 Economic analysis The global cost was selected to assess the cost effectiveness of renovation strategies (TABLE 5). Before the renovation stage, the global cost is lower than in other scenarios because the calculations do not take into account the maintenance costs. If the pre-renovation stage is taken as the reference point, the escalation should be 9% for the global cost to decrease in the renovation scenarios. The implemented low-budget major renovation has the lowest global cost values in the renovation strategies. Low-energy renovation with extensions has ca 15% higher global cost than the low-energy renovation without additional extensions. Demolishing of an existing building and building a new one has ca four times higher global cost than the low-energy renovation and the low-energy renovation with extensions. TABLE 5. Global incremental cost values Renovation Global cost, /net m 2 Scenario Escalation 1% Escalation 3% Escalation 5% Escalation 7% Escalation 9% Without renovation Major renovation Low-energy Low-energy (extensions) New building Environmental analysis Embodied energy of a renovation scenario can be expressed as a sum of CO 2 emitted from the production of the construction materials and CO 2 emitted from the energy production that a building consumes during the period of 20 years. Results presented in TABLE 6 indicate that the smallest embodied energy (2924 tons) is achieved by renovating an existing building to the low-energy level, being also lower than the reference case before renovation. In that case only 51 tons of embodied energy originate from the materials (1,8% of total); that is close to major renovation but the impact from energy consumption has decreased notably, especially from electricity. Extensions in the case of low-energy lead to higher need for materials and energy that ends up with a higher total value because of deteriorated compactness. Construction of a new building with the same volume has about nine times higher need of resources for materials; that leads to higher total embodied energy compared to low-energy renovation, despite of a small energy consumption during exploitation. Full papers - NSB 2014 Page 1218

49 TABLE 6. Emission of CO 2 from the production of construction materials and from energy production Scenario CO 2 emissions, t Materials District heating Electricity Total Without renovation Major renovation Low-energy Low-energy (extensions) New building Subdivision of material emissions in case of a new building shows that most of the emissions originates from load-bearing walls, slabs and roofs (all together ca 89% of total). Percentage of insulation materials constitutes about 8% and windows approximately 3% of total material emissions. 4. Discussion Our results showed that in terms of energy efficiency, economic viability, and embodied energy, no direct reasons exist to demolish old concrete element buildings and build new apartment buildings. Energy performance of existing low-energy and low-energy buildings with extensions is close to that of a new building; however, the construction cost of a new building is about four times higher. Also, the environmental impact of a new building as a renovation scenario is the highest (15% of embodied energy comes from the materials) and the improvement of energy performance during the exploitation will have no further impact. Our result that demolition and constructing a new building has higher environmental impact refers in principle to similar conclusions found in previous studies (Ireland 2008 and Yates 2006). They report that equivalent refurbishment can be as green as new buildings but difference is rather small and depends on a case and chosen time period. Since all renovation scenarios have lower total environmental impact compared by status quo (without renovation), we have proven the need for renovation of older housing stock from environmental aspect. It should be noticed that some of the additional factors related to a new building (transportation, HVAC systems, construction waste management) were excluded in our analysis, in the opposite case, the difference between renovation vs new building would even have been larger. Load bearing structures are not a critical issue as the condition of the main load bearing structures was found to be sufficient (Kalamees 2011). Therefore, the condition and low energy efficiency of old concrete element buildings are not the reasons to consider their demolition. Tallinn Vision Council has pointed out that floor planning of these old dwellings is unsuitable for families (Sarv 2013) because of small-sized bathrooms and kitchens. In addition, in the five-storey buildings, narrow staircases and absence of elevators restrict movement of families with small children and older people. Demolition is a plausible solution when some region is intended to be thoroughly renewed. At higher volumes, the construction costs would be lower and a larger macro-economic impact would be also an important factor, but here further detailed analysis is required. On a single building level, renovation is substantially cheaper than building a new dwelling. The number of old concrete element buildings reveals a potential solution in favor of renovation due to enormous construction capacity. Power (2008) has stated that even with the highest feasible level of demolition, the existing stock would remain the dominant energy challenge in the built environment far into the future. Focus should be on sustainable design from the materials that contain a low amount of energy, on the use of local materials and the durability of buildings during both renovation and new construction. 5. Conclusions Main findings of this study reveal that in the current case an existing building can be renovated to meet the same energy-efficiency levels as a new building and demolition of an existing building and Full papers - NSB 2014 Page 1219

50 construction of a new one raises the global cost at least four-fold as compared to renovation. Analyses of embodied energy via CO 2 emissions from the materials and the energy production used in a building during 20 years demonstrate that a new building has higher environmental impact than lowenergy renovation. Therefore, energy efficiency, economic and environmental issues show no support to the idea of demolition instead of renovation. 6. Acknowledgements This work was supported by institutional research funding IUT1 15 Nearly-zero energy solutions and their implementation on deep renovation of buildings of the Estonian Ministry of Education and Research, Reducing the environmental impact of buildings through improvements of energy performance, AR12059 (financed by SA Archimedes) and Research European with Social Foundation financing task Cooperation of Universities and Innovation Development, Doctoral School project Civil Engineering and Environmental Engineering code References Construction Production Regulation Regulation (EU) No 305/2011 of the European Parliament and of the Council of 9 March 2011 EN 15459, Energy performance of buildings economic evaluation procedure for energy systems in buildings, November Estonian Government s Ordinance No. 68. Minimum requirements for energy performance of buildings. RT I, , 4, Hegger, M.; Fuchs, M.; Stark, T.; Zeumer, M. Energy Manual. Birkhäuser Verlag AG Ilomets, S.; Kalamees, T.; Agasild, T.; Õiger, K.; Raado, L.-M. (2011). Durability of concrete and brick facades of apartment buildings built between in Estonia. International Conference on Durability of Building Materials and Components. Porto, Portugal:, Ireland, D New Tricks with Old Bricks. The Empty Homes Agency, London. Kalamees, T.; Kurnitski, J Estonian Test Reference Year for Energy Calculations, Proceedings of the Estonian Academy of Science, Engineering 12 pp Kalamees, T.; Õiger, K.; Kõiv, T-A.; Liias, R.; Kallavus, U.; Mikli, L.; Lehtla, A.; Kodi, G.; Arumägi, E Technical condition of Prefabricated Concrete Large Panel Apartment Buildings in Estonia. International Conference on Durability of Building Materials and Components. Porto, Portugal:, Kurnitski, J Lessons learnt from Viikki Synergy building sustainable development design competitions: proposed criteria for sustainability. Ministry of Economic Affairs and Communications Ordinance No. 63. Methodology for calculating the energy performance of buildings; RT I, , 1, National Register of Construction Works, Power, A Does demolition or refurbishment of old and inefficient homes help to increase our environmental, social and economic viability? Energy Policy 36 pp Sarv, M Mustamäe maha, püsti uus Mustamäe! Eesti Päevaleht, 19 January, (in Estonian). Yates, T Sustainable Refurbishment of Victorian Housing. BRE Press, Bracknell. Full papers - NSB 2014 Page 1220

51 Full paper no: 152 Building Performance Simulation software for planning of energy efficiency retrofits Thomas Fænø Mondrup, Ph.D. Jan Karlshøj, Associate Professor Flemming Vestergaard, Associate Professor Department of Civil Engineering, Technical University of Denmark, Denmark KEYWORDS: Retrofitting, energy efficiency, building performance simulation, information flow SUMMARY: Designing energy efficiency retrofits for existing buildings will bring environmental, economic, social, and health benefits. However, selecting specific retrofit strategies is complex and requires careful planning. In this study, we describe a methodology for adopting Building Performance Simulation (BPS) software as an energy conscious decision-making tool. The methodology has been developed to screen buildings for potential improvements and to support the development of retrofit strategies. We present a case study of a Danish renovation project, implementing simulation-based approaches to energy efficiency retrofits in social housing. To generate energy savings, we focus on optimizing the building envelope. We evaluate alternative building envelope actions using procedural solar radiation and daylight simulations. In addition, we identify the digital information flow and the exchange requirements for each simulation. 1. Introduction 1.1 Background to study The impetus to energy efficiency comes from a variety of sources. In the European Union (EU), the Commission has adopted an action plan aimed at achieving 20% reduction in primary energy consumption by 2020, the goal. This reduction will require major improvements in the energy efficiency of buildings, which represent around 40% of the EU s total consumption (European Union 2009). Recently, the EU s drive to reduce consumption mainly focused on new buildings. However, considering that the average lifetime of a building is over 50 years, and a complete renewal of the existing European building stock would take more than 100 years (Kaderják et al. 2012), a substantial reduction in total consumption will not occur if no energy is saved through retrofitting existing buildings (Verbeeck et al. 2005). Selecting specific retrofit strategies is a complex endeavor with many actions to be considered. A decision support approach is therefore needed (Kolokotsa et al. 2009). Here, Building Performance Simulation (BPS) software has an important role to play (Peltormäki 2009). With the evolution of information technology (IT), virtual models have been developed to simulate the performance of buildings (Doukas et al. 2009). Consequently, today s simulation software allows any aspect of a retrofit strategy to be simulated and assessed before it is built, helping project team members to understand the implications of their choices and to make informed decisions (Beaven 2011). 1.2 Multifaceted study Based on the above, this study has two goals. The first is to explore the current approaches to energy efficiency retrofits in the Architecture, Engineering, and Construction (AEC) industry. The second is Full papers - NSB 2014 page 1221

52 to describe a methodology to facilitate software as an energy conscious decision-making tool. The methodology has been developed to screen buildings for potential improvements and to support sustainable retrofit strategies. Using an integrated and experience-based approach (Towns 2001), the study goals are addressed by: (1) a review of trends in the field of energy efficiency retrofits to establish a knowledge base, and (2) a case study of a Danish renovation project to explore the effect of BPS software as an informative decision-making aid. 2. Methodology 2.1 Review of current approaches A review of current approaches to energy efficiency retrofits has been conducted and included articles and research conducted by academic institutes; industry work practice; and guidelines generated by government institutions. The review was carried out to understand and identify current trends in energy efficiency retrofits, and specifically focused on the uptake of integrating BPS software as a tool for design decision-making. 2.2 Qualitative case study research A qualitative case study of energy efficiency retrofits in Danish social housing has been conducted. The case study approach facilitates in-depth, multi-faceted explorations of complex issues in their real-life settings (Crowe et al. 2011). In the present case study, multiple context-specific retrofit actions were compared to identify and evaluate trade-offs and post-retrofit benefits, which were defined as reduced energy consumption and improved indoor environmental quality. To achieve improvements, the case study retrofit actions focused on optimizing the building envelope. A key feature of this case study was that BPS software was used to predict the influence of the investigated retrofits. In particular, the researchers used a comprehensive suite of solar radiation and daylight simulations to show how building performance is affected by specific retrofit choices. Solar radiation simulations were performed using Autodesk Ecotect Analysis (Autodesk 2011); daylight simulations were performed using IES Virtual Environment (IES 2012). Both Ecotect and IESVE use data interpolation from the EPW weather file. As part of the simulation process, the case study identified the task/tool-specific exchange requirements for each simulation, that is the required data input for each solar radiation [Ecotect] and daylight simulation [IESVE]. Based on these simulations, knowledge was provided prior to the decision-making/retrofit planning, thereby facilitating an informative decision approach (Attia 2012). Notably, the primary purpose of this case study was to demonstrate the benefits of adopting BPS software as an informative decision-making tool for pre-retrofit investigations, not to present specific building performance figures. Based on a Triple Helix of university-industry-government interactions, an interdisciplinary project team of clients, project managers, contractors, architects, engineers, and manufacturers collaborated in the case study (Etzkowitz 2003). Here, representing the university and engineering link, the corresponding author of this article was involved in simulation and design activities. 3. Review 3.1 Energy efficiency retrofits Retrofitting is the process of modifying something after it has been manufactured (City of Melbourne 2013). For buildings, energy efficiency retrofits are defined as actions that allow an Full papers - NSB 2014 page 1222

53 upgrade of the building s energy and environmental performance to a higher standard than was originally planned (Jaggs et al. 1999). An overview of potential retrofit strategies, and retrofit actions which may improve performance figures, is illustrated in FIG 1 below. FIG 1. Retrofit strategies/actions [Inspired by (Kolokotsa et al. 2009)] An example of a retrofitting action is the upgrading of insulation levels. Here, re-insulation of the building envelope external walls, roofs, and floors will improve the energy consumption of the building by reducing thermal losses through the fabric. Another example is the replacement of traditional single/double glazed windows with energy efficient triple glazed windows. As with the insulation upgrade, triple glazed windows will improve the building s thermal performance. Replacing or changing the position, size, and shape of the windows may also result in improved solar gains, and better daylight exploitation, thereby reducing heat consumption and electrical lighting consumption respectively (Bokel 2007). Furthermore, a key feature of retrofit is the objective of improved indoor environmental quality, usually measured by occupant comfort. Indoor environmental quality (IEQ) is determined by several factors, including air quality, acoustics, temperature, and lighting conditions. Consequently, some retrofit strategies integrate natural ventilation for better air quality, thermo-active building systems for thermal stability, and natural lighting for a better quality of illumination (Osso et al. 1996) (Paul et al. 2008). The green agenda is generally a powerful tool in a retrofit argument. However, retrofits also allow an upgrade of functionality, architectural quality, and aesthetic value of the building (Kalc 2012). 3.2 Retrofit performance criteria The planning and evaluation of energy efficiency retrofits depend on well-defined project goals and carefully constructed criteria (Jaggs et al. 1999). Accordingly, the main criteria for efficiency and sustainable performance in a retrofit include: (1) improvement of energy consumption, (2) limited impact on global environment, (3) improvement of indoor environmental quality (IEQ), and (4) upgrading of functionality, architectural quality, and aesthetic value. Furthermore, the expected cost of a specific retrofit is key to its effective value. In this study, however, cost-effectiveness is not included as a criterion for retrofit evaluations. Several of these criteria often appear to be in conflict, for example, energy consumption improvements versus architectural quality. Therefore, finding the optimum retrofit strategy is an optimization procedure. Here, the optimization involves iterative evaluations of proposed retrofit strategies/actions against selected criteria. Therefore, because optimization is complex, efforts for energy efficiency retrofits often focus on specific strategies/actions without the adoption of a holistic approach (Kolokotsa et al. 2009). 3.3 Simulation-based decision-making methodology Generally, decisions taken during the early phases of the design process, where the impact of design decisions on building performance is more significant than decisions made in later design phases, can Full papers - NSB 2014 page 1223

54 determine the success or failure of a retrofit. For this reason, ensuring informed decision-making in the early design phases of both new builds and retrofitting is important (Shaviv et al. 1996). Here, intelligent models and simulation-based approaches can be supportive. In the simulation-based process, a virtual model is developed to identify the most beneficial retrofit strategies/actions through performance simulations. More specifically, BPS software is used to simulate the performance of a virtual model representing a specific retrofit strategy/action. Then, simulation results are evaluated against predefined performance criteria. If the results are not satisfactory, the retrofit strategy/action is modified and the simulation process is repeated (Attia 2012). This iterative procedure is illustrated in FIG 2 below. FIG 2. Iterative decision methodology [Inspired by (Kolokotsa et al. 2009)] 3.4 Simulation-based retrofit design process In contrast to design processes aimed at new-build, the retrofit design process is strongly influenced by the conditions of an existing building. The simulation-based retrofit design process is illustrated in FIG 3 below, here integrating the above mentioned simulation-based decision methodology. As illustrated, the simulation-based retrofit design process consists of three phases: (1) analysis of existing conditions, (2) development of retrofit strategies/actions, including evaluation against performance criteria, and (3) implementation of retrofit strategies/actions. FIG 3. Retrofit design process A key challenge to simulation-based retrofit design processes is the digital information flow between functional phases. In most cases, this information flow is defined by task/tool-specific exchange requirements, that is the required data input for specific BPS software applications. 4. Case study 4.1 Analysis of existing conditions The framework of this case study was directed toward the Gate 21 pilot project Building Envelope Retrofits (GATE ). The aim of this project was to investigate the benefits of energy efficiency Full papers - NSB 2014 page 1224

55 building envelope retrofits in Danish social housing, referring to Strategy 1, implementing retrofit actions related to the building envelope and design aspects. In particular, Gate 21 was looking for creativity in developing multiple exemplar building envelope designs, with the aim of identifying successful actions that could be adopted into future building envelope retrofit projects. Another issue that was highlighted was that of developing building envelope designs optimized for solar radiation and daylight exploitation. FIG 4. Pre-retrofit conditions of case study house The dwelling used for the retrofit case study was a precast concrete construction, 1970s single storey house in Albertslund, Denmark (55.39 N E). Pre-retrofit buildings typically have aging window units, poor insulation, air leakage, and mould growth due to surface condensation. These factors result in increased energy bills and poor indoor environmental quality. FIG 4 shows the house exterior and plan. 4.2 Development of building envelope retrofit actions Based upon review findings, the practice procedure for the development and evaluation of optimized building envelope retrofit actions followed five steps: (1) definition of performance criteria, (2) development of retrofit strategies/actions, (3) building performance simulations, (4) evaluation of simulation results, and (5) retrofit proposals Definition of performance criteria Case study performance criteria were defined to establish a basis for evaluation. Performance criteria were generated with two main purposes: (1) to improve energy consumption by optimizing the exploitation of solar radiation and (2) to improve indoor environmental quality by optimizing the exploitation of daylight. In many cases, such performance criteria will be some combination of potential improvements. For example, optimizing the exploitation of solar radiation may not only improve energy consumption figures, but also indoor environmental quality levels by supporting occupants thermal comfort. Equally, optimizing the exploitation of daylight may not only improve indoor environmental quality levels, but also energy consumption figures Development of retrofit strategies/actions In collaboration with the case study project team, a list of retrofit actions was developed. Since the aim of this case study was to investigate the effects of multiple building envelope designs, basic retrofit actions included re-insulation of external walls and upgrading of existing windows. Specifically for this case study, the influence of selected building envelope design variables was investigated, particularly that of alternative window positions, sizes, and shapes to investigate the resulting effects on solar gains and daylight conditions. The retrofit options consisted of nine building envelope designs/retrofit actions: Action 1: Energy efficient windows. Action 2: Energy efficient windows + increased window width. Action 3: Energy efficient windows + increased window height. Full papers - NSB 2014 page 1225

56 Action 4: Energy efficient windows + extra window section at patio doors. Action 5: Energy efficient windows + double patio doors. Action 6: Energy efficient windows + small skylight in living room. Action 7: Energy efficient windows + large skylight in living room. Action 8: Energy efficient windows + extra window section in living room. Action 9: Energy efficient windows + extra window section in master bedroom Building performance simulations BPS software was used to investigate the retrofit actions. Simulations were performed on two levels: (1) simulation of solar radiation striking exterior surfaces [Ecotect] and (2) simulation of interior solar gains and daylight distribution [IESVE]. Before simulating, specific exchange requirements for each simulation were identified: Site: Global coordinates, weather data, elevation, 3D geometry, context [Ecotect + IESVE]. Building: Global coordinates, orientation, elevation, 3D geometry [Ecotect + IESVE]. Spaces: Elevation, 3D geometry, space boundaries, [IESVE]. External constructions: 3D geometry, U-values, [IESVE]. Internal constructions: 3D geometry, U-values, surface reflectance values [IESVE]. Windows: Orientation, 3D geometry, U-values, g-values, VT-values, shadings [IESVE]. In FIG 5 below, selected solar radiation simulations are illustrated. Here, average hourly solar radiation is mapped over existing conditions, hours in question 06-18, all year, summer, and winter, contour range Wh/m 2. The Ecotect case study models were kept simple, representing outer volumes/exterior surfaces only. As illustrated, surrounding vegetation was not included. FIG 5. Incident solar radiation on exterior surfaces (south view) In FIG 6 below, selected daylight simulations are illustrated. Here, average annual solar gains and daylight distribution are mapped over existing conditions, retrofit Action 1 with energy efficient windows, and retrofit Action 7 with energy efficient windows and a large skylight in the living room, contour range LUX. The base-case model was created to understand the existing conditions of the case study building. This model was used as a reference to estimate improvements from retrofit actions 1 to 9. FIG 6. Correlation of interior solar gains and daylight distribution (top view) Full papers - NSB 2014 page 1226

57 4.2.4 Evaluation of simulation results Based upon the BPSs, several kinds of correlations were demonstrated. For example, FIG 5 shows that an obstructed context greatly influences radiation values. As illustrated, the surrounding wooden fence causes overshadowing, particularly during winter when the sun is low. Therefore, upper parts of the façades and freely exposed roofs should be prioritized when optimizing the exploitation of solar radiation. FIG 6 shows that the replacement of existing windows with energy efficient windows brings significant improvements. Energy efficient windows have smaller frames, allowing more sunlight and daylight to penetrate. In addition, the installation of the large skylight further improves the solar gains and daylight distribution and is particularly effective at bringing solar radiation and daylight into deep spaces/darker areas of the case study building Retrofit proposals The evaluation of simulation results forms a solution space for potential building envelope retrofit actions. This solution space does not define any specific optimum retrofit, rather a wide range of applicable retrofit actions. Nevertheless, installing large window openings will improve solar radiation and daylight exploitation. Note, however, that high intensity solar radiation is the commonest cause of overheating in buildings and should therefore be controlled, for example with adjustable external solar shading. 4.3 Implementation of retrofit strategies/actions The final step is to implement specific building envelope retrofit actions into the case study building. For implementation, the case study project team members should select specific retrofit actions within the developed solution space. This selection process is currently being conducted. 5. Conclusions In the decision-making process of selecting specific retrofit strategies, multiple actions are available. The decision maker has to take into consideration energy, environmental, functional, architectural, and financial aspects to develop a sustainable retrofit strategy. For this purpose, a decision support approach is needed. In this study, the critical role of Building Performance Simulation (BPS) software as an energy conscious decision-making tool was emphasized. In the case study, this was particularly illustrated by solar radiation and daylight simulations results. Based upon this tendency, BPS software is generally evaluated as a useful methodology for planning of energy efficiency retrofits. 6. Acknowledgements The authors would like to thank all of the people and organizations involved in this study, particularly the case study project team members. The work presented in this study was performed in the scope of the pilot project Building Envelope Retrofits, funded by Gate 21. References Attia S A Tool for Design Decision Making Zero Energy Residential Buildings in Hot Humid Climates Université catholique de Louvain, UCL, Department for Architecture and Climate. 298 p. Autodesk Full papers - NSB 2014 page 1227

58 Beaven M Building Information Modelling Across Arup, digital collaboration is redefining the possible in performance and design. ARUP. Bokel R.M.J The Effect of Window Position and Window Size on the Energy Demand for Heating, Cooling and Electrical Lighting. Proceedings of Building Simulation City of Melbourne What is a Building Retrofit? City Of Melbourne, 1200 Buildings. Crowe S., Cresswell K, Robertson A., Huby G. & Avery A., Sheikh A The Case Study Approach. BMC Medical Research Methodology. 11: Doukas H., Nychtis C. & Psarras J Assessing Energy-Saving Measures in Buildings Through an Intelligent Decision Support Model. Building and Environment. 44: Etzkowitz H Innovation in Innovation: The Triple Helix of University-Industry-Government Relations. Social Science Information. 42: European Union Summaries of EU legislation. GATE About GATE IES VE-Pro. Jaggs M. & Palmer J Energy Performance Indoor Environmental Quality Retrofit a European Diagnosis and Decision Making Method for Building Refurbishment. Energy and Buildings. 31: Kaderják P., Meeus L., Azevedo I., Kotek P., Pató Z., Szabó L. & Glachant J How to Refurbish All Buildings by 2050 Final Report June THINK. 72 p. Kalc I Energy Retrofits of Residential Buildings Impact on Architectural Quality and Occupant s Comfort. Norwegian University of Science and Technology, NTNU. 83 p. Kolokotsa D., Diakaki C., Grigoroudis E., Stavrakakis G. & Kalaitzakis K Decision Support Methodologies on the Energy Efficiency and Energy Management in Buildings. Advances in Building Energy Research. 3: Osso A., Walsh, T., Gottfried, D. & Simon, L Sustainable Building Technical Manual: Green Building Design, Construction, and Operations. Public Technology, Inc, USA. 292 p. Paul W. & Taylor P A Comparison of Occupant Comfort and Satisfaction between a Green Building and a Conventional Building. Building and Environment. 43: Peltormäki A ICT for a Low Carbon Economy Smart Buildings. European Commission. Brussels. 48 p. Shaviv E., Yezioro A., Capeluto I.G., Peleg U.J. & Kalay Y.E Simulations and Knowledgebased Computer-aided Architectural Design (CAAD) Systems for Passive and Low Energy Architecture. Energy and Buildings. 23: Towns M.H Kolb for Chemists: David A. Kolb and Experimental Learning Theory. Journal of Chemical Education. 78: Verbeeck G. & Hens H Energy Savings in Retrofitted Dwellings: Economically Viable? Energy and Buildings. 37: Full papers - NSB 2014 page 1228

59 Full paper no: 153 Uncertainty of indoor boundary conditions at calculation of energy consumption for heating of residential buildings determined by inhabitants behaviour Peter Matiasovsky, Dr. Ing 1 Pavol Hrebik, PhD 2 Peter Mihalka, PhD 1 1 Institute of Construction and Architecture, Slovak Academy of Sciences, Slovakia 2 Slovak University of Technology in Bratislava, Bratislava KEYWORDS: Inhabitants behaviour, indoor temperature, air change rate, internal heat gains, uncertainty SUMMARY: The paper is focused on an analysis of the indoor boundary conditions uncertainty influencing the reliability of energy assessment of the buildings for housing. The distribution of input parameters that enter into a calculation of the heat demand for heating, indoor temperature, air change rate and internal heat gains given by the inhabitants behaviour were analysed. From available literature data and from the results of measurements it was proved that the uncertainty of the input energy simulation parameters can be modelled with suitable distribution functions. 1. Introduction The sensitivity analysis carried out for residential buildings energy rating by Corrado and Mechri (2009) highlights that only a few factors are responsible of most energy rating uncertainties given by the inhabitants behaviour: in decreasing order of importance, indoor temperature, air change rate, number of occupants, metabolism rate, and equipment heat gains. Thus, in this study a further investigation was performed on the factors. The indoor temperature, air change rate and internal heat gains were analysed for the best assessment of the energy needs calculation uncertainty. These three factors were supposed to be determined by the behaviour and habits of a household. Each of the factors was specified by the parametric probability distribution. 2. Indoor temperature Thermal comfort is defined as that state of mind in which satisfaction is expressed with the thermal environment (Olesen, 1982). At creating the satisfactory thermal environment the homoiotermia of human organism (keeping the body temperature at certain level of 37 C) is required. The environment is evaluated by the psychophysical scale in which PMV (Predicted Mean Vote) index expresses supposed mean value of thermal feeling of larger group of persons. Fanger (1970) developed the dependence of PMV index on certain combination of activity and clothing and four factors of thermal comfort: air temperature, mean effective temperature of ambient surfaces, relative humidity and air velocity. When all factors of thermal comfort except the temperature are supposed to be constant it is possible to express PMV as a linear function of temperature. At the modelling of indoor temperature uncertainty in various dwellings we considered that each household tries to maintain the constant indoor temperature satisfying its particular thermal comfort requirements. These requirements are Full papers - NSB 2014 page 1229

60 TABLE 1. Values of PMV index for indoor temperature values at activity 1 met, clothing 1 clo and air velocity < 0.1 m/s PMV Temperature ( C) expressed by thermal index PMV. Considering the activity 1 met, clothing 1 clo and air velocity < 0.1 m/s according to the Fanger s thermal comfort equation (Kaclík 1984) PMV = 0 at temperature of 23 C. In Table 1 there are PMV values corresponding the temperatures interval C. The PPD index (Predicted Percentage of Dissatisfied) expresses supposed percentage of dissatisfied, which in given environment will feel thermal discomfort. PPD index (%) is determined from PMV according to standard (ISO 7730:2005): 4 2 PPD = exp( 0, PMV PMV ) (1) With use of Equation (1) from values in Tab. 1 we get the distribution of predicted percentage of dissatisfied with temperature (Tab. 2), which we transformed to the predicted percentage of satisfied with temperature PPS = 1 PPD, in order to express the accepted temperature probability. In Figure 1 the satisfaction of inhabitants with indoor temperature, PPS is approximated as the normal distribution with standard deviation 4 C. A comparison of this distribution with distributions of indoor temperature measured in real dwellings shows that real distributions are significantly narrower. The data determined by Piršel (1989) have the means of and C and standard deviations of 1.82 and 2.18 C for living rooms and bedrooms respectively. The model indoor temperature distribution proposed by Corrado and Mechri (2009) is the normal distribution with the mean of 22 C and standard deviation of 2 C for Sweden. The results of survey presented by Haldi (2010) give the values of 23.5 C and 1.5 C for the mean and standard deviation respectively. The standard (ISO 7730:2005) recommends the following limits for acceptable thermal environment: < PMV < 0.5 or PPD < 10 %, which corresponds the temperature interval from 21.5 to 24.5 C. That recommendation represents a reduction of the interval of acceptable temperatures given by the standard deviation 4 C, to the interval of recommended values with standard deviation 0.5 C (Fig. 1). This, 0.5 C value is supposed by (Jaraminienea & Juodis 2006). The comparison of relative distributions of acceptable, recommended and real indoor temperatures results in a conclusion that real temperatures in dwellings lie in the interval determined by acceptable and recommended values. A real uncertainty of indoor temperature can be modelled by normal distribution with standard deviation 1.5 C. The distribution with standard deviation 0.5 C represents an ideal situation with reliable control system. (2) TABLE 2. Probability of temperature unacceptability PPD or acceptability PPS Temperature ( C) PPD (%) PPS (%) Full papers - NSB 2014 page 1230

61 0,9 0,8 0,7 0,6 (- ) c e n0,5 re c u o e 0,4 tiv la e R 0,3 Living rooms 1987 Bedrooms 1987 PPS normal distribution, SD = 4 C Normal distribution, SD =1.5 C - model Normal distribution, SD = 0.5 C 0,2 0, Temperature ( C) FIG 1. Normal distributions of acceptable (SD = 4 C), recommended (SD = 0.5 C) and real - model (SD = 1.5 C) indoor temperature compared with data from (Piršel 1989) 3. Air change rate The air change in a household depends on weather conditions, air permeability of envelope structures, operating of technical equipment and activity of occupants. The air change in buildings consists of two components: the basic air change provided by infiltration and the air change due to occupants behaviour. The windows opening is characterised by the number of windows open during monitored period per dwelling (IEA-Annex VIII 1987) defined as: N N N t i i i= o = 1 (3) tm Where N total number of windows (-) t i opening duration of particular window (h) monitored period (h) t m The air change rates can be calculated according to universal empirical relationship (De Gids & Phaff 1982) adopted in standard EN (2007) for case of single-sided ventilation. The hourly values of total air change rate in a dwelling n (1/h) represent the sum of minimum required air change rate n 0 due to infiltration and air change rate Δn due to open windows: Full papers - NSB 2014 page 1231

62 2 ( v ( θ θ ) 0.01) 3600 n = n0 + Δ n = Sw 0.5 No i a H + V µ (4) Where S w window area (m 2 ) µ coefficient reducing window area to equivalent open window area (-) H window height (m) V volume of a dwelling (m 3 ) θ i,θ a indoor, outdoor temperature respectively ( C) V wind velocity (m/s). The ventilation heat loss (kwh) is then calculated as: Q v =.361. n. V.( θ θ ) (5) 0 i a Equation (4) was verified in the case study (Matiašovský P. & Koronthályová O. 2003) where a relationship between real duration of the windows opening and the air change rate in identical dwellings characteristic in Slovakia was estimated. In Fig. 2 there is a linear dependence between the number of open windows and the air change rate, for the average weekly values. The found regression corresponds to Equation (4) and gives the actual minimum air change rate n 0 = 0.57 with the regression coefficient a = 0.57.After inserting the actual parameters of dwellings and the actual mean values of wind velocity and air temperature during a heating season into Equation (4) the resulting regression coefficient a = 0.58 which confirms an universality of Equation (4). In Annex VIII (1987) the distribution of average air change rates in dwellings due to windows opening during a heating season was determined. The distribution has an asymmetric character, with minimum 0.1 h -1, median 0.14 h -1 and maximum 0.8 h -1. For arbitrary household, the average air change rate in dwelling during heating season is determined by dwelling inhabitants behaviour and can be expressed as the value of random distribution, introducing the relative average air change rate in a dwelling K: n = n + a N = n0 + a m o N o K 0 (6) 0.5 to e u d te ra g e a n ch a ir k ly e w g e ra v e A ) / h (1 g in n e p o s w o d in w 2,0 1,8 1,6 1,4 1,2 1,0 0,8 0,6 0,4 0,2 0,0 y = 0,5663x + 0,5717 R2 = 0, ,5 1 1,5 2 Average number of windows open during week (- ) FIG 2. Dependence of average weekly air change rate on average weekly number of windows during week (Mihálka & Matiašovský 2007) Full papers - NSB 2014 page 1232

63 0,60 0,50 (- ) 0,40 ce n re cu 0,30 o e tív la e0,20 R Average air change rate by windows opening in flats during heating season Average weekly air change rate due to windows opening in flat Average number of windows open in flat during week Windows opening duration in flat Model 0,10 0, Relative air change rate K (- ) FIG 3. Distribution and uncertainty model of relative average air change rate due to windows opening - exponential distribution (λ = 0.69) The relative air change rate by windows opening in a dwelling K is defined as the ratio of air change rate to median air change rate, or the ratio of number of open windows in a dwelling to median number of open windows: Δn N K = = m m Δn o N o The introduction of relative air change rate K enables to compare the distributions of relative air change rate in dwellings due to windows opening during a heating season with the distributions of relative air change rate and weekly values of number of open windows and windows opening duration in dwellings, presented in (Mihálka & Matiašovský 2007). The compared distributions are similar and compatible mutually, which indicates that the distribution of all windows opening activities lies in one general interval, independently on its time interval scale. In Fig. 3 there is a comparison of the analysed distributions of relative air change rates due to windows opening, as well as relative numbers of open windows and durations of windows opening in the heating season K. As the uncertainty λ K model, the exponential distribution f ( K, λ) = λ e, with λ = 0.69, was identified. The model proposed by Corrado and Mechri (2009) allows the air change rates smaller than minimum required air change rate 0.5 h Internal heat gains Modelling the internal heat gains is based on an assumption that households use the electric energy solely. The analysis issues from the data on yearly consumption of electric energy in households (kwh/year) and on their electrical appliances equipment (Stanek kps.fsv.cvut.cz/file_ download.php?fid=2413), based on the results of detailed analysis of statistical survey of ČSÚ (2003, 2005 and 2011), the outputs of project REMODECE (2008) and the values received from the report JRC-IES (2007). Considering the data presented in Tables 3 and 4 the heat gains from electrical equipment were expressed as a function of the number of household members. The internal heat gains were then calculated as a sum of the heat gains from electrical appliances and constant metabolic heat (7) Full papers - NSB 2014 page 1233

64 gains. The hourly courses of heat gains from electrical appliances represent 90 W per person in an average household and they were developed on the basis of type household electricity load profiles of class 4 ( - the consumption without the use of electricity for heating per year in Slovakia, defining the load expressed by the values of relative hurly loads in the year within the range from 0 to 1. The heat produced by household members was considered as 1 met per person and it was reduced by coefficient 0.8 expressing the percentage of assumed presence in dwelling. The resulting relation for hourly data (kwh) has the form: Q i 2 = TDO 4 90 ( P 0,032 P ) P (8) Where TDO4 type household electricity load profiles of class 4 (-) P number of persons in household (-) The polynomial of second order expresses the dependence of consumption on number of household members in Fig. 4. It is developed from the data in Table 4 where is the energy consumption per household member in relation to the energy consumption of member of average household (Tab. 3) for dwellings with various numbers of persons in household. The distribution of household members has developed during last decades. In Fig. 5 there is the comparison of household size distribution in Slovakia in years: 1961 (Keilman 1987), 1981 and 2004 (Dol & Haffner 2010), 1991 and 2001 (Dzianová 2001), 2010 (Senaj & Zavadil, 2012). The model of household size distribution was chosen from the results of last survey from 2010 (Senaj α α 1 βp β P e & Zavadil, 2012). It is gamma distribution f ( P, α, β ) =, with parameters α = 2.85, β = Γ( α) 1. The model proposed by Corrado and Mechri (2009) supposes the one person households as the most probable ones. TABLE 3. Parameters of average household Period Electricity consumption Number of persons Electricity consumption per person (kwh/person.year) (kwh/year) (-) TABLE 4. Yearly consumption of electric energy in household in dependence on number of inhabitants Number of persons in household Consumption Consumption per person Consumption per person/consumption per member of average household (-) (kwh/year) (kwh/persons.year) Full papers - NSB 2014 page 1234

65 FIG 4. Energy consumption per person in relation to energy consumption per person in average household in dependence on number of persons 0,35 0,30 (- ) 0,25 ce n re 0,20 cu o e0,15 tiv la e0,10 R 0,05 0, Number of persons per household (- ) Model FIG 5. Distribution of household size in Slovakia Uncertainty model of number of persons in household - gamma distribution (α = 2.85, β = 1) Conclusions The consumption of energy for heating of residential buildings is significantly determined by the behaviour of their inhabitants, determining the indoor boundary conditions. The uncertainty of indoor boundary conditions in households: indoor temperature, air change rate and internal heat gains was analysed with the following results. The household temperature uncertainty can be characterised by a normal distribution with the mean value determined by a combination of activity, clothing, relative humidity, air velocity and by the standard deviation determined by an acceptability of the deviation from mean value by household inhabitants. An analysis of the uncertainty of ventilation habits in households showed the correlation of the windows opening activities and air change rates, independently on the analysed time interval. The normalisation of their distributions, dividing by their medians enables to use the exponential distribution of the relative air change rate by opening windows as a universal model. Full papers - NSB 2014 page 1235

66 The internal heat gains have two components. The first of them are the gains from appliances, which can be modelled under the assumption of electricity use only. The appliances amount in a dwelling is a function of the household members. The distribution of household members is developing and changing during longer periods. The contemporary distribution for Slovakia can be modelled by gamma distribution with one resident household as the mode. 5. Acknowledgements The authors wish to thank the Slovak Research and Development Agency APVV, project No for the financial support of this work. References Corrado V. & Mechri H. E Uncertainty and Sensitivity Analysis for Building Energy Rating. Journal of Building Physics 33, De Gids W. & Phaff H Ventilation rates and energy consumption due to open windows: A brief overview of research in the Netherlands. Air infiltration review 4, 4-5. Dol K. & Haffner M Housing Statistics in the European Union. OTB Reasearch Institute for the Built Environment, Delft University of Technology. Dziranová O Households and families according to Census of population and housing in Statistical Office of the Slovak Republic. (in Slovak) EN Ventilation for buildings - Calculation method for the determination of air flow rates in buildings including infiltration. Brussels: CEN. Fanger P. O Thermal Comfort: Analysis and applications in environmental engineering. McGraw-Hill. Haldi F Towards a Unified Model of Occupants' Behaviour and Comfort for Building Energy Simulation, École Polytechnique Fédérale de Lausanne (Thesis). Hrebik P Impact of uncertainty of selected parameters for calculation of heat demand for heating, Slovak University of Technology in Bratislava, PhD thesis (in Slovak). IEA- Annex VIII Inhabitants Behaviour with Respect to Ventilation, Summary. ISO 7730:2005 Ergonomics of the thermal environment - Analytical determination and interpretation of thermal comfort using calculation of the PMV and PPD indices and local thermal comfort criteria. Jaraminienea E. & Juodis E Heat demand uncertainty evaluation of typical multi-flat panel building Journal of Civil Engineering and Management 12, Kaclík J Economical heating of buildings, ALFA Bratislava. Keilman N Recent trends in family houshold composition. European Journal of Population 3, Matiašovský P. & Koronthályová O Passive solar gains versus ventilation by openning windows. Proceedings of 2nd International Conference on Building Physics. Balkema. Lisse, Mihálka P. & Matiašovský P Water vapour production and ventilation regimes in large panel building flats. Building Research Journal 55, Full papers - NSB 2014 page 1236

67 Olesen B. W Technical Review No. 2: Thermal Comfort. Piršel L. et al Determinants of thermal comfort in buildings. Final Report II-8-4/02,1,1. STU in Bratislava, Faculty of Civil Engineering, Bratislava. (in Slovak). Senaj M. & Zavadil T Results of survey of financial situation of Slovakian households. Casual study of National Bank of Slovakia. Bratislava. (in Slovak) Stanek K. Perspectives of further research kps.fsv.cvut.cz/file_download.php?fid=2413. (in Czech) Richardson, I., Thomson, M. and Infield, D. (2008) A high-resolution domestic building occupancy model for energy demand simulations. Energy and Buildings 40, Full papers - NSB 2014 page 1237

68 Full paper no: 154 Effect of orientation on the hygrothermal behaviour of a capillary active internal wall insulation system Valentina Marincioni, M.Sc. 1 Hector Altamirano-Medina, Ph.D. 1 1 University College London, UK KEYWORDS: internal insulation, capillary active, relative humidity, orientation, solar radiation SUMMARY: UK authorities are promoting energy efficiency schemes to improve the performance of buildings as a result of the high levels of energy consumption and consequent CO2 emissions. A quarter of these emissions are due to requirements for space heating. Installation of insulation is one of the most common alternatives to thermally improve buildings, especially on buildings built of solid masonry (~20 percent of the housing stock). However, the thermal improvement of buildings located in conservation areas, listed buildings, decorative façades, or traditional buildings could be only achieved through the use of internal wall insulation. Solid masonry walls with high surface water absorption coefficients have a higher dependence on external climate conditions (e.g. rain, solar radiation), which are likely to affect the performance of internal wall insulation. This paper examines the effect of walls orientation on the hygrothermal behaviour of an internally insulated 16th century building. External walls have been insulated with a capillary active system that allows moisture movement towards indoor environments. Sensors to monitor relative humidity and temperature between the existing brick wall and the insulation were installed in the north-facing and south-facing walls of the building. Both walls are exposed to the same internal environmental conditions (teaching area). The study showed that drying of the south-facing wall occurred faster than drying of the northfacing wall and that drying of the south-facing wall was enhanced by the effect of direct solar radiation. 1 Introduction Improving the energy efficiency performance of the building stock is a priority if the UK is to meet at least 80% carbon reduction by Wall insulation is one of the measures considered for buildings built of solid masonry (~20 percent of the housing stock) (DECC, 2011). Due to planning requirements, traditional buildings built before 1920 (generally buildings located in conservation areas, listed or with decorative façades) and characterised by solid walls are likely to be thermally improved through the installation of internal wall insulation. However, internal wall insulation can undermine the durability of a building, increasing the risk of mould growth and timber decay. Solid walls would be exposed to lower temperatures and, depending on the insulation system applied, there may be an increase in the vapour resistance of the building envelope. This paper presents the results of a case study where the effect of orientation on the heat and moisture transfer within internally insulated solid walls was assessed. The building analysed is a grade II listed barn, with 330 mm solid brick walls and internally insulated with a capillary active insulation system made out of dense woodfibre. Full papers - NSB 2014 page 1238

69 2 Methodology 2.1 The building The building studied is a grade II listed 16 th century barn located in Maidenhead, west of London. The building was refurbished in 2011 and converted into an education centre, featuring a large teaching area, open-plan offices and one exhibition area. The external walls (solid brick) were insulated internally with 100 mm of Pavadentro, a composite board formed by woodfibre and a mineral layer that creates a light vapour diffusion resistance; the total equivalent air layer thickness of the dry composite board is s d = 1.6 m. The existing wall was levelled with a lime-based coat before the insulation was applied and then a bonding coat was used to provide full contact between the insulation and the existing wall. External façades were left exposed, without any impregnation treatment or installation of damp proof courses; therefore not limiting moisture penetration into the building. The building durability could be maintained if potential moisture accumulation were counterbalanced by subsequent moisture drying (a process formed by evaporation and vapour transfer). a) b) FIG 1. Building and external walls insulated a) north facing wall; b) south facing wall (access to teaching area) TABLE 1. Wall assembly Wall construction (outside to inside) Thickness (mm) Brick 330 Levelling coat (3:1 NHL and sand) 0 to 6 Bonding coat (Lime plaster) 5 Pavadentro (composite insulation) woodfibre board 20 mineral layer 1 woodfibre board 80 Internal finish (Lime plaster) 8 Full papers - NSB 2014 page 1239

70 2.2 Monitoring method The aim of the study was to estimate the effect of orientation on the hygrothermal behaviour of internally insulated walls. Four thermocouples and resistive probes (sensors) to measure temperature (T) and relative humidity (ϕ) respectively were installed at the interface between the existing wall and the insulation; sensors (NH, NL) were installed at 30 cm (low L) and at 200 cm from the ground (high H) on the north-facing wall and on the south façade (SH, SL) respectively, as shown in figure 1. Sensors were set with a sampling interval of Δt = 30 min and an accuracy of T = ± 0.7 ºC and ϕ = ± 5 % at 21 ºC. Data were collected for 22 months starting on 01/11/2011 and ending on 11/09/2013. NH, NL SH, SL FIG 2. Building plan with focus on the teaching area and locations of sensors 2.3 Climate The building is located within the Wind Driven Rain zone 2 (where 33 l/m 2 < WDR < 56.5 l/m 2 per spell) in an area where the prevalent wind direction is WSW. Minimum and maximum mean daily temperatures (data from the period ) vary from 1 ºC to 9 ºC and from 7 ºC to 23 ºC respectively. Monthly average sunshine fluctuates from 50 to 200 hours. (Met Office, 2013a) The main source of internal moisture comes from the people occupying the teaching room, generally large groups for short periods of time; occupation patterns depend on the frequency of school visits. The room is naturally ventilated (all windows are openable). 3 Results and discussion Temperature and relative humidity at the interface between the existing wall and the insulation were collected and the profiles related to the north and south orientations compared. The reduction and increase of relative humidity (drying and wetting period respectively) at the interface were analysed and the causes of such events identified. Profiles of relative humidity at the four interstitial locations studied are shown in Figure 3. The initial relative humidity varied from 82 % to 94.6 % as sensors were applied on the surface of the insulation board and surrounded by the bonding coat (wet when applied). The initial moisture of the construction system started drying out after 2 months on the south elevation and 4 months on the north elevation. As shown in Figure 3, there was a reduction of the interstitial relative humidity in both insulated walls, however with a higher reduction observed in the south-facing wall. In both walls, the relative humidity was lower at the end of the study at high level (NH and SH). Drying occurred approximately from January to September, whereas wetting occurred from September to January. Full papers - NSB 2014 page 1240

71 FIG 3. Relative humidity and temperature measured at four location of the existing wall-insulation interface. Drying periods in 2012 and 2013 were characterised by events of sudden relative humidity reduction; sudden reductions were defined as events when at least one sensor shows a decrease of relative humidity ϕ (%), in function of time t (d), for minimum 7 days and with an average rate of 0.5 % per day or above: 0.5 %/d (1) Full papers - NSB 2014 page 1241

72 According to these criteria, three major events occurred on year 1, followed by two on year 2. TABLE 2. Events of sudden decrease in relative humidity ( / t) and difference in the temperature averages Event Start date End date Duration (days) Δϕ/Δt south wall (%/d) Δϕ/Δt north wall (%/d) High Low High Low 1 24/3/12 6/4/ /5/12 3/6/ /7/12 1/8/ /4/13 9/5/ /7/13 20/7/ T south wall - T north wall (ºC) The events of sudden relative humidity reduction lasted from 9 to 14 days, with a rate of decrease from 0.43 to 0.82 % per day in the south-facing wall and from 0.14 to 0.37 % per day in the northfacing wall. A difference in temperature at the interface of the existing wall and insulation was noted during the events of sudden relative humidity reduction (Figure 3, below), which was not observed during the wetting period. Temperatures at the four locations were comparable; however in the drying period, and in particular during the relative humidity reduction events, the temperature at the interface of the existing wall and the insulation on the south-facing wall was higher than the temperature within the north-facing wall. Table 2 shows the average difference of interstitial temperature between the sensors located in the south wall and the ones in the north wall. The difference in temperature seems to be related to orientation and more specifically to solar radiation. Direct radiation would allow an increase in surface temperature in the south-facing wall and a higher rate of evaporation compared to the north-facing wall, which is only affected by diffuse radiation. The observed sudden reductions in the internal humidity of the south-facing wall coincided with events reported by the Met Office (2013b) as remarkably sunny periods: from the 23 rd to 30 th March 2012 and from the 21 st to the 28 th May The same happened in July 2012, a cold and wet month until the 21 st, when the weather became warmer and sunnier. In Figure 4, the monitored relative humidity for the period November 2011 August 2012 was plotted along with daily averaged solar radiation intensity (W/m 2 ); events of sudden relative humidity reduction match periods of consistently high solar radiation intensity. FIG 4. Events of sudden relative humidity (dashed line) reduction and solar radiation (solid line) Full papers - NSB 2014 page 1242

73 4 Conclusion This paper examines the effect of walls orientation on the hygrothermal behaviour of internally insulated walls of a traditional building. Temperature and relative humidity at the interface between the existing wall and the insulation were monitored for a period of two years. It was observed that the walls have a different hygrothermal performance. The south-facing wall presented a faster reduction in relative humidity compared to the north-facing wall. Discrepancy in the performance was found to be associated to wall orientation; in particular solar radiation. The hygrothermal performance of a capillary active internal wall insulation system was found to allow dry-out of moisture within the building envelope enhanced by changes of temperature in the external wall due to longer and constant periods of solar radiation. Other studies have shown that capillary active insulation is beneficial to the building structure helping to reduce the risk of mould growth (Häupl, Fechner et al. 2006, Wegerer and Bednar, 2011). The effect of solar radiation and wind driven rain on the performance of various conventional and capillary active insulation systems is being further investigated. Acknowledgments The authors would like to thank Natural Building Technologies and Technology Strategy Board who funded the research project and to Oxley Conservation for providing the case study building. References DECC Extra help where it is needed : a new Energy Company Obligation. Crown copyright. Häupl P., Fechner H. et al Moisture atlas for building envelopes. 3rd International Conference on Building Physics, Montreal, QC. Wegerer P. & Bednar T Long-term Measurement and Hygrothermal Simulation of an Interior Insulation Consisting of Reed Panels and Clay Plaster. 9th Nordic Symposium on Building Physics, Tampere, Finland. Met Office a. South England: climate. [ accessed on 22/11/2013] Met Office b. UK climate summaries. [ accessed on 28/11/2013] Full papers - NSB 2014 page 1243

74 Full paper no: 155 Hygrothermal Performance of TES Energy Façade at two European residential building demonstrations Comparison between Field Measurements and Simulations Carl-Magnus Capener, Ph.D. 1 Stephen Burke, Ph.D. 2 Simon Le Roux, Architect 3 Stephan Ott, Architect 4 1 SP Technical Research institute of Sweden, Energy Technology, Sweden 2 NCC Construction Sverige AB, NCC Teknik, Sweden 3 Aalto University, Department of Architecture, Finland 4 TU Munich, Chair of Timber Structures and Building Construction, Germany KEYWORDS: E2ReBuild, monitoring, hygrothermal simulations, TES-system SUMMARY: In this study, the retrofitted facades of two European multi-unit residential buildings built in the 1950 s and 1980 s are investigated. The demonstration buildings, situated in Munich, Germany and Oulu, Finland, are part of the EU FP7 project E2ReBuild, a European collaboration project, researching and demonstrating industrialised energy efficient retrofitting of residential buildings in cold climates. The demonstration project in Munich, Germany, consisted of two blocks of residential multi-storey buildings in the suburb of Sendling, built in The buildings were typical examples of the concrete brick constructions, built throughout Germany in the post-war era. The pilot building in Oulu, northern Finland is one of five student apartment buildings in a housing corporation. The building was completed in 1985 according to a Finnish industrialized building system developed in the late 1960's using prefabricated concrete elements for residential buildings, called the "BES system". To improve their energy performance, the retrofit included a façade refurbishment with the TES method utilizing timber based, prefabricated façade elements for the renewal of the building envelope and improved thermal insulation. As part of an advanced monitoring programme, hygrothermal gauges were installed in the walls and they have been monitored for more than one year after the retrofitting. This paper presents the results from the in-situ measurements of the two demonstrations and compares the findings to calculated transient hygrothermal 2D-simulations of the facades utilising the monitored data from the sites in Finland and Germany. 1. Introduction According to European standards and the EU s energy roadmap, the energy performance of multi-unit residential buildings from the 1950 s and later in Europe is poor. External thermal insulation systems are commonly used to improve the thermal performance of such buildings, and for the two selected buildings of the E2ReBuild project the TES-method was chosen for improving the building envelope performance. The TES-system and method utilises timber based and insulated prefabricated façade elements for the renewal of the building envelope and to improve its thermal performance (Lattke 2011, Cronhjort 2014). In this study, the hygrothermal effects caused by the refurbishment are investigated and the TES-system is monitored. This paper presents the findings from two investigated European multi-unit residential demonstration buildings. The demonstration buildings, situated in Munich, Germany and Oulu, Finland, are part of Full papers - NSB 2014 page 1244

75 the EU FP7 project E2ReBuild, a European collaboration project, researching and demonstrating industrialised energy efficient retrofitting of residential buildings in cold climates. The demonstration project in Munich, Germany, consisted of two blocks of residential multi-storey buildings in the suburb of Sendling, built in The buildings were typical examples of the concrete brick constructions, built throughout Germany in the post-war era. The pilot building in Oulu, northern Finland is one of five student apartment buildings in a housing corporation. The building was completed in 1985 according to a Finnish industrialized building system developed in the late 1960's using prefabricated concrete elements for residential buildings, called the "BES system" (Cronhjort 2014). To improve the buildings energy performance, the retrofit included a façade refurbishment with the TES method utilizing timber based, prefabricated façade elements for the renewal of the building envelope and improved thermal insulation. As part of an advanced monitoring programme, hygrothermal gauges were installed in the walls and they have been monitored after the completion of the retrofitting. This paper presents the results from the in-situ measurements of the two demonstrations and compares the findings to calculated transient hygrothermal 2D-simulations (Künzel 1995, Holm 2000) of the facades utilising the monitored data from the sites in Finland and Germany. 2. Description of the demonstration buildings and field measurements The FP7 project E2ReBuild includes seven demonstration buildings throughout northern and central Europe and this paper investigates the hygrothermal performance of timber element system (TES) (Lattke 2011) as external thermal insulation method and compares measured hygrothermal performance with simulation results from WUFI 2D models. In this paper two of these demonstration buildings are presented and monitoring results from their retrofitted north-facing walls are shown. 2.1 Demonstration buildings Background on Munich demo The Munich demonstration was built in 1954 and is located in the suburb of Sendling. It consists of two blocks of residential buildings. They are examples of typical concrete block constructions built throughout Germany after World War 2. ( The heating demand of the building after the retrofit is calculated to be about 21 kwh/m²a. This is about 38 % lower, than the national requirement, EnEV2009. The overall energy demand after the retrofit is equivalent to a primary energy use of 23.5 kwh/m²a. This number is low because it includes a bonus for regenerative energy sources like solar thermal collectors, and a primary energy factor of fp = 0.7 for district heating. The refurbishment concept includes a significant dismantling of the existing dwellings, built from light weight concrete block walls and concrete ceilings. The building was stripped down to the primary structure and the roof was taken off, see Figure 1, left. Additional changes in floor plan layout and new circulation cause interventions on the interior walls as well as on the window openings. A new attic floor and a roof were added together with an entire new building envelope made from TES Energy Façade elements. Full papers - NSB 2014 page 1245

76 Figure 1: Left: Dismantled structure of Munich demonstration with new elevator shaft (Picture: Lichtblau Architects). Right: Oulu demonstration during assembly of prefabricated TES elements (Picture: Simon Le Roux). The highly insulated exterior wall with triple glazed windows is the backbone of the building envelope, see Table 1. The heating system is supplied from the district heating grid. On sunny days it is supported by solar thermal panels on the roof with a large accumulator tank containing litres of water as buffer. Room heating is done by radiators. The apartments have decentralised ventilation units with plate heat exchangers. The highly insulated building envelope, together with a modern and efficient ventilation system with heat recovery, means that the tenants enjoy an energy-efficient apartment with a high level of thermal comfort. Table 1 Facts about thermal performance of envelope and building services, Munich demonstration. before after Exterior walls and roof 1.8 W/m²K 0.15 W/m²K Windows 2.5 W/m²K 0.9 W/m²K Basement ceiling 1.55 W/m²a 0.45 W/m²K Heating energy (calculated) 280 kwh/m²a 21.2 kwh/m²a Primary energy (calculated) 343 kwh/m²a 23.5 kwh/m²a Background on Oulu demo The E2ReBuild demo building in Oulu, northern Finland is one of five student apartment buildings in a housing corporation. ( The Finnish demonstration building underwent a complete retrofitting of the envelope, see Table 12. The old façade layers of the previous BES-systems were removed leaving only the inner concrete layer in place. A new façade was retrofitted using prefabricated timber based elements, see Figure 1, right. The old roof was replaced completely by a new timber truss roof and a new thermal insulation layer of 550 mm resulting in a U-value of 0.08 W/m2K. ISOVER blown loose fill mineral wool, λ=0.041 W/mK. The existing ground floor slab was replaced, with a new in-situ concrete ground floor slab with 200 mm BASF Neopor EPS insulation. Table 2 Facts about thermal performance of building envelope, Oulu demonstration. before After Exterior walls and roof 0.28 W/m²K 0.11 W/m²K Windows 2.1 W/m²K 0.8 W/m²K Ground slab 0.24/0.36 W/m²a 0.11/0.15 W/m²K Roof 0.22 W/m²a 0.08 W/m²a Full papers - NSB 2014 page 1246

77 2.2 On-site hygrothermal monitoring of the facades Part of this project includes an analysis of the TES Energy Façade elements with regards to hygrothermal performance, or the temperatures and moisture performance of the exterior wall. Figure 2 shows where each of the GE HygroTrac sensors was placed in the north façade of the Munich demonstration. wood shuttering formwork 24 mm air layer / lathing horizontal 24 mm gypsum fibre board 15 mm construction wood / cellulose 200 mm adaption layer cellulose 60 mm membrane, Sd-value = 5 m exist. plaster, lime-cement plaster 25 mm existing exterior wall, light-weight concrete building blocks 300 mm exist. plaster, lime-cement plaster 15 mm Figure 2: Monitoring positions of the presented Munich TES Façade element. For the Oulu demonstration in Finland there was a similar set up of monitoring positions of the facades, as shown in Figure 3. Also facing north, the retrofitted wall construction consisted of: TES façade element description: US10YK Declared U-value of the finished wall is 0.11 W/m 2 K 7mm corrugated fibre cement cladding 44 mm air gap x100mm timber battens 9mm gypsum wind barrier mm glass mineral wool slab (Lambda 0,033 W/mK) 42x48mm c600mm horizontal timber battens 42x198mm c600mm timber load bearing frame 9 mm plywood board 50 mm soft thermal insulation 80mm existing precast concrete Figure 3: Monitoring positions of the presented Oulu TES Façade element Measured data For the Munich demo temperature, RH, and moisture content were measured at the measurement points shown in Figure 3 between 2012 and This data was measured every hour and was uploaded to GE s homepage where it could be monitored and downloaded. The sensors are wireless and had difficulties in sending their data every hour during the measurement period so a number of data points are missing. The sensors have also stopped recording data between February and April, Full papers - NSB 2014 page 1247

78 The data has been downloaded and sorted to match the times from the WUFI calculations for comparison purposes. Similarly, for the Oulu demo, temperature and relative humidity is monitored at the measurement points shown in Figure 3 since February Hygrothermal modelling and simulation The hygrothermal behaviour of the demonstration buildings facades has been modelled by the twodimensional hygrothermal building envelope tool WUFI 2D 3.3. The software has been experimentally verified for many types of building component assemblies (Künzel 1995, Karagiozis 2001) and similar set-ups (Holm 2000, Tariku 2006). Material data and initial moisture conditions were supplied from material databases such as MASEA Datenbank (Materialdatensammlung für die energetische Altbausanierung) and the IBP Fraunhofer Material Database. As the buildings are between 30 to 60 years old, the German demonstration originates from the early 1950-ies, there is some lack of precise historic material data and appropriate assumptions had to be made. 3.1 Munich demonstration WUFI 2D simulations were done using the drawing shown in Figure 2 together with measured climate data during the period of January 1, 2012 to October 28, Material properties were mostly taken from the Default materials database in WUFI. The specific material in the existing wall is unknown with unknown thermal and moisture properties. It is known that the material is a type of Leca block with aerated aggregate. The real lambda value of the old wall was calculated using the measured indoor temperatures, temperatures in the adaption layer and temperatures in the exterior part of the mineral wool. These calculations showed that the lambda value of the old wall in reality is between 0.09 and 0.12 W/mK, which is similar to the thermal properties of the default materials Light Expanded Clay Aggregate and Aerated concrete. The default material s lambda value in WUFI 2D was modified to the measured value for the actual wall. Other thermal and moisture properties were obtained based on a report from an on-line database U- wert ( The default moisture properties for Light Expanded Clay Aggregate were used in the calculation since the actual moisture properties of the existing wall are unknown. At the beginning of the calculation, the initial moisture levels of all materials were set to about 80 % RH, based on measured values. 3.2 Oulu demonstration For the Finnish demonstration the WUFI 2D simulations were set-up according to the drawing shown in Figure 3 together with measured climate data for the period of March 2013 to March An initial relative humidity throughout the existing construction of 60 % was assumed since this was an old construction and should not contain any excess moisture and a modest ventilation rate in the air gap behind the cladding of 5 air changes per hour (5 ACH) was selected. 4. Results Both the Munich and Oulu simulated results agree quite well with the measured results in the new wall. The temperatures correlate very well with the measured results. The moisture levels also correlate very well, however the measured data shows much more variation in moisture levels than the simulated data. Full papers - NSB 2014 page 1248

79 4.1 Munich The measured data indicates that the moisture measurement points are in a mineral wool layer. The simulated results are also taken from mineral wool. It is interesting to see that even though they are the same points, the calculations show a much more stable moisture level in the wall than in reality. Figure 4: Temperature and relative humidity for the Munich TES Façade elements insulation, inner (10n) and outer (11n) locations. The moisture levels in the exterior of the wall correlate well however, the calculated moisture level in the middle of the wall (point 9n in Figure 2) does not match the measured values. Further calculations seem to indicate that the plastic may be punctured at the sensor. If we calculate the moisture levels of the wall with the vapour barrier (SD 5m), the moisture level at the interior of the vapour barrier is both more stable and much higher than calculated, while the remaining moisture levels show very good correlation between measured and calculated values. If we lower the SD value of the vapour barrier to that of a weather barrier (SD = 0,1m) the simulated moisture values at point 9n agree with the measured data, however all the other measurement points have more error than simulations with the vapour barrier. This seems to indicate that measurement point 9n is affected by the exterior climate more than it should be if the vapour barrier was complete. Figure 5: Temperature and relative humidity for the Munich TES Façade elements timber stud (12n). In both the calculated and measured results, there does not appear to be a significant moisture risk associated with TES Energy Façade in Munich over the long term. The trend that can be seen in the WUFI calculation is that the construction dries out over time. Full papers - NSB 2014 page 1249

80 4.2 Oulu The measured results correlate well with the simulated figures. It is clear that the initial moisture level has a great influence on the simulated values, the relative humidity is too low for the outer part of the wooden studs but for the inner part the simulation is well in accordance with the monitored result. After some months however, the levels are very close to the measured values, as the monitored wall gets drier. Also, the ventilation rates seem to have an influence, at least for the outer parts of the simulated wall. The low, 5 ACH, ventilation rate give accurate readings during many periods, but often seem to be underestimated as larger fluctuations can be seen in the relative humidity on the inside of the outdoor gypsum board. Another possible reason for the fluctuations can indicate insufficient air tightness over the wind barrier, causing larger fluctuations in the monitored results of the insulation compared to the simulated results. Figure 6: Temperature and relative humidity for the Oulu TES Façade elements insulation, inner (TE/ME260) and outer (TE/ME261) locations. It is clear that the extra insulation placed outside the wooden studs has a beneficial influence on the temperature and relative humidity of the studs. Comparing the relative humidity for the monitoring position of the TES outer insulation (ME261) to the relative humidity at the TES outer part of the timber studs (ME263), clearly shows the reduction in relative humidity. Not only does the insulation break the thermal bridge, it also raises the temperature of the studs outer parts compared to a case without extra insulation, and this gives lower relative humidity and risk of moisture damage. Figure 7: Temperature and relative humidity for the Oulu TES Façade elements timber studs, inner (TE/ME262) and outer (TE/ME263) locations. Full papers - NSB 2014 page 1250

81 5. Conclusions In conclusion, the two cases from Munich and Oulu show that WUFI 2D is a good tool to determine the moisture performance of the TES Energy Façade after a renovation however, the results are very sensitive to the input data such as the existing wall, climate data, the new building materials and if there is any problem with the quality of the work. The demo cases also show that the risks for moisture damage in the form of mould growth in the TES Energy Façade are quite low in both cases for the measured climate. This gives an excellent possibility to evaluate TES Energy Façade with different modifications and in new locations using WUFI 2D before actually starting the retrofit of a building. Initial moisture can pose a risk for wooden construction, both elevated moisture contents in the TES wooden studs themselves, but also in the interior and existing wall material where the TES will be placed can be a source of excess moisture, especially for materials such as concrete and lightweight concrete. For future work it would be interesting to see the effect of built-in moisture in the existing wall on the hygrothermal performance of the external TES timber studs and the risk of moisture damage this would impose. Also, the robustness and sensitivity of the system to moisture from envelope leakage or from transport to construction site is a topic that needs further investigation, as well as the effect of the extra layer of insulation on the timber studs in the Finnish demo compared to the German TES build-up without extra insulation. 6. Acknowledgements This paper is based on research and results from the EU FP7 funded project E2ReBuild Industrialised energy efficient retrofitting of resident buildings in cold climates. The project started in January 2011 and ends in June The project is coordinated by Christina Claeson-Jonsson, NCC AB. The research leading to these results has received funding from the European Union's Seventh Framework Programme (FP7/ ) under grant agreement n References Cronhjort, Y., Le Roux, S Holistic retrofit and follow-up through monitoring: Case Virkakatu, Oulu, Finland, Submitted to the Nordic Symposium on Building Physics 2014, Lund, Sweden Holm, A., Künzel, H.M Two-Dimensional Transient Heat and Moisture Simulations of Rising Damp with WUFI 2d, 12th Int. Brick/Block Masonry Conf. Proc. Vol. 2, Madrid, Spain Karagiozis, A., Künzel, H.M., Holm, A WUFI ORNL/IBP A North American Hygrothermal Model: Proceedings of the Thermal Performance of the Exterior Envelopes of Whole Buildings VIII (Buildings VIII), Clearwater Beach, Florida, USA, December 2-7, 2001 Künzel, H.M Simultaneous Heat and Moisture Transport in Building Components One- and two-dimensional calculation using simple parameters, IRB Verlag, Germany Lattke, F., Larsen, K., Ott, S., Cronhjort, Y TES Energy Façade prefabricated timber based building system for improving the energy efficiency of the building envelope, funded by: Woodwisdom Net, Research project from Tariku, F., Kumaran, M.K Hygrothermal modelling of aerated concrete wall and comparison with field experiment, Proceedings of the 3rd International Building Physics Conference, Montreal, Canada, August 27, 2006, pp Full papers - NSB 2014 page 1251

82 Full paper no: 156 New sustainable and insulating building material made of cattail Martin Krus 1 Werner Theuerkorn 2 Theo Großkinsky 1 Hartwig Künzel 1 1 Fraunhofer Institute for Building Physics IBP, Holzkirchen, Germany 2 typha technik, Naturbaustoffe, Schönau, Germany KEYWORDS: Insulation, sustainability, cattail, half timbered framework, heritage SUMMARY: Due to the special structural properties of cattail (typha) building materials can be produced offering a combination of insulation and strength, which is unique on the market. The leaf mass of typha is especially suited due the structure of the plant. The leaves have a fiber-reinforced supporting tissue filled with soft open-cell spongy tissue providing for amazing statics and an excellent insulating effect. In the past few years, the Fraunhofer Institute for Building Physics investigated various product developments in cooperation with the inventor Dipl.-Ing. Werner Theuerkorn. The newly developed magnesite-bound typha board has an extremely high strength and dynamic stability despite a low thermal conductivity of about W/mK and can solve energetic as well as static problems. This innovative building material possesses a lot of positive properties. With the typha board as infill of the timber frames and as an additional inside insulation layer an extremely slender exterior wall construction with wall heating is realized. Due to the simple processability and inherent stiffness the material could be adjusted to the irregular inclined walls. The suitability of the wall structure has been investigated over a measuring period of 1.5 years. The U-value of the whole building (infill and timber construction) is about 0.35 W/m²K. The low level of moisture applied by the mortar and plaster dried out fast to a constant moisture contents in the wooden supports of below 20 M.-%. 1. Introduction Cattails are, due to their enormous growth rate and yield, optimally suited as raw material for industrial use. Typha stock (Fig. 1, left) comprises resilient, natural monocultures with an annual production rate of 15 to 20 tonnes of dry matter per hectare. This corresponds to four to five times the amount that local evergreen forests produce. Cattail crops create ecologically precious wetlands, which fulfill other important functions besides the absorption of nutrients and CO 2 (Faulstich 2012). Cultivation in lowland moors and valley plains in Germany would offer a sufficient basis to cover the total demand for insulation and wall construction materials. The special structural characteristics of cattails support the production of construction materials that offer a unique combination of load bearing capacity and insulation. The plant s structure (Fig. 1, right) entails the particular suitability of the typha leaf mass for creating innovative building materials (Pfadenhauer 2001, Theuerkorn 1998). Due to the combination of tensile strength of stem fibre and elastic sponge-like tissue, leaves are tear and break resistant, flexible and maintain their shape even in dried condition. These characteristics provide remarkable load-bearing capacity and excellent insulation properties. Behaviour of leaf mass under tensile and compressive stress is completely different along the leaf axis from base to tip than perpendicular to it: along the axis, the leaf material resists high compression loads of approximately 1 N/ mm and even higher tensile stress. Perpendicular to this axis, elastic deformation sets in already at very low stress of 0.01 N/mm and predominantly remains within reversible ranges. Full papers - NSB 2014 page 1252

83 Fig. 1. Typha sprout with leaf fan, left picture by TU Munich and section of a typha leaf, right picture by Chr. Gruber BLfD The special qualities of typha insulation panels originate in these diverse characteristics. Their production is based on laying out typha-leaf particles randomly, yet parallel to the panel plane and binding them with magnesite. The result is a material that can be created within a relatively simple procedure. This product contains only plant ingredients, purely mineral-based adhesive and no further additives. Thus, it is completely compostable. At the same time, it features a beneficial ratio of compressive strength along the panel plane, thermal conductivity, vapour diffusion properties, as well as storage mass for summertime heat protection. Variations in strength values and insulation capacities depend on bulk density and the percentage of magnesite (Table 1). Table 1:Measured heat conductivities and bearing loads for different bulk densities and parts of mangesite bond, from Theuerkorn (2013). Type Density [kg/m³] Magnesite part [%] Bearing load [N/mm²] Heat conductivity* [W/Km] 1a ,54 0,055 1b ,46 0,058 1c ,34 0,053 2a ,36-2b ,36-2c ,29 0,048 3a ,01-3b ,76 0,061 * orientative, not normative measurement The complete range of hygrothermal key values was registered for a material sample with a particularly effective combination of stability and thermal conductivity. The material, despite relatively high bulk mass and high solidity, features a comparably low thermal conductivity of W/mK and displays capillary action at a medium vapour diffusion rate (Table 2). By using it, vapour barriers could be avoided completely in many applications. Moreover, this innovative building material possesses a lot of other positive properties: renewable building materials with a very high resistance to mould growth good protection against fire, noise control and thermal insulation in summer simple processability with all common tools relatively diffusion open and capillary active low energy consumption in production recyclability. Full papers - NSB 2014 page 1253

84 Table 2: Hygrothermal material properties of type 1a. Material property Unit Result Bulk density kg/m³ 270 Porosity Vol.-% 75 Diffusion resistance - 28 dry-cup (23 0/50) Wet-cup (23 50/93) - 20 Water absorption coefficient Kg/m² h 1.1 Sorption moisture content Vol.-% C 65 % r. H: 23 C 80 % r. H: Vol.-% C 93 % r. H: Vol.-% C 97 % r. H: Vol.-% 6.9 Capillary saturation Vol.-% 59 Heat conductivity W/mK Insulation of half timbered framework Due to the immanent energy transition, new requirements are also placed on historic buildings. In general, they can hardly be met. This is why it is necessary to develop new materials and concepts for energy optimization meeting historic preservation needs specifically for such objects. The newly developed building material made of cattails seems appropriate for the task. In the case of a timbered building in Nuremberg with asymmetrical design and inadequate bracing of the structural frame (see Fig. 2, left) the timbers were supposed to be made visible again - while maintaining EnEV 2009 regulations as well as historic preservation requirements (Theuerkorn 2013; Fritsch 2013), and at the same time, providing stability to the building. These requirements were met by employing typhapanels. The model project made use of the material and was supported by the Federal German Foundation for the Environment and the Bavarian State Office for the Preservation of Historical Monuments. After comprehensive research on the existing construction, the details for the wall composition were designed and realized in close cooperation with construction management and craftsmen. A model frame served to develop a slender exterior wall construction of only 20 cm depth featuring integrated wall heater (Fig. 2, right). Fig. 2. Condition of the wooden framework before restoration (left) and sheme of the insulation (right), pictures by Alexandra Fritsch. The framework timbers were covered with slats according to carpentry standards for renovating such structures. The typha panels were cut to provide a continuous 10 mm wide groove. First, external panels with a thickness of 60 mm were fixed on the outside to the slats with drywall screws and Full papers - NSB 2014 page 1254

85 washers. A second panel with 60 mm thickness was fit internally and connected to the external panels with screws, while leaving a gap between the edge of the interior panels and the timbers. To enable wind proofing and force-fit connections, the gaps between timbers and typha panels are infilled with a typha-based joint compound. The seams are covered and smoothed with a taping knife. By including ground typha material, the joint compound can expand if water is introduced later on. After creating a uniform plane wall surface by use of compensating panels, additional 40 mm thick typha panels were attached to the interior wall. Due to panels being screw-tight and easy to render, their surface served to directly mount wall heat pipes with screw connectors. Voids were infilled with a lime-gypsum based mortar. The finishing coat is a loam render enriched with cattail seed parachutes. This render reinforcement is an effective means to ensure crack resistance without a fabric lining. Fig. 3 shows the layout of the construction. Fig. 3. Layout wall structure planning guide, picture by Fritsch+Knodt&Klug. The functional capability of the wall construction was tested by monitoring during a one and a half years testing period. For this purpose, sensors were distributed along the cross section of a selected infill area to determine temperature, relative humidity, wood moisture, and heat flow (Fig.4). Fig. 4. Outer sensors at layer 1 beneath the outside rendering (left) as well as Temperature sensor, heat flow wafer and humidity sensor on barrier layer 3, beneath internal insulation on the second insulation board layer (right). The course of boundary layer temperatures as hourly average values from January 2011 to September 2012 is displayed in Fig 5, left. The typical thermal stratification from the interior (room air temperature RLLT / interior surface temperature IOFT) to the exterior along boundary layer 4 (GS4T) is evident. Prior to operating the wall heater, the space was heated via an open door to the heated neighbouring room. The initial adjustment attempts of the renter after begin of operation clearly display excessive use (marked by a blue circle). Fig 5, right side, shows the measured heat flow behind the interior insulation (boundary layer 3) and the temperature distribution along this area, as well as the exterior along boundary layer 1. The thermal insulation characteristics of the wall construction can be calculated based on this heat flow in relation to the temperature difference. Two selected measuring periods (indicated in orange) Full papers - NSB 2014 page 1255

86 result in a heat transfer coefficient (U-value) of 0.26 W/m²K, taking the additional interior insulation into account. For the entire construction including frame timbers the result is a value of 0.31 W/m²K. Since measurements also include thermal gains due to solar intake (including diffuse irradiance), this heat transfer resistance is called relational U-value. Fig. 5. Measured course of the temperature as daily mean values for the period from January 2011 to September 2012 (left) and heat flow measured behind the internal insulation (layer 3) with course of the temperature at layer 1 and 3. The bars show time periods suited for assessing the thermal resistance (right). When calculating the real U-value based on the material properties, the outcomes are slightly higher values of 0.29 W/m²K for the infill and 0.35 W/m²K for the entire construction. Due to the building moisture that is introduced via the exterior render and the relatively good absorption capacity of wood, the results show a very high initial wood moisture content of more than 100 %. However, drying occurs quickly, and the wood moisture content of all four measurement areas along the wood surface located immediately behind the exterior render layer decreased to 20 % (Fig. 6). Fig. 6. Course of the wood moisture (left) and view of the building after restoration (right). By adding the typha panels as a combination of infill insulation and interior insulation, a heat transfer coefficient of approximately 0.35 W/ m²k was achieved for this wall construction - at an overall wall thickness not exceeding 20 cm, including wall heater. For a timbered building, this is an extraordinarily good result. The measurements of temperatures and air humidity undertaken across two heating periods prove the suitability of the construction in regard to building physics. Altogether, the entire insulation procedure based on applying magnesite bonded typha panels comprises an extremely effective solution in terms of building physics and historic preservation. Full papers - NSB 2014 page 1256

87 3. Calculational investigations for the magnesite-bound typha board as internal insulation on masonry The half-timbered building in Pfeiffergasse in Nuremberg has a ground floor of massive masonry as many other buildings of the same kind. For the energetic restoration of this building it was planned to install an internal insulation by magnesite-bound typha boards of a material thickness of 4 cm furnished with fiber-reinforced clay plaster. The boards were installed edge to edge and fixed by dowels on the internal surface of the external wall. Since an accompanying measurement was not planned due to financial reasons hygrothermal calculations were carried out to assess whether this measure will be free of damage. The IBP developed a proved and frequently validated onedimensional and two-dimensional computer program WUFI -Pro (Künzel 1994) for the calculation of coupled heat and moisture transfer processes. Previous descriptions of the moisture transfer behavior of building materials by means of this method have achieved good compliance of calculation and practical investigations of the test specimen (Krus 1996; Künzel 1999). Climate data of Holzkirchen are used as climate boundary conditions allowing assessments on more unfavorable weather conditions than those in Nuremberg. Living conditions with normal moisture load (meaning normal use of living areas) serve as indoor climate. The heat transmission coefficients are 8 W/m²K on the inside and 17 W/m²K on the outside. Material parameters of the masonry and clay plaster are taken from the WUFI material database. Calculations of the typha board are based on the hygrothermal material parameters determined before. If necessary, it is possible to assess by means of the prognosis tool WUFI -Bio (Sedlbauer 2001; Sedlbauer 2003), whether mould growth may occur. Since an all-over contact of the typha board with the brick work cannot be secured a thin air layer was suggested between insulation and masonry. If the most critical point of the wall between external wall and internal insulation is considered, the characteristic seasonal fluctuations of temperature and humidity can be observed. Maximum relative humidity of approx. 65 %, however, is achieved in this point (Fig. 7, left). Therefore, mould growth can be excluded. The calculation results show that no moisture damage will occur in case of careful implementation with good permanent convection tightness. 30 Monitorpos Monitorpos. 2 Temperature [ C] Temperature [ C] Monitorpos Monitorpos. 2 Rel. Humidity %] Rel. Humidity %] Time Zeit [Jahre] [a] Time Zeit [Jahre] [a] Fig. 7. Course of temperature (top) and relative humidity (bottom) between external wall and internal insulation with careful implementation (left) and without perfect tightness with back flow of 1 l/mh (right). Full papers - NSB 2014 page 1257

88 Since it cannot be suggested that this is always the case, the tolerance of the construction must be investigated. To find an answer to the question the following investigations were based on a defined leakage with a back flow of the insulation board by warm and humid air from the interior. The following calculations were based on a back flow of 1 liter per hour and running meter wall length. Fig. 7, right side, shows the situation behind the insulation. Due to the inflowing warm air minimum temperatures of barely below 0 C are raised to scarcely above freezing temperature. The rel. humidity doesn`t exceed 80 % despite the back-flow. Fig. 8, left side, shows the result of a further increase of the back flow up to one cubic meter per day and running meter wall length. In winter, however, 80 % r. h. is temporarily exceeded so that mold growth could no longer be excluded. For verification the calculated course of temperature and relative humidity at this point is used for the mould growth prognosis program WUFI -Bio. As the results in Fig. 8, right side, show the spore water content never exceeds the limit water content. Despite back flow no mould growth must be expected under these conditions. Therefore, the wall structure with this kind of internal insulation shows a considerable tolerance of untightness, one reason may be the relatively low thickness of the insulation material. 30 Monitor po s. 2 Temperature [ C] Monitor po s. 2 Rel. Humidity %] Time [a] Zeit [Jahre] Fig. 8. Course of temperature (top) and relative humidity (bottom) between external wall and internal insulation at a back flow of 1 m³/md (left) and calculated results by WUFI -Bio for the area behind the internal insulation. Full papers - NSB 2014 page 1258

89 4. Conclusions A product has been developed by means of optimizations concerning the structure of the board and the material properties showing numerous positive characteristics. For the first time ever, a material is available showing a relatively high bearing capacity and simultaneously good insulation properties, which is manufactured from renewable material at low energy consumption and entails great advantages for the environment. Moreover, susceptibility to mould growth in practical application is relatively low, what is frequently a problem in case of renewable insulation materials. The material is sufficiently diffusion-open to support drying-out processes but diffusion-tight enough to work without any vapor barrier in many applications. Calculations of the internal insulation showed that at least in case of moderate thickness the typha boards can be directly doweled to the wall and coated by an internal plaster, and thus can function free of damage from the building physical point of view and have a considerable tolerance for errors in installation. The installation of the typha boards in the framework building in Nuremberg allowed the achievement of a thermal transmittance coefficient of approx W/m²K with a total thickness of the wall structure of only 20 cm including wall panel heating by a combination infilling with the typha board and internal wall surface area insulation. This result is extremely good for a framework building. The measurements of temperatures and humidity in various depths of the structure carried out at the object during two heating periods prove the building physical suitability. In the beginning, slightly higher initial humidity occurred due to the built-in moisture added by internal and external plasters, and then the framework dried out rapidly and humidity remained low and uncritical. The measurements of the heat flow confirmed the positive results of the calculations. The measurements of the moisture contents of the timber also showed that due to the rapid drying-out only relatively low and uncritical moisture occurred in the timber. All in all, an extremely positive result is achieved from the building physical point of view and as concerns the preservation of historical monuments by the installation of magnesite-bound typha boards as insulation measure. 5. Acknowledgements Thanks to Altstadtfreunde Nürnberg e.v. and to Deutsche Bundesstiftung Umwelt who have not only financially supported these investigations. References Faulstich, M SRU-Umweltgutachten 2012, Verantwortung in einer begrenzten Welt, Berlin. Fritsch, A. & Theuerkorn, W Fachwerksanierung und Energieeffizienz. In Denkmalpflege Informationen. Krus, M. & Künzel, H.M Vergleich experimenteller und rechnerischer Ergebnisse anhand des Austrocknungsverhaltens von Ziegelwänden. Internationales Symposium of CIB W67 Energy and Mass Flow in the Life Cycle of Buildings. Wien, August 1996, S Künzel, H.M Verfahren zur ein- und zweidimensionalen Berechnung des gekoppelten Wärmeund Feuchtetransports in Bauteilen mit einfachen Kennwerten. Dissertation Universität Stuttgart Künzel, H.M. (1999): Praktische Beurteilung des Feuchteverhaltens von Bauteilen durch moderne Rechenverfahren. WTA-Schriftenreihe, Heft 18, Aedificatio Verlag. Full papers - NSB 2014 page 1259

90 Pfadenhauer, J.& Heinz, S Multitalent Rohrkolben -Ökologie, Forschung, Verwertung Broschüre zum Abschlussbericht des DBU-Projektes Rohrkolbenanbau in Nierdermooren - Integration von Rohstoffgewinnung, Wasserreinigung zu einem nachthaltigen Nutzungskonzept im Donaumoos , TU München, Lehrstuhl für Vegetationsökologie. Sedlbauer, K Vorhersage von Schimmelpilzbildung auf und in Bauteilen. Dissertation Universität Stuttgart (2001). Sedlbauer, K. & Krus, M Schimmelpilze in Gebäuden Biohygrothermische Berechnungen und Gegenmaßnahmen. Berlin : Ernst und Sohn Verlag S , Bauphysik-Kalender Theuerkorn, W., Reizky, Lenz, Kleyn Rohrkolben, ein nachwachsender Rohstoff. Theuerkorn, W.; Fritsch, A.; Mach, M.; Krus, M; Großkinsky, Th.; Fitz,C. Theuerkorn, D. Knodt, H. Walter, U Neuer Baustoff für umweltfreundliche und bautechnische Sanierung in der Denkmalpflege. DBU-Bericht (Förderkennzeichen AZ 27918). Full papers - NSB 2014 page 1260

91 Full paper no: 157 Cross Laminated Timber vs. timber frame walls in water damage comparing drying and mould growth Kristine Nore, Ph.D. 1 Johan Mattsson, Cand.Scient 2 Mari Sand Austigard, Ph.D. 2 1 Norwegian Institute of Wood Technology, Norway 2 Mycoteam AS, Norway KEYWORDS: Water damage, CLT, Cross Laminated Timber, wood, mould growth, drying. SUMMARY: In Norway water damages were reported to insurance companies in Risk management and renovation procedures are needed in order to reduce the costs generated by these incidents. In the presented experiment one square meter dwelling partition walls were submerged for 48 hours prior to full assembly and drying in a dry environment. Dwelling partition walls are between two separate residential apartments, and have to fulfil requirements regarding fire safety, sound insulation, moisture control and more. The edges of the wall samples were sealed in order to obtain drying mainly through the wall surfaces. The results show different drying patterns in standard light weight timber frame walls and Cross Laminated Timber (CLT) walls. The CLT walls were not fully wetted after 48 hours, and could still absorb water from the soaked mineral wool between the CLT elements. The CLT partition walls therefore gave lower risk of mould growth, and lower actual mould growth. However, the drying period was longer compared to the timber frame wall. 1. Introduction Water damages are common in Norway, reaching water damages reported to insurance companies in 2013 (Finance Norway, 2014). There is always a risk of mould growth following water damage. Mould is unacceptable in dwellings, and avoiding mould damage is a major aspect in handling water damages. The crucial factor in order to avoid mould growth is to obtain quick drying of the wetted materials. It has been shown that log walls and massive timber walls are damaged by mould fungi after 3-4 weeks, while light timber frame walls are damaged after about one week of wet conditions due to more accessible cellulose in the gypsum boards (Mattsson & Stensrød 2009). Mould growth depends on available water at the surface of the substrate (Rayner and Boddy, 1988, Samson et al. 2004). It is therefore of great importance to establish dry surfaces as fast as possible, at least within one week, in order to avoid mould damages after water damages. The main question is how to handle water damages in such a way that mould growth can be avoided in every single case, regardless of construction type. Despite the large number of water damages every year, there is a general lack of understanding regarding how to perform a satisfactory handling of water damages. Mattsson & Stensrød (2009) reported that about 50% of investigated water damages where inadequately managed by the contracted professional companies. Given the general lack of knowledge about how to handle well-known structures, there is great uncertainty regarding how to handle water damages in new and more unknown materials and structures, e.g. in CLT assemblies. It is therefore important to clarify how wet these constructions might get during leakage incidents, how fast they dry out, and how extensively they may be attacked by mould. Full papers - NSB 2014 page 1261

92 CLT elements are still emerging as a construction concept, and experience is limited. The first erected CLT buildings are now ten-fifteen years old. Risk handling in the event of water damage is not developed for CLT buildings, neither in terms of survey methods, adequate instruments for moisture measurements or suitable drying equipment. Knowing the hygrothermal performance after water damage will allow better procedures for risk handling actions. Hygrothermal performance for external walls, including common CLT build ups, is tested and calculated in McClung (2013). She concludes that built in moisture alone is not likely to cause building failure. One limit found is the ability to calculate actual performance for different climates. Available simulation tools are not sufficiently accurate for this purpose, and experimental data are needed. Two experiments were performed: One large scale drying test on square meter dwelling partition wall elements and one smaller scale test of mould growth on smaller wall elements. Dwelling partition walls used in multistory wooden buildings often consist of separate load bearing walls, most often with additional insulation in between for sound insulation. 2. Measurement set-up The CLT provided was 120 mm thick. This is usually more than needed for loadbearing inner walls, but was considered an advantage concerning moisture uptake and drying out period. The light weight timber frame was built using standard mm wood, assembled as shown in BDS (2002). In each test element two wall elements were combined, with insulation in between. All elements used in both tests were insulated with mineral wool, which mainly provides sound insulation in such partition walls. The drying test was performed using large-scale elements measuring B H = m. Three CLT partition wall elements and one timber framed reference element were used. The mould growth test was performed on smaller elements measuring 0.6 x 0.37 m. Four CLT elements and four timber frame elements were used. 2.1 Large scale laboratory drying test The large scale buildup is shown in Figure 2. The elements were submerged for 48 hours. After wetting the pieces were assembled to dwelling partition walls. In order to ensure the correct distance between the separate wall elements in each test element, chipboards were attached to all four edge surfaces. PE foil was inserted between the chipboards and the elements in order to seal the edges. For additional sealing, an acrylic sealing compound was applied along the edges of the chipboards. FIG 1. A thermal photograph of two of the CLT elements after wetting. Colder areas (blue) have higher water content in the wood than warmer (yellow) areas. The elements have mainly absorbed water along their outer edges, leaving the middle of each element relatively dry. Full papers - NSB 2014 page 1262

93 a FIG 2. Construction principle for the large-scale drying test. a: Two wall elements combined to form a dwelling partition wall, with mineral wool in between for sound insulation. b: the elements suspended in the weighing rig. The reference timber frame test element is seen in front. The elements were instrumented with weight cells and Hygrotrac sensors (Omnisense, 2014). Each element was suspended from a weight cell, so that their weight could be monitored continuously. Unfortunately the weight cells did not operate prior to January 8 th. The Hygrotrac sensors measure air temperature, relative humidity and wood moisture content. In each CLT element three Hygrotrac sensors were mounted, on the inside (wet part of partition wall), in the middle of the element (holes drilled from outside) and on the outside (also measuring laboratory climate), as shown in Figure 2a. After 80 days in the drying rig, air sampling for total count of mould spores was performed in all four test elements. The method used was the Air-O-Cell air sampling cassette in which 75 litres of air was sucked from inside each element using a calibrated wall pump (BioAire), see Figure 3. The air is passed through a sticky surface where the mould spores and other airborne particles are captured, and the filter is subsequently analysed under a microscope. b FIG 3. Sampling of air from inside a CLT element, via a hole drilled in the particleboard on the edge surface of the element. Full papers - NSB 2014 page 1263

94 2.2 Mycotest The elements used in the mould growth test were contaminated by soaking the elements in a mould spore suspension. The suspension was a mixture of commonly occurring mould fungi, e.g. Cladosporium spp., Penicillium spp., Aspergillus versicolor and Chaetomium globosum of about spores/ml water. After contamination, Hygrotrac instruments were installed and the elements were assembled and the edges sealed using PE-foil and moisture tight sealing tape (Figure 4a). The elements were placed in climate chambers with 20 C and 70% relative humidity. Temperature, relative humidity and wood moisture in each test element was measured every 30 minutes. Each week one CLT and one timber frame element were dissembled for assessment of mould growth (Figure 2b). The amount of mould growth was assessed using microscopic analysis of tape lifts. a b FIG 4a. Timber frame test element for mould growth after contamination and assembly. b. Timber frame test element after dissemble for mould growth assessment. 3. Results 3.1 Drying test The climate in the room where the drying test was performed is not controlled. The temperature and relative humidity varied throughout the test as shown in Figure 5. The temperature was [19-24] C and the relative humidity [25-60] %. The standard deviation is included in the graph to show the variation of the six sensors. The test elements were placed only 10 cm apart in the drying rig. This caused higher air humidity in the in-between spaces, especially in the beginning of the test. This is seen in the standard deviation of the relative humidity in Figure 5. The drying process, which is most rapid in the beginning of the test, is responsible for the lower air temperature shown in the start of the drying. This is due to the fact that the transfer of bound water moleculesto free water molecules in the air reduces the moisture enthalpy, which again requires energy. This energy is taken from heat in the air and in the wood, reducing temperature of both in the process. The temperature and relative humidity inside the partition walls are shown in Figure 6. The temperature follows the room temperature, only with smaller amplitudes. A time lag is clearly evident in the CLT elements, as well as even smaller amplitudes than in the timber frame element. This is due to the larger hygrothermal capacity of the CLT elements. The fact that the relative humidity reaches values above 100 % is obviously a measurement failure. Anyhow, the moisture inside the timber frame element levels out at a level that is probably close to saturation. The relative humidity in the timber frame element probably falls to a level slightly below saturation on February 6 th, but no further drying seems to take place during the test period. The air inside the CLT elements increases in humidity during the first few days, and then dries steadily throughout the test period. The CLT elements clearly had more moisture capacity to remove the moisture from the locked void of the walls. Full papers - NSB 2014 page 1264

95 The moisture content of the wood in the CLT elements is shown in Figure 7. The wood moisture content was decreasing from the inside towards the outside of the element throughout the test period. The moisture content at the outer surface very quickly decreased below 10 weight percent. In the middle of the element the moisture remained relatively stable at 15 weight percent, while the inner surface dried from around fiber saturation point to 20 weight percent during the 12 weeks of drying. Figure 8 shows the development in weight for all elements. The CLT weighed around 130 kg, while the timber frame element weighed around 70 kg. The timber frame element lost 5 litres from Jan 8 th to Feb 24 th. The CLT elements lost inaverage 1.5 litres in the same period. FIG 5. Climate around the test walls given by six temperature and relative humidity sensors and presented as average values with standard deviations. FIG 6. Temperature and relative humidity inside the test elements. The values for the CLT elements are average data from the three elements. The temperature delay for the CLT walls is due to their larger hygrothermal inertia. Full papers - NSB 2014 page 1265

96 FIG 7a: Wood moisture content in weight percent on the inner surface, in the middle of the CLT elements and on the outer surface. The moisture gradient shows that the elements are drying from the inside towards the outside. b. Development of the weight in all four elements in the drying test. The timber frame wall element had a lower initial weight than the CLT elements, and lost weight at a higher rate. 3.2 Mould growth test Logging of relative humidity in the test elements showed clearly lower relative humidity inside the CLT elements compared to the timber framed elements. The relative humidity in the timber frame elements was around 100% throughout the test period, while the relative humidity in the CLT elements was 70-80%. The analyses of the test elements dissembled after 1, 2, 3 and 4 weeks of exposure show large differences in mould growth between the timber framed and the CLT elements (Table 1). The growth of mould fungi was well established in the timber framed elements during the first week of the mould growth test and increased to extensive growth during the second week. The mould growth was mainly found on the gypsum boards, and on the wood surfaces directly attached to the gypsum boards. In the CLT elements mould growth was slow to establish and only reached sparse growth in the fourth and final week of the test period. Results from air sampling in the test elements in the drying test show after 80 days a large amount of mould spores was found in the timber frame element, mould spores/m 3. In the CLT elements the number of mould spores was similar to that in the ambient air, counting 108 mould spores/m 3 air. TABLE 1. Mould growth in the test elements after one, two, three and four weeks. Element Mould growth One week Timber framed element Moderate growth CLT element No growth Two weeks Timber framed element Extensive growth CLT element No growth Three weeks Timber framed element Extensive growth CLT element No growth Four weeks Timber framed element Extensive growth CLT element Sparse growth Full papers - NSB 2014 page 1266

97 4. Discussion Water damages occur in all type of buildings and constructions. The main questions are how to handle such damages and which risk there is for mould growth in various situations. Mattsson & Stensrød (2009) reported that about 50% of controlled water damages were inadequately managed. Greater knowledge of the development of mould growth in different damage situations should make proper damage management easier to obtain. The results display the influence of hygrothermal inertia. Even after 48 hours of submersion the water did not reach more than centimeters into the CLT as shown in Figure 1. This proves slow water uptake, and the corresponding slow drying is also shown in Figure 6 and 7 a and b. This corresponds to former research like given in Hameury (2006). Regarding water leakages this effect is beneficial. The water is slowly absorbed and dried, not providing free surface moisture. This gives time for renovation. However, it is of great importance never to hinder diffusion drying in wet biological material, including CLT elements. Thus, a risk management procedure in case of water leakage in such structures should be developed. Our results in this project show that CLT wall constructions dry out more slowly than timber frame walls. Nevertheless, the results regarding mould growth after deliberate contamination show that the risk for mould damages is significantly lower in the CLT construction compared to the more commonly used timber frame wall structure which had established mould damage within one week as expected. There was no active contamination of the large elements that were soaked for monitoring of the drying process. The air sampling for mould spores showed that a natural contamination occurred in the test elements, even though the test was started in December when there are very few mould spores in the air (Mattsson et al 2014) and only new materials without any previous moisture problems or mould growth were used. This shows that that the mould spores available in the ambient air and tap water are sufficient to cause extensive mould damage, even in winter when spore counts are low. This is also what is found in normal water damages; given a little time, mould growth almost always occurs. The evident occurrence of mould damages in the timber frame wall construction and evident absence of similar damages in the CLT elements also confirms the same pattern as the controlled test with active contamination of mould fungi. The positive result from the CLT elements regarding mould growth despite a very long exposure time can be explained by the large amount of wood in the CLT. Wood is a highly hygroscopic material with a vast number of available sorption sites, efficiently transporting water molecules from the air into the wooden structure. Thus the water activity at the surface is kept at a low level, and mould spores are unable to germinate (Samson et al 2004). The extensive mould growth in the timber frame elements is probably caused by a combination of two factors: lasting high relative humidity and the presence of gypsum boards. Gypsum boards are extremely susceptible to mould growth, due to a combination of accessible cellulose in the paper and good water retaining properties of the gypsum (Mattsson & Stensrød 2009). 4.1 CLT elements as construction material We have only tested standard CLT elements. In this type of element water slowly if not at all penetrate between the lamellae in the element, and water uptake takes place mainly at the outer surfaces of the element. Timber elements fastened by wooden plugs, screws or other methods might allow water to penetrate into the element, and water uptake can take place in each lamella. This could give more hospitable conditions for mould fungi. Different insulation materials can be expected to affect the relative humidity inside the partition walls. Further studies should be performed using different massive timber elements and different insulation materials as well as different modes of wetting, Full papers - NSB 2014 page 1267

98 simulating leakages, wetting in the construction process and/or other common sources of water damage. 5. Concluding remarks The testing of CLT elements compared with standard timber framed wall shows promising results regarding the risk of mould growth after water leakages. Despite a clearly slower drying process in the CLT elements, there was significantly less mould growth. This shows that CLT elements do not increase the risk of mould damages after possible water damages, and even indicates that the risk may be smaller where CLT elements are used. This is important, as mould growth has a negative impact on indoor air quality. However, there is still need to develop proper risk management procedures in case of water damage in CLT constructions. The authors foresee a development of pre-applied drying out slots available for mounting drying equipment in case of water leakage. 6. Acknowledgements The authors highly acknowledge Innovation Norway who party funded this project and Massiv Lust who provided the CLT. References BDS Partition walls between row house dwellings. Building Design Sheet. SINTEF Building and infrastructure, Trondheim, Norway (in Norwegian). Finance Norway (FNO). Water damages statistics Cited January 20 th Hameury S., The hygrothermal Inertia of Massive Timber Constructions. Doctoral Thesis. Royal Institute of Technology, Architecture and the Built Enviroment. Stockholm, Sweden. McClung, V. R. (2013). Field study of Hygrothermal Performance of Cross-Laminated Timber Wall Assemblies with Built-In Moisture. Thesis and dissertation Paper Ryerson University, USA. Mattsson J, Muggsopp i bygninger. Forekomst, påvisning, vurdering og utbedring. Mycoteam, Oslo (in Norwegian). Mattsson J, Grønli I, Whist CM, Ødegaard AT, 2014.Muggsoppsporer i luftprøver. Agarica 2014, vol. 34, (in Norwegian). Mattsson J, Magnusen K, NBI-blad Muggsopp i bygninger. SINTEF Byggforsk, Oslo (in Norwegian). Mattsson J, Stensrød O, Håndbok om vannskader. Årsak, undersøkelser, tiltak og gjenoppbygging. Mycoteam, Oslo (in Norwegian). Nunez M, Sivertsen MS, Mattsson J, Substrate and construction preferences for Actinomycetes and 20 mould genera. Proceedings in Healthy Buildings 2012, Brisbane. International society of indoor air. Rayner ADM, Boddy L, Fungal decomposition of wood: Its biology and ecology. John Wiley & sons, Chichester. Samson RA, Hoekstra E, Frisvad J, Filtenborg O, Introduction to food- and airborne fungi. Centraalbuerau voor schimmelcultures, Utrecht. Full papers - NSB 2014 page 1268

99 Full paper no: 158 Retrofitting a brick wall using vacuum insulation panels: measured hygrothermal effect on the existing structure Pär Johansson, Lic.Tech. 1 Stig Geving, Professor 2 Carl-Eric Hagentoft, Professor 1 Bjørn Petter Jelle, Professor 2, 3 Egil Rognvik 3 Angela Sasic Kalagasidis, Assistant Professor 1 Berit Time, Ph.D. 3 1 Chalmers University of Technology, Sweden 2 Norwegian University of Science and Technology (NTNU), Norway 3 SINTEF Building and Infrastructure, Norway KEYWORDS: listed building, brick wall, interior insulation, vacuum insulation panel, measurement, laboratory, driving rain SUMMARY: Old listed buildings need to be retrofitted to reduce the energy use for heating. Vacuum insulation panels (VIPs) require less thickness than conventional insulation materials to reach the same thermal resistance. The aim of this paper is to investigate the hygrothermal performance of a brick wall with wooden beam ends after it was insulated on the interior with VIPs. The paper presents the first part of a laboratory study where a brick wall was built in the laboratory and exposed to simulated driving rain. Different measurement techniques of the relative humidity in the construction have been used. The relative humidity in the wall increased substantially when exposed to driving rain. The moisture content in the wooden beams also increased. However, it has not been possible to fully determine the influence by the added insulation layer. It is clear that the drying capacity to the interior side is substantially reduced. These investigations are ongoing and will be reported in future publications. 1. Introduction In Europe, the majority of the future building stock has already been built. The increasing energy prices and the pressure to reduce the energy use in society urge for energy retrofitting measures in the existing building stock (IEA 2013). External walls of old buildings in Sweden and Norway often have a low thermal resistance in comparison to current standards. In Swedish buildings built before 1960, the average U-value of the walls is 0.58 W/(m 2 K) (Boverket 2009) while it is 0.9 W/(m 2 K) for at least Norwegian buildings from before 1945 (Thyholt et al. 2009). For retrofitted walls the general target U-value is 0.18 W/(m 2 K) in Sweden (Boverket 2011) and 0.22 W/(m 2 K) in Norway (KRD 2010). Many old buildings are considered to be of great historical value and are protected for their external appearance which limits the possible retrofitting measures. Retrofitting on the exterior side of the wall is, in many cases, not allowed so the only possible solution is to add interior insulation. The available additional thickness of the wall is limited by the allowed reduction in rentable internal floor area. Novel highly efficient thermal insulation materials such as vacuum insulation panels (VIPs) increases the thermal resistance of the wall compared to conventional insulation materials with the same thickness. The thermal resistance of a VIP is 5-10 times higher than for conventional insulation materials (Baetens et al. 2010), reducing the required thickness to reach a targeted thermal resistance. Full papers - NSB 2014 page 1269

100 VIPs are rigid panels which, unlike most insulation materials, cannot be adapted on the construction site and have to be preordered in the correct dimensions. They are sensitive to damages which could lead to puncturing and a fivefold increase in thermal conductivity. Therefore special care has to be taken in all stages of the construction process to avoid damaged VIPs. Also, thorough hygrothermal investigations are needed to ensure that the relative humidity in the wall is below the critical levels for mold growth and dry rot fungi in wood and freeze thaw damages in brick and mortar. When retrofitting old buildings, the prerequisites are given by the existing construction. The intermediate floors in old brick buildings are often carried by wooden beams which are embedded in the brick. Mold and dry rot can damage the wooden beams and the risk for that is higher when interior insulation is added because of the higher relative humidity in the wall. Driving rain raises the moisture content in the wall and wooden beam ends, increasing the risk for damages. Also air leakages from the interior into the area around the wooden beam ends can transport moist air from the interior which will raise the moisture content even higher (Kehl et al. 2013). Unprotected brick walls may also have freeze-thaw damages. The movement of water through brick masonry has many important consequences in building constructions and it has therefore been studied by a number of authors, e.g. Hall (1977) and Brocken (1998). While the majority of these studies involved water suction experiments from a free water surface, large scale experiments where water suction in brick walls is studied during a real or artificial rain load, such as presented by Abuku et al. (2009) and Piaia et al. (2013), are rare. To the knowledge of the authors, similar studies for brick masonry are not available. The aim of this paper is to investigate the hygrothermal performance of a brick wall with wooden beam ends after the wall was insulated on the interior side with VIPs. The brick wall was built in laboratory according to the methods used in the late 19 th century to the early 20 th century in Sweden and Norway. Wooden beam ends were studied since these are a known risk area when insulating brick walls (Kehl et al. 2013; Rasmussen 2010; Ueno 2012). The wall was tested in a large-scale building envelope climate simulator where it was exposed to a temperature gradient and cycling climate with driving rain. In the first sequence, the wall without interior insulation was tested. Before the second sequence VIPs were added to the interior of the wall. A pre-study using the hygrothermal simulation tool WUFI 2D was presented by Johansson et al. (2013) where suitable materials, wall layout and testing climate were proposed. The study is part of a research project which is run in cooperation between Chalmers University of Technology in Gothenburg, Sweden, the Norwegian University of Science and Technology (NTNU) and SINTEF Building and Infrastructure, in Trondheim, Norway. 2. Wall layout and material selection A common wall thickness in brick buildings from the late 19 th century is 380 mm which is equal to 1.5 bricks thick walls. The brick walls of multiple floors often have wooden beams inserted around 200 mm into the brickwork to carry the intermediate floors (Kvande & Edvardsen 2013). In the prestudy, Johansson et al. (2013) studied three wall thicknesses, 120 mm, 250 mm and 380 mm, to investigate the possibility to decrease the (expensive) testing time by using a thinner wall in the laboratory study. It was found that the moisture accumulation rate was not decreasing linearly with increasing thickness, but had a more exponential relationship. However, the same water flow was found during wetting for the different wall thicknesses. Therefore, the same conclusions could be drawn by using a 250 mm thick brick wall as for using a 380 mm thickness. The schematics of the brick wall built in the laboratory and investigated in this paper are presented in FIG 1. Two VIP sizes, 20 mm thick, were used in the study, larger 600x1 000 mm and smaller 500x600 mm, as shown in FIG 1. Three types of sensors were installed in the wall to monitor the wetting and drying; 10 relative humidity sensors (E+E Elektronik EE060), 8 Sahlén sensors (wood moisture sensors) and 12 resistance moisture meters (pin-type). The relative humidity sensors measure the relative humidity and temperature in the range of 0-100% and C. They were located in the mortar between the bricks, see FIG 2b, together with the Sahlén sensors. These measure the mass percentage moisture in Full papers - NSB 2014 page 1270

101 birch wood inside the sensor which gives a measurement range of % relative humidity. The size of the relative humidity sensor is 116x12 mm (length, diameter) and the Sahlén sensor is 40x13 mm. The resistance moisture meters were made of two insulated metal pins located 25 mm apart, installed on three different locations in the wooden beams as shown in FIG 1. One sensor was drilled into the center of the beam at the interior surface of the wall (a). Two other sensors were installed 10 mm from the end of the beam, one with insulated pins in the center of the beam (b) and one on the wooden surface (c). The relative humidity sensors were monitored hourly in the first sequence and every 6 minutes in the second sequence by a computerized system. The Sahlén sensors and resistance moisture meters were monitored daily at the start of the climate sequence, and later every 2 days. FIG 1. Left: schematics of the brick wall tested in the laboratory. The sensor locations at different depths of the wall are indicated by a, b and c. Right: measurements of the wall with the locations and sizes of the VIPs and sensor positions. RH = RH sensors, S = Sahlén sensors and W = resistance moisture meters. A layer of mineral wool was located around the two lower wooden beams. The horizontal dashed black line, indicated by an arrow, shows the symmetry line of the wall where a rubber strip was installed on the exterior side of the wall to break the water run-off. The brick wall was built inside a 3x3 m steel frame to allow it to be moved from the laboratory to the climate simulator. The lower part of the frame was filled with 200 mm cellular glass insulation to insulate the lower boundary from the steel frame. The size of the bricks was 226x104x60 mm (length x width x height) and the thickness of the mortar joints between the bricks were mm, see FIG 2a. The water accumulation in the upper part of the wall could interfere with the measurement results for the lower part of the wall. Therefore a rubber strip was installed on the exterior side of the wall to stop liquid water from being transported along the wall. Mortar was applied on the entire interior brick surface to make an even surface for attaching the VIPs. Four voids of each 100x225 mm, see FIG 2c, were created where the wooden beams were installed after the wall had dried. The gaps between the brick and wooden beams were sealed with a mix of modelling clay and beeswax. Around the two lower wooden beams, a layer of mineral wool was added to simulate the thermal resistance of the intermediate floor while the space around the two upper beams was left as it was, see FIG 1. A polyethylene foil was wrapped around the beams and over the space between them to simulate the vapor resistance of an intermediate floor, see FIG 2d. In the second sequence of the laboratory study, the interior side of the brick, between the intermediate floors, was covered by VIPs, see FIG 2e. To resemble the properties of an old brick wall in the laboratory study it was essential to use a brick and mortar similar to what was used in Sweden and Norway in the late 19 th century to the early 20 th century. The modern brick types are formed by dry-pressing, molding or extruding the clay to form the wanted size and shape (Brick Industry Association 2006), giving other properties to the bricks than what manual production methods does. The liquid water transport coefficient is lower for modern bricks than for the historical bricks which are targeted in this study. The bricks and mortar used in the laboratory study were chosen based on the conclusions from the pre-study by Johansson et al. (2013). Full papers - NSB 2014 page 1271

102 FIG 2. Photos from the construction of the wall in the laboratory. a: brick and mortar laid out, b: relative humidity sensor installed in the mortar in the middle of the wall, c: finished brick wall with the four voids for the wooden beams marked with black rectangles, d: wooden beams installed and wrapped in polyethylene foil, e: installation of the VIPs which were glued to the wall with taped edges. The mortar type used to bind the bricks together has varied over time. Hydraulic lime mortar was used in the early days but later replaced by cement mortar and finally by mixtures of lime and cement mortar (Kvande & Edvardsen 2013). The liquid water transport coefficient is substantially larger for lime and cement mortars and hydraulic lime mortars than for pure cement mortar. Hydraulic lime mortar requires a longer curing time than a mixture of lime and cement mortar would. It was also expected to be more difficult to fully control the adhesion of the hydraulic lime mortar. Therefore, a lime and cement mortar was chosen which resembles the hygrothermal properties of the historic hydraulic lime mortar and minimizes the curing and adhesion times, allowing for faster construction. 3. Measurements of hygric properties of the brick and mortar The hygric properties of the brick and mortar were tested in the laboratory of NTNU and SINTEF Building and Infrastructure in Trondheim, Norway. The procedures in the standards NS-EN ISO and NS-EN were followed using 6 samples each of the brick and mortar. The average values and standard deviations for the brick and mortar are presented in TABLE 1 TABLE 1. Measured hygric properties of the brick and mortar used in the laboratory study. The uncertainties are given as the standard deviation of the mean with a confidence interval of 68.3%. Material Density (kg/m 3 ) A w -value (kg/(m 2 s 0.5 )) Moisture content at 75% RH (wt-%) Moisture content at saturation (wt-%) Brick ± ± ± ± 0.6 Mortar ± ± ± ± 0.2 The A w -value is defined in NS-EN ISO as the short term liquid water absorption coefficient which is a measure of the rate of water absorption by e.g. driving rain on a material. To assess the liquid water transport coefficient, D ws (m 2 /s) dependent on the moisture content, w (kg/m 3 ), there is an approximate relationship between A w and D ws which is used in WUFI 2D (Fraunhofer IBP 2010): 2 w w f 1 ( ) Aw D ws w = (1) w f where A w short term liquid water absorption coefficient (kg/(m 2 s 0.5 )) w moisture content (kg/m 3 ) w f moisture content at saturation (kg/m 3 ) In order to use Equation (1), the moisture sorption isotherm for the material has to be defined. The approximate equation for the sorption isotherm based on measured data is (Fraunhofer IBP 2010): w ( ϕ) = w f ( b 1) ϕ b ϕ (2) Full papers - NSB 2014 page 1272

103 where w f moisture content at saturation (kg/m 3 ) b fitting parameter (-) φ relative humidity (-) The liquid water transport coefficient was calculated using Equation (1), based on the measurement results. The liquid water transport coefficients for the brick used in the laboratory study compared to the data from WUFI 2D (Fraunhofer IBP 2010) at 80% relative humidity (assuming D ws = 0 m 2 /s at 0% RH with a linear relation to 80% RH) and at saturation are presented in TABLE 2. TABLE 2. Liquid water transport coefficient at 80% relative humidity and at saturation for the brick used in this study compared to data from WUFI 2D (Fraunhofer IBP 2010). Material D ws at 80% RH 10-9 (m 2 /s) D ws at saturation 10-6 (m 2 /s) Measured brick Masonry Extruded Historical Hand-formed Vienna 1900s WUFI 2D The liquid water transport coefficient increases with a factor of when the moisture content in the brick used here becomes saturated. The relation is similar for the other brick types. Historical has the highest liquid water transport coefficient. The brick used in this study has a 100 times lower liquid water transport coefficient which is more in line with the properties of Masonry at saturation. The moisture diffusion resistance factor, µ (-), was not measured here but it is around for most bricks (Fraunhofer IBP 2010). The measured properties of the lime and cement mortar was similar to what is found in the WUFI 2D database for mortars of this type. The sorption isotherm was very similar to the sorption isotherm of the brick, but with a maximum moisture content of 195 kg/m 3 compared to 237 kg/m 3 in the brick. The liquid water transport coefficient was at 80% relative humidity and at saturation which is 60% and 18% of the brick. A parametric study of these different properties was performed by Johansson et al. (2013) which showed that the capillary active bricks gave a faster wetting of the wall. 4. Climate simulator and climate sequence To make the results of the laboratory study applicable to the conditions in Gothenburg (Sweden) and Bergen (Norway), the climate sequence was based on the climate in these cities. They are both located close by the sea which means a large portion of driving rain will hit the façades of the buildings. Measurements of the amount of rain during a rain event were used as input. The climate simulator is designed to generate a controlled dynamic climate condition on both sides of the brick wall. On the interior side a constant temperature of 25 C and relative humidity of 40% was chosen. The temperature and relative humidity was 5 C and 70% on the exterior side in the first sequence. In the second sequence it had to be changed to 10 C and 90% relative humidity because the equipment was overloaded and broke down during the first sequence. The rain period in the first sequence was supposed to be four hours long, but due to equipment malfunction the rain was not turned off in time but first after 14 hours. In the second sequence the rain period was reduced to 30 minutes since the rain intensity and the rate of the capillary suction were much higher than anticipated. The rain amount hitting the wall was 5 mm/hour, i.e. 5 l/(m 2 h). After the wetting sequence, the drying climate was 10 C and 60% on the exterior side and kept at 25 C and 40% on the interior side. After 1 month drying with VIPs on the interior side of the wall, the panels were removed to allow for drying out from both sides of the wall. The climate was then changed to 40 C and 10% relative humidity on both sides. Full papers - NSB 2014 page 1273

104 5. Results of the hygrothermal measurements in the climate simulator The plans for the measurements in the climate simulator had to be changed during the course of running the experiment. The malfunction during the first climate sequence meant that the wall was saturated after a very short time and also the interior side of the wall was wet. All the relative humidity sensors in the wall showed a relative humidity of 100% within less than 36 hours. The temperature could not be kept constant during this sequence so the temperature in the wall varied between 8.2 C and 21.2 C which makes an evaluation of the moisture measurement results complicated. During the second sequence, the equipment worked more in line with the expectations. The temperature was kept constantly around 10 C on the exterior side of the wall which gave a temperature between 9.4 C and 11.2 C in the middle of the wall. The measurements of the relative humidity in the wall during the second climate sequence are presented in FIG 3. FIG 3. Relative humidity in the wall. Left: relative humidity measured with the relative humidity sensors. Right: relative humidity measured with the Sahlén sensors translated from the wood moisture content by using the sorption isotherm for birch wood. The sensors are located in the mortar in the middle of the wall (RH3b, RH4b, S3b, S4b) and in the mortar in the inner part of the brick wall (RH3a, RH4a, S3a, S4a). The Sahlén sensor only measures relative humidity above 60%. It is clear that the moisture sensors in the middle of the wall are reached by the moisture faster than the sensors in the inner part of the brick wall. The Sahlén sensors show a significantly slower increase in relative humidity in the wall since the sensors uses a wooden material that has to absorb the moisture before the sensor can register the increasing moisture content. Also here, the relative humidity increases slower in the inner part of the brick wall in the first few days, but is then equal as in the middle of the wall. The relative humidity in the four wooden beams is shown in FIG 4. FIG 4. Relative humidity in the beams translated from the wood moisture content by using the sorption isotherm for spruce, in the middle of the beam end (b) and at the surface of the beam end (c). Left: lower beams (with mineral wool). Right: upper beams (without mineral wool). Full papers - NSB 2014 page 1274

105 There is a large difference between the relative humidity in the beams in the upper and lower part of the wall. This difference may partly be caused by the mineral wool insulation which is placed around the two lower beams, but not around the upper ones. The higher temperature in the wall results in a lower relative humidity, but this effect cannot explain the large difference on its own. The relative humidity sensors and Sahlén sensors in the upper part of the wall also showed a significantly lower relative humidity than what is shown for the lower part in FIG 3. The reason for this behavior is not clear. Part of it might be caused by the force of gravity on the liquid water flow. 6. Conclusions A laboratory study with a brick wall built in the laboratory and exposed to simulated driving rain in a large-scale building envelope climate simulator was conducted. The relative humidity in the wall increased substantially when exposed to driving rain. As expected from the simulations in the prestudy, the moisture increased faster in the mortar in the middle of the wall compared to in the inner part. The brick and mortar was more capillary active than expected. The lower part of the wall had a higher relative humidity which could be caused by the force of gravity, acting on the liquid water flow. The different sensors gave consistent results, although the Sahlén sensor is more appropriate to be used in measurements where the relative humidity is expected to change slower than what was the case in this study. In the wooden beams, the moisture content increased more in the end of the beam than close to the interior brick surface. Due to equipment malfunction, data from important measurement periods are missing, making it difficult to draw conclusions from the first sequence of the laboratory investigations. In the next phase of this project, a wall without VIPs will again be tested and compared to the wall with VIPs. The simulations will be compared to hygrothermal simulations. 7. Acknowledgements The work is supported by The Swedish Research Council Formas, the Lars Hierta Memorial Foundation, and finally the Research Council of Norway and several partners through The Research Centre on Zero Emission Buildings ( Porextherm Dämmstoffe GmbH is acknowledged for supplying the vacuum insulation panels and Wienerberger and St-Gobain Weber are acknowledged for supplying the brick and mortar. References Abuku, M., Blocken, B., & Roels, S Moisture response of building facades to wind-driven rain: field measurements compared with numerical simulations. Journal of Wind Engineering and Industrial Aerodynamics, 97(5-6), Baetens, R., Jelle, B. P., Thue, J. V., Tenpierik, M. J., Grynning, S., Uvsløkk, S., & Gustavsen, A Vacuum insulation panels for building applications: A review and beyond. Energy and Buildings, 42(2), Boverket Så mår våra hus - redovisning av regeringsuppdrag beträffande byggnaders tekniska utformning m.m. (The state of our buildings - report of governmental mission on the technical design of buildings etc.). [In Swedish]. Karlskrona, Sweden: Boverket. Boverket Regelsamling för byggande, BBR 2012 (Regulations for construction, BBR 2012). [In Swedish]. Karlskrona, Sweden: Boverket. Brick Industry Association Manufactoring of Brick. Reston, Virginia, USA: The Brick Industry Association. Brocken, H. J. P Moisture Transport in Brick Masonry: The Grey Area Between Bricks (Dissertation). Eindhoven, The Netherlands: Eindhoven University of Technology, Faculty of Architecture, Building and Planning, and Faculty of Applied Physics. Full papers - NSB 2014 page 1275

106 Fraunhofer IBP. (2010). WUFI 2D Transient Heat and Moisture Transport (Version 3.3.2) [Computer Program]. Holzkirchen, Germany: Fraunhofer IBP. Hall, C Water movement in porous building materials - I. Unsaturated flow theory and its applications. Building and Environment, 12(2), IEA Policy Pathway: Modernising Building Energy Codes. Paris, France: OECD/IEA and New York, NY, USA: United Nations Development Programme (UNDP). Johansson, P., Time, B., Geving, S., Jelle, B. P., Sasic Kalagasidis, A., Hagentoft, C.-E., & Rognvik, E Interior insulation retrofit of a brick wall using vacuum insulation panels: design of a laboratory study to determine the hygrothermal effect on existing structure and wooden beam ends. Proceedings of the 12th International Conference on Thermal Performance of the Exterior Envelopes of Whole Buildings, Clearwater Beach, Florida, USA, December 1-5, Kehl, D., Ruisinger, U., Plagge, R., & Grunewald, J Wooden beam ends in masonry with interior insulation - A literature review and simulation on causes and assessment of decay. Proceedings of the 2nd Central European Symposium on Building Physics, Vienna, Austria, September 9-11, KRD Byggteknisk forskrift (TEK 10). FOR Forskrift om tekniske krav til byggverk (Building Code. Regulations on technical requirements for construction). [In Norwegian]. Oslo, Norway: Kommunal- og regionaldepartementet, Bolig- og bygningsavd. Kvande, T., & Edvardsen, K. I Eldre yttervegger av mur og betong. Metoder og materialer (Older exterior brick and concrete walls. Methods and materials). [In Norwegian]. Oslo, Norway: SINTEF Building and Infrastructure. Piaia, J. C. Z., Cheriaf, M., Rocha, J. C., & Mustelier, N. L Measurements of water penetration and leakage in masonry wall: Experimental results and numerical simulation. Building and Environment, 61, Rasmussen, T. V Post-Insulation of Existing Buildings Constructed Between 1850 and Proceedings of the 11th International Conference on Thermal Performance of the Exterior Envelopes of Whole Buildings, Clearwater Beach, Florida, USA, December 5-9, NS-EN :2002. Methods of test for mortar for masonry - Part 18: Determination of water absorption coefficient due to capillary action of hardened mortar. Brussels, Belgium: European Committee for Standardization (CEN). NS-EN ISO 15148:2002. Hygrothermal performance of building materials and products - Determination of water absorption coefficient by partial immersion. Geneva, Switzerland: International Organization for Standardization (ISO). Thyholt, M., Pettersen, T. D., Haavik, T., & Wachenfeldt, B. J Energy Analysis of the Norwegian Dwelling Stock. Subtask A - Internal working document. IEA SHC Task 37 Advanced Housing Renovation by Solar and Conservation: International Energy Agency, Solar Heating and Cooling Programme. Ueno, K Masonry Wall Interior Insulation Retrofit Embedded Beam Simulations. Proceedings of the Building Enclosure Science & Technology Conference, BEST 3: High Performance Buildings - Combining Field Experience with Innovation, Atlanta, GA, USA, April 2-4, Full papers - NSB 2014 page 1276

107 Full paper no: 159 Energy savings in the Danish building stock until 2050 Kim B. Wittchen, Senior scientist, M.Sc., Civ. Eng. 1 Jesper Kragh, Senior scientist, M.Sc., Civ. Eng. 1 1 Danish Building Research Institute, Aalborg University, Denmark KEYWORDS: Renovation, Existing buildings, Energy savings, Regulation compliance, Energy projection SUMMARY: (Style: Summary Heading) A study has been conducted analysing the energy savings for space heating and domestic hot water in the Danish building stock due to renovation of building components at the end of their service life. The purpose of the study was to estimate the energy savings until 2050 as building components are energy upgraded according to the requirements stipulated in the Danish Building Regulations Furthermore, scenario analyses was made for the potential impact on the energy consumption of introducing different levels of tightening of the energy requirements for existing buildings in the Danish Building Regulations. Compliance with the requirements in the Danish Building Regulations will potentially result in energy savings for space heating and domestic hot water around 30 % until Further tightening of the component insulation level requirements will only result in marginally higher savings, due to the level of the current requirements. Higher energy savings can, though, be achieved e.g. by setting requirements for balanced mechanical ventilation with heat recovery and use of solar heating for domestic hot water. 1. Introduction As something relatively new, requirements for the minimum insulation level of building components have been introduced in the Danish Building Regulations 2010 (BR10) to be applied when the building components are being renovated, e.g. replacement of roof covering. However, it is a question how much these requirements will influence the energy consumption in the existing Danish building stock in Furthermore, the impact of further tightening of the requirements needs to be investigated. A calculation model for the net heating consumption in the entire existing Danish building stock until 2050 have been established in order to investigate the consequences of continuing with the current requirements and the effect of introducing stricter rules. The purpose of the analyses was to clarify the energy savings until 2050 if the building components are being upgraded according to the requirements stipulated in the Danish Building Regulation Upgrading is assumed to be introduced when the building components need renovation anyway due to the building materials used having reached the end of their service life. Additionally, the analyses were targeted at an investigation of the effect of introducing stricter requirements for the energy upgrading of building components in combination with planned refurbishment. Assumptions of when building components are going to be replaced originate from the Danish building and dwelling stock register (BBR) which holds information about the building materials used in facades and roofs. This information combined with knowledge about the age of the building and estimates for the probable service life of different building materials gives an Full papers - NSB 2014 page 1277

108 estimate of the replacement rate and hence the rate of energy upgrading of the existing building stock. 2. Background In the Danish Building Regulations 2010 there are ultimate minimum energy requirements that must be followed when replacing windows. For roofs, external walls etc. there are requirements that must be met if it is economically, architecturally and technically feasible to meet them as part of a retrofitting process. The examples in Table 1 are specifically listed in BR10 as normally being economically feasible. Table 1. Examples, mentioned in the Danish Building Regulation 2010, BR10, as often being economically feasible. Building component Insulation thickness that is economically feasible to upgrade [mm] Total insulation thickness after upgrading [mm] Accessible attic < Sloping walls and ceiling to ridge < Space under the roof slope < Flat roof < Lightweight external wall < Cavity masonry wall Uninsulated Cavity wall insulation Massive external wall in brickwork External wall in lightweight concrete < Floor structure above unheated basement - Insulation between beams Floor above unheated basement < Floor above accessible crawl space < Floor above free space < Slab on ground Uninsulated 250 In the analyses, the above values imply that an accessible attic is insulated if the total U-value of the roof construction is above 0.20 W/m²K, and that it will have a U-value of 0.15 W/m²K after upgrading. BR10 also set rules for the replacement of windows, and here new windows must have an annual average energy balance of no less than -33 kwh/m² per year calculated for a standard window regarding size, configuration and average orientation. This requirement will become stricter in 2015 when the energy balance should be at least -17 kwh/m² per year, and even stricter in 2020 where the energy balance should be at least 0 kwh/m² per year. The table below shows an overview of the energy requirements for building components that should be complied with in combination with renovation and extensions of existing buildings. Compliance is mandatory unless it is not economically, architecturally or technically feasible to comply with the level of the requirements. In these cases, the building components should be insulated up to the feasible level. Full papers - NSB 2014 page 1278

109 Table 2. Requirements to building components in combination with renovation and extensions of existing buildings as stated in the Danish Building Regulations Building component W/m²K External walls and basement walls towards the ground 0.2 Internal walls and floors towards unheated rooms or rooms heated to a 0.4 temperature more than 5 K lower than the current room Slab on ground, basement floors towards ground and floors above outdoor 0.12 spaces or ventilated crawl spaces Ceiling and roof constructions, including walls towards spaces under the 0.15 roof, flat roofs and sloping walls directly towards the roof External doors, gateways, hatches, removable windows and dome lights 1) 1.65 Windows kwh/m² per year In facades 2) -33 In roofs 2) -10 1) When replacing windows after 1 January 2015, the U-value (incl. frame) should not exceed 1.40 W/m²K. 2) When replacing windows after 1 January 2015, the energy balance over the heating season should not be lower than -17 kwh/m² per year and for roof windows not lower than 0 kwh/ m² per year. When replacing windows after 1 January 2020, the energy balance for facade windows should not be lower than 0 kwh/m² per year. 3. Method and assumptions The purpose of the analyses was to estimate the energy savings to be expected until 2050 if buildings and building components are being upgraded according to the requirements laid down the in the Danish Building Regulations 2010, when they have to be replaced or renovated for other reasons. A model of the energy consumption in the existing Danish building stock was set up, based on information extracted from the database of the Danish energy performance certification scheme (EMO) and extrapolated to cover all Danish buildings using data from the Danish building and dwelling stock register (BBR). The model compares the calculated energy consumption of the building stock according to registrations made by energy certification experts with the theoretical energy consumption in the same buildings after energy upgrading. Energy flows included in the model are: space heating energy, ventilation, and domestic hot water. The calculated energy consumption for the building stock as-is was compared with the 2011 energy statistics by the Danish Energy Agency (Energy statistics, 2011) and showed a discrepancy of 6 % in comparable building categories. Potential energy savings from improvements of the technical installations in the buildings are thus not part of the analyses. Often, there will be architectural considerations in combination with external insulation of external walls made of masonry, which is the predominant building material for external walls in Denmark. This, in combination with the long service life of this kind of external walls, sets limits for the share of masonry external walls that are expected to have external insulation. For older blocks of flats in major cities constructed before 1950, there is a potential for external insulation of the walls on the back of the buildings without violating the architectural impression of the street view. The amount of this area is difficult to estimate from the available information. For other types of materials used in external walls, e.g. concrete and lightweight facades, there will normally not be the same architectural constraints against adding external insulation. It is thus assumed that 0.5 % of the masonry walls are being energy upgraded every year until The share of energy upgrades of slabs on ground is evaluated to be modest and normally related to establishing floor heating, e.g. in bathrooms. In contrast to this, floors above basements and Full papers - NSB 2014 page 1279

110 accessible crawl spaces are expected to be subject to energy upgrading. It is not possible to identify a certain point in the lifetime of a building when these floors will be upgraded. Therefore, it was estimated that 15 % of the floors above basements and accessible crawl spaces will be upgraded up to Energy upgrading The material used as roof covering is registered in the BBR. From knowledge about the year of construction and average service life of different building materials (GI, 2013), the future replacement rate of roofs are estimated and used to calculate the upgraded energy performance of the existing building stock. Roof covering of older buildings is expected already to have been replaced, but not necessarily energy upgraded, one or more times since the building was constructed. The share of these roofs that are being replaced every year is estimated to be 1 divided by the average service life of the roofing material. Similar assumptions are made for the other building components. The insulation level for the existing building stock is based on registrations made by building experts in the building energy certification scheme. The average insulation levels were calculated for different typical construction periods and building types as area-weighted U-values Ventilation systems and solar thermal systems Establishing balanced mechanical ventilation systems with heat recovery will have a growing relative impact as the insulation level of the buildings increases. It is therefore expected that these systems will increase in numbers over time. The effect of heat recovery was calculated by introducing an average efficiency in combination with an estimated airchange rate. It is further assumed that airtightness of the renovated buildings are being dealt with in combination with replacement of windows and external doors. As a starting point, mechanical ventilation is not anticipated in combination with ordinary renovation works as it is not mandatory according to BR10. The effect of this measure was evaluated in a special scenario where mechanical ventilation with heat recovery was assumed to be installed in combination with replacement of roof covering of buildings with sloping roofs. The same consideration is valid for the installation of solar thermal systems for covering of a share of the energy consumption for domestic hot water. 3.2 Energy model for the existing building stock The Danish building stock is divided into different types of buildings and furthermore into different typical age classes. Within each type and age class, the original buildings energy performance is assumed to be more or less uniform. Based on statistical data from the EMO scheme regarding the current insulation levels and areas per unit for each of the buildings components (facades, roofs, floors, windows and doors), it is possible to create a model for each type and period, e.g. single family houses constructed between 1961 and From the calculated average energy consumption in each type and age class, it is possible to extrapolate to the consumption of the entire Danish building stock. The results of these calculations were then compared with and tuned according to the energy consumption in the same groups in the Danish Energy Agency s energy statistics (Energy statistics, 2011). The model for the energy consumption in each class includes heat losses through the thermal envelope, ventilation losses and energy used for domestic hot water. On the gains side, loads from persons, appliances and sun through windows are included. A degree-day method was used to calculate the energy consumption in the existing building stock and for calculation of energy savings in combination with planned building refurbishment. Full papers - NSB 2014 page 1280

111 3.3 Service life for building materials The service life for building components is base for the expected life time for the various building materials used for external constructions. The expected life times for building materials is extracted from GI (2013) and Larsen (1992). The life time for a material depends on many factors like roof covering, roof boarding and exposure. The average material life times are shown in the table below. Table 3. Estimated average life times for external building materials used in the analyses. Roof covering Life time [years Flat roofs 35 Asphalt board (sloping roofs) 35 Fibre concrete, incl. asbestos tiles and slates 40 Cement tiles 60 Roof tiles 60 External wall covering Bricks (clay, lime-sand stone, cement stone) 75 Lightweight concrete (light blocks, porous concrete) 60 Sheets of fibre cement, incl. asbestos 45 Wooden boards 40 Concrete components 40 Windows 25 For new windows, especially those made of plastic or combined wood/metal, the life time of the frame can be longer, but in this analysis it is not of major importance for the energy savings by The figure below indicates the works associated with energy upgrading of roof insulation by The drop in activity after 2044 is due to the fact that some roof coverings are changed twice during the period and only the first replacement results in energy upgrading. Under this assumption, 81 % of the total roof area will be thermally upgraded by FIG 1. Estimated development in roof covering replacements that lead to energy upgrading of the roofs on the existing buildings. Over the period 82 % of the roofs are estimated to be replaced and consequently energy upgraded. Full papers - NSB 2014 page 1281

112 Over the period leading up to 2050, all windows are estimated to be replaced and 18 % of the external walls to be thermally upgraded. The low ratio of upgraded external walls is due to the long lifetime of the dominant masonry wall and because of architectural constraints for changing the appearance of these kinds of walls. A fixed share of 0.5 % of the available area of external masonry walls is assumed to be upgraded every year, and this dominates the total refurbishment works on external walls. 4. Calculation results The table below shows results from 11 different scenarios. The scenarios set out to investigate the effects of different suggestions for more strict requirements to building components and other initiatives to increase energy savings in the existing building stock by 2050 and compare this with what happens if rules will remain at the same level as the current requirements. Furthermore, one scenario analysed the consequence of a prolonged service life of the roofing materials. In the business-as-usual scenario A0, only 80 % of the potential area is assumed to be upgraded due to architectural or technical constraints. The A scenarios analyses different levels of implementing building energy upgrading by The B scenarios analyse different tightening of the component requirements in the Danish Building Regulations. Table 4. Calculated net heating energy consumption by 2050 in each of the different scenarios. Scenario Energy consumption in 2050 TJ/year Energy savings compared with today % Energy savings compared with scenario A0 %-point Status A0 Business-as-usual % - A1 Full BR compliance % 3.7 % A2 90 % BR compliance % 1.8 % A3 Longer life of roofs 1) % -3.4 % A4 All roofs insulated before % 1.5% A5 Fast implementation of A windows 2) % 0.0 % B1 More tight requirements for roofs + A % 2.4 % B2 More tight requirements for external walls + A % 2.7 % B3 More tight component requirements + A % 3.2 % B4 Extra tight requirements for roofs + A % 2.7 % B5 Extra tight requirements for external walls + A % 3.5 % B6 Requirements for A+ windows + A % 4.3 % B7 Automation and effectiveness + A % 3.5 % B8 B9 Extra tight component requirements = B4+B5+B % 6.9 % More tight component requirements and A+ windows = B1+B2+B % 5.7 % B10 Automation and effectiveness + B % 7.4 % C1 BMV with VGV + B % 19.2 % Faster implementation of stricter requirements when replacing windows (Scenario A5) will not result in lower energy consumption by 2050, unless new window types are invented. The reason Full papers - NSB 2014 page 1282

113 for this is the short lifetime of windows that ensures that all windows have been upgraded to comply with the 2020 requirements before 2050 even for windows replaced in The general insulation level of existing Danish buldings is rather high, and only a limited number of the traditional energy-saving measuers are thus economically feasible. To be able to meet the government s target that Denmark is to become free of fossil fuels by 2050 while for heating of buildings by 2035, more rigid requirements need to be considered. Among those are requirements for implementing balanced mechanical ventilation with heat recovery in residential buildings with a sloping roof (enough free space in the attic to install the ventilation system) in combination with roof retrofit (Scenario C1). The estimated service life of roof covering materials can be questioned and a special scenario have thus been made to analyse the effects of a longer service life of these materials. A 25 % extension of the service life for roofs (Scenario A3) only results in a decrease of the retrofitted area of about 5 %, which only has a marginal influence on energy savings by Furthermore, there is a hump of roofs on buildings constructed in the 1970s, which even with a 25 % prolonged service life will need to be replaced before The next figure shows the development of energy use for space heating, ventilation and domestic hot water in each scenario. It is clear that installation of balanced mechanical ventilation with heat recovery in combination with renovation of sloping roofs will have a significant effect on the energy consumption by This is not surprising as only a marginal share of the existing building stock has been equipped with this kind of systems yet. To be able to save this amount of energy, it is however a pre-condition that airtightness of the buildings have been improved at the same time, e.g. in combination with replacement of windows. FIG 2. Development in net energy consumption for space heating, ventilation and domestic hot water in the existing Danish building stock as analysed in the A Scenarios. Energy saving measures are assumed to be implemented at the same rate as the building components are being retrofitted due to the end of their service life. Full papers - NSB 2014 page 1283

114 Generally, all curves bend around 2037 and that is the time when all windows have been upgraded at least once during the period, and no further energy savings can be expected from window upgrading except if further technical improvements of window technology are implemented and requirements in the Danish Building Regulations are being tightened further. 5. Conclusions It is the aim of the Danish government that Denmark should be free of fossil fuels by 2050 while for heating buildings this should happen in To be able to reach that goal, it is estimated that the energy consumption in the existing building stock should be reduced by about %. Following the current path, with energy upgrading of building components in compliance with the requirements in the Danish Building Regulations 2010 (BR10) when retrofitting the buildings due to termination of service life for the building components, will not result in enough energy savings (about 30 % of the 2011 national energy use in buildings) to reach that goal. To be able to come closer to that goal there is a need for more strict requirements in combination with refurbishment works and also improvements of the energy performance of windows. It is possible to get more energy efficient windows with an annual energy balance value of +15 kwh/m² per year for facade windows. Introduction of mechanical ventilation with heat recovery can also contribute to fulfilment of the goal. The analyses have not taken into account demolishing of existing buildings and replacement with new buildings by Historically, about 1 % of the Danish building stock is replaced every year. If this trend continues over the next 30 years, about one third of the buildings will be newly built by 2050 and have a significant lower energy need. 6. Acknowledgements The work has been funded as part of the Danish Energy Agency s activities for establishing a national strategy for energy upgrading of the existing building stock. 7. References Energy statistics 2011, Danish Energy Agency, Available at: Larsen, E. S Service life of concrete constructions (In Danish: Betonkonstruktioners levetid). SBi report 225. Danish Building Research Institute, Hørsholm, Denmark. GI: Danish Building Owners Investment Fund (GI: Grundejernes Investeringsfond), Danish Building Regulations 2010 (December 2010). The Danish Ministry of Economic and Business Affairs, Copenhagen, Available at: Wittchen K.B., Kragh J. and Aggerholm S. (March, 2014). Potential heating energy savings in the Danish building stock until (In Danish: Potentielle varmebesparelser ved løbende bygningsrenovering frem til 2050). SBi Danish Buildings Research Institute, Aalborg University, Copenhagen, Denmark. Full papers - NSB 2014 page 1284

115 Full paper no: 160 User behaviour impact on energy savings potential Jørgen Rose, MSc. Civ. Eng., Ph.D 1 1 Danish Building Research Institute, Aalborg University, Denmark KEYWORDS: User behaviour, energy upgrading, energy savings potential, indoor temperature, internal heat gain, domestic hot water consumption, air change rate SUMMARY: (Style: Summary Heading) When buildings are to undergo energy upgrading in Denmark, the national compliance checker, Be10, is often used to calculate expected energy savings for different energy-saving measures. The Be10 calculation is, however, very dependent on a variety of standard assumptions concerning the building and the residents' behaviour and if these defaults do not reflect actual circumstances, it can result in non-realisation of expected energy savings. Furthermore, a risk also exists that residents' behaviour change after the energy upgrading, e.g. to obtain improved comfort than what was possible before the upgrading and this could lead to further discrepancies between the calculated and the actual energy savings. This paper presents an analysis on how residents behaviour and the use of standard assumptions may influence expected energy savings. The analysis is performed on two typical singlefamily houses corresponding to different levels of energy consumption. The purpose of the analysis is to identify the importance of each of the four primary user-related parameters in terms of their relative and combined impact on the overall energy needs before/after upgrading; 1) Indoor temperature, 2) Internal heat gain, 3) Domestic hot water consumption and 4) Air change rate. Based on the analysis, a methodology is established that can be used to make more realistic and accurate predictions of expected energy savings associated with energy upgrading taking into account user behaviour. 1. Introduction User behaviour plays an important role for a building s energy consumption and in connection with energy upgrading of existing buildings, user behaviour may lead to non-realisation of expected energy savings. Most often failure to achieve energy savings occur because users gain the possibility and focuses on increased comfort instead, e.g. through a slight increase in temperature or air change rate. User behaviour influence on energy consumption in buildings has been dealt with in numerous articles and reports and is not a new topic, e.g. (Lundström, 1986). A state-of-the-art review on occupants influence on the energy consumption in buildings was performed by Larsen et al (2010). This analysis was performed as part of the Danish Energy Agency Network for Energy Renovation aiming to support the establishing of future energy-policies in Denmark. The purpose of the analysis is to identify the importance of four primary user-related parameters in terms of their relative and combined impact on the overall energy consumption before/after the energy upgrading: 1. Indoor temperature 2. Internal heat gain 3. Domestic hot water consumption 4. Air change rate Based on the analysis, a methodology is established that can be used to make more realistic and accurate predictions of expected energy savings associated with energy renovation taking into account Full papers - NSB 2014 page 1285

116 user behaviour. The purpose is to develop a method which provides an energy calculation, based on a specific combination of parameters corresponding to a specific family in a specific building. Furthermore, the purpose of the analysis is to provide an overview of how user behaviour affects the expected energy savings in buildings, and thus try to establish a method for determining the expected energy savings, taking into account user behaviour. 2. Method The analysis is performed using 2 buildings representing typical single-family houses from 2 different periods; the 1930s and the 1960s. The following gives a brief description of the 2 buildings s The house is a typical bungalow from 1932 with a gross heated area of 103 m 2. The building has a full, unheated basement less than half below ground level with a gross area of 103 m 2. The total window area on the ground floor is 16% of the floor area. The total glass area on the ground floor is 12.0 m 2. FIG 1. Typical single-family house from 1930s U-values for building constructions The U-values are summarised in Table 1. Note that windows are assumed to have been changed during the 1960s, and now correspond to traditional double-glazed windows. TABLE 1. U-values for building constructions Building construction U-value [W/m 2 K] Floor separation 1.02 Exterior wall, mean 1.45 Ceiling 0.55 Windows 2.70 Basement wall 1.24 Basement floor 0.40 Full papers - NSB 2014 page 1286

117 2.1.2 Heating and ventilation The house has an old oil boiler in the basement connected to a 2-pipe heating system. All pipes are insulated with 10 mm insulation. The hot water tank holds 200 l with 30 mm insulation. The building has natural ventilation and can be categorised as leaky, which means that the total air change rate in the house is set at 0.45 l/s per m Calculated energy consumption Calculation of energy consumption is based on the Danish compliance checker, Be10 (Aggerholm and Grau, 2012). The calculation covers energy consumption for heating, cooling, ventilation and domestic hot water. The 1930s house has a calculated energy consumption of 417 kwh/m 2 per year s The house is a typical single-family house from the 1960s with a gross heated area of 108 m 2. The house consists of lounge, kitchen/dining area, utility room/bathroom, hall, toilet and 3 bedrooms. The total window area is 22% of the floor area. The total glass area is 19.9 m 2. FIG 2. Typical single-family house from 1960s U-values for building constructions The U-values are summarised in Table 1. TABLE 2. U-values for building constructions Building construction U-value [W/m 2 K] Exterior wall, heavy 0.46 Exterior wall, light 0.49 Ceiling 0.39 Windows 2.70 Slab on ground 0.30 Full papers - NSB 2014 page 1287

118 2.2.2 Heating and ventilation The heating distribution system is a 2-pipe heating system with a flow temperature of 80 C and return temperature of 60 C. The heating system is an old oil boiler unit located in the utility room. All pipes are insulated with 30 mm insulation. Domestic hot water is produced in a 200 l hot water tank with 30 mm insulation. The building has natural ventilation and can be categorised as leaky, which means that the total air change rate in the house is set at 0.45 l/s per m Calculated energy consumption The single-family house has a calculated energy consumption of 240 kwh/m 2 per year. 2.3 Energy-saving measures For each building, 2 packages of energy-saving measures are suggested s, energy-saving measures, Package 1 The following measures are carried out: 1. Floor separation: 70 mm clay replaced by 75 mm insulation 2. Ceiling: New 300 mm insulation 3. Exterior wall: 150 mm exterior insulation 4. Windows: Replaced by Class A windows (U = 0.9 W/m 2 K and g = 0.62) 5. Heating supply: Connection to district heating instead of old oil boiler The total energy consumption is reduced from kwh/m 2 per year to kwh/m 2 per year s, energy-saving measures, Package 2 The following measures are carried out: 1. Floor separation: 70 mm clay replaced by 75 mm insulation 2. Ceiling: New 300 mm insulation 3. Exterior wall: 150 mm exterior insulation 4. Windows: Replaced by Class C windows (U = 1.3 W/m 2 K and g = 0.62) 5. Air tightness: Improved (from 0.45 to 0.30 l/s per m 2 ) 6. Mechanical ventilation: 90% heat recovery 7. Heating supply: District heating instead of old oil boiler The total energy consumption is reduced from kwh/m 2 per year to kwh/m 2 per year s, energy-saving measures, Package 1 The following measures are carried out: 1. Exterior wall: 160 mm insulation for the heavy wall and 125 mm insulation for the light wall 2. Ceiling: New 200 mm insulation added to existing 100 mm 3. Windows: Replaced by Class A windows (U = 0.9 W/m 2 K and g = 0.62) 5. Air tightness: Improved (from 0.45 to 0.30 l/s per m 2 ) 6. Mechanical ventilation: 90% heat recovery 7. Heating supply: Ground source heat pump instead of old oil boiler The total energy consumption is reduced from kwh/m 2 per year to 78.5 kwh/m 2 per year. Full papers - NSB 2014 page 1288

119 s, energy-saving measures, Package 2 The following measures are carried out: 1. Exterior wall: 160 mm insulation for the heavy wall and 125 mm insulation for the light wall 2. Ceiling: New 200 mm insulation added to existing 100 mm 3. Windows: Replaced by Class A windows (U = 0.9 W/m 2 K and g = 0.62) 5. Air tightness: Improved (from 0.45 to 0.30 l/s per m 2 ) 6. Heating supply: District heating instead of old oil boiler The total energy consumption is reduced from kwh/m 2 per year to 97.7 kwh/m 2 per year. 3. Energy savings as a function of user behaviour Chapter 2.3 has shown expected energy savings for two different buildings and for different energy saving measure packages. These calculations are based on standard assumptions concerning user behaviour, e.g. 20 C indoor temperature etc. To evaluate influence of user behaviour, new sets of calculations are performed where four primary user-related parameters vary. Table 5 shows variations. TABLE 5. Variation of parameters in calculations Parameter Variation Unit Indoor temperature C Internal heat gain W/m 2 Domestic hot water consumption l/m 2 Air change rate (natural ventilation) l/s per m 2 Air change rate (infiltration) l/s per m 2 Calculations are performed for each individual package for the parametric variations shown in Table 5. Calculation results for the 1930s single-family house are shown in Tables 6 9. TABLE s, Package 1. Relative energy savings in % as a function of indoor temperature before/after energy upgrading. Indoor temperature after [ C] Indoor temperature before [ C] TABLE s, Package 1. Relative energy savings in % as a function of internal heat gain before/after energy upgrading. Internal heat gain after [W/m 2 ] Internal heat gain before [W/m 2 ] Full papers - NSB 2014 page 1289

120 TABLE s, Package 1. Relative energy savings in % as a function of domestic hot water use before/after energy upgrading. Domestic hot water use after [l/m 2 ] Domestic hot water use before [l/m 2 ] TABLE s, Package 1. Relative energy savings in % as a function of air change rate before/after energy upgrading. Air change rate [l/s per m 2 ] Air change rate [l/s per m 2 ] The tables show the relative energy savings, e.g. if the indoor temperature is 20 C before the energy upgrading and 22 C after, then the relative energy savings are 93.6% of the expected energy savings. The before situation could also correspond to a situation where no data is available and therefore a standard value is assumed. Similar calculations are performed for the remaining packages, i.e. Package 2 for 1930s and Packages 1 and 2 for 1960s. A cross comparison shows that the lower the total energy consumption is, the more the relative savings are influenced, i.e. the expected energy savings for the 1960s building are more sensitive to discrepancies between parameters in the before and after situations. 4. Discussion 4.1 Indoor temperature The indoor temperature greatly affects the energy consumption of the building, and the analysis shows that for every degree the inside temperature deviates from standard assumptions, the energy consumption is increased/decreased by 6 8%. This applies regardless of the level of the total energy consumption. The analysis also shows that the higher the indoor temperature, the greater energy savings will be achieved in connection with an energy upgrading. If the indoor temperature changes in connection with an energy upgrading, e.g. 2 C, then the relative savings are reduced by 4 7% in a house from the 1930s and 8 12% in a house from the 1960s. 4.2 Internal heat gain The internal heat gain greatly affects the energy consumption of the building and the lower the energy consumption of the building, the greater the relative importance of variations in the internal heat gain. In the non-upgraded buildings, 1 W/m 2 deviation in the internal heat gains influences the energy Full papers - NSB 2014 page 1290

121 consumption by 2 3%, for the upgraded buildings from the 1930s about 4 5% and the upgraded buildings from the 1960s about 5 6%. The analysis also shows that the energy savings achieved are largely independent of the level of the internal heat gain if it is the same after as before. The energy-saving potential is thus largely independent of the internal heat gain. If the internal heat gain changes in connection with energy upgrading, then the relative savings are reduced by 2 3% in a house from the 1930s and 3 4% in a house from the 1960s for each 1 W/m 2 change. 4.3 Domestic hot water consumption Consumption of domestic hot water affects the total energy consumption of the building with the same level, regardless of the building's overall energy state, and therefore the deviations in hot water consumption is most important in buildings that have undergone extensive energy upgrading. The analysis shows that for every 50 litres/m 2 per year, the consumption of hot water differs from the standard assumption of 250 litres/m 2 per year, the total energy consumption is increased/decreased by approximately 1% for the non-upgraded buildings and approximately 2 3% for the energy-upgraded buildings. The analysis also shows that the energy savings achieved are largely independent of the consumption of domestic hot water, if the level of consumption is the same after as before. The energy-saving potential is thus largely independent of the consumption of domestic hot water. If the consumption of domestic hot water changes in the course of an energy upgrading e.g. increases by 50 litres/m 2 per year, then the relative savings are reduced by approximately 1% in a house from the 1930s and 1-2% in a house from the 1960s. 4.4 Air change rate The air change rate affects the energy consumption of the building to some extent and the lower the energy consumption the greater the significance of the air change rate. In the 1930s house an increase in air change of 0.05 l/s per m 2 results in an increase in energy demand of approximately 4%. In the 1960s house, an increase in air change results in an increase in energy demand of approximately 6%. The analysis also shows that the lower the air change rate, the greater the savings that are achieved in the context of an energy upgrading. The energy-saving potential is thus dependent on air change rate. If the air change rate increases, e.g l/s per m 2 in the context of an energy upgrading, then the relative savings are reduced by approximately 2% for the 1930s house and approximately 4% for the 1960s house. 4.5 Combined effects Domestic hot water consumption does not influence the energy balance of the building and effects can be calculated independently of other parameters. The other three parameters are, however, interdependent, and the overall impact on the building's energy needs cannot be determined by a simple summation of individual effects. However, the effect of combining parameters is still quite limited and the only case where it is actually necessary to adjust the total energy savings is for the combination of indoor temperature and ventilation rate. This can be achieved by a simple calculation of the extra ventilation heat loss that occurs based on the change in temperature (compared to 20 C) and the change in air change rate (compared to 0,13 l/s pr. m 2 ), i.e.: kg / m 1,205J / kgk ( v 0,13) l / s pr. m ( T 20 K v a a ) Where v a is the actual ventilation rate in l/s pr. m 2 and T a is the actual temperature in C. The error introduced by simply adding the individual effects but taking into account the abovementioned correction for combinations of indoor temperature and ventilation rate will be in the Full papers - NSB 2014 page 1291

122 order a few percent maximum. This way a method for predicting the energy saving potential can be based on similar analysis for different types of buildings and building use. 5. Conclusions This analysis has shown that the levels of the internal heat gain and the consumption of domestic hot water only has a modest impact on the energy-saving potential of buildings, as long as the value of the parameters does not change in the course of an energy upgrading. None of these parameters are directly related to comfort, and therefore they will typically not be changed in the process. The indoor temperature and air change rate in the building both affect the energy-saving potential. Both parameters are directly related to the comfort in the buildings and are therefore parameters that could potentially be changed in connection with an energy upgrading. The indoor temperature is clearly the more important of the two parameters, and the results of the analysis shows that for every degree the indoor temperature is raised after an energy upgrading, the expected savings are reduced by approximately 6 8%, i.e. the lower the total energy demand of the building the more significant the influence of the indoor temperature. The air change rate is less important, but it is clear that if there are large differences between the assumption of level and actual level it can affect energy savings significantly. The significance of the air change rate is highly dependent on the building's total energy consumption and the lower the energy consumption, the greater the significance of the air change rate. The results of the analysis show that for every 0.01 l/s per m 2 difference between the air change rate before and after the energy upgrading, the expected energy savings are reduced by approximately 0.4 to 0.8%, depending on the overall energy consumption. This may not sound of much, but if the air change rate changes from the minimum requirement for new buildings (0.30 l/s per m 2 ) to a level where it can be categorised as a leaky building (0.45 l/s per m 2 ), the expected energy savings are reduced by up to 12%. Based on the analysis a relatively simple method for determining of energy savings for energy upgrading measures can be developed. The aim would be to develop a method that can predict energy savings for specific energy saving measures in a specific building, taking into account user behaviour. 6. Acknowledgements This work was financed by the Danish Energy Agency (Energistyrelsen). References (Times New Roman 14 pt bold, Style: References Heading) Lundström, E. 1986; Occupant influence on energy consumption in single-family dwellings. Swedish Council for Building Research, Document D5:1986, Sweden. Larsen, T. S., Knudsen, H. N., Kanstrup, A. M., Christiansen, E., Gram-Hanssen, K., Mosgaard, M., Brohus, H., Heiselberg, P. and Rose, J. 2010; Occupants influence on the energy consumption of Danish domestic buildings - State of the art. DCE Technical Report No Aalborg University, Denmark. Aggerholm, S. and Grau, K. 2005; Bygningers energibehov - Pc-program og beregningsvejledning. (Building energy demand PC program and user guide) SBi-Anvisning 213. Statens Byggeforskningsinstitut (SBi), Hørsholm, Denmark. Full papers - NSB 2014 page 1292

123 Full paper no: 161 Façade integrated active components in timber-constructions for renovation - a case study Fabian Ochs, Dr.-Ing. 1 Georgios Dermentzis, Dipl.-Ing. 1 Dietmar Siegele, Dipl.-Ing. 1 Alexandra Konz, Dipl.-Ing. 1 Wolfgang Feist, Prof. Dr. 2 1 University of Innsbruck, Unit for Energy Efficient Buildings, Austria 2 Passive House Institute, Darmstadt, Germany KEYWORDS: Façade integrated active components Deep Renovation SUMMARY: Deep renovation to high energy efficiency plays a key role in saving energy and in reducing CO 2 - emissions. In the framework of the EU project inspire (fp7), renovation kits are developed and energy efficient renovation packages are investigated with the aim to achieve a primary energy demand of maximum 50 kwh/(m2 a) for heating, domestic hot water, auxiliary energies and lighting). As one approach a façade integrated micro-heat pump (mechanical ventilation with heat recovery and exhaust-air heat pump) is being developed. A prototype is measured in PASSYS test cells and will later monitored in a demo building in Ludwigsburg, Germany. In the paper the approach of the micro-heat pump is discussed considering building physics and energy performance. The potential of the system is investigated by means of building and system simulation. The façade integrated micro-heat pump is a promising concept for renovated and new buildings with a very low heating demand. With the prefabricated elements a minimum invasive renovation is enabled. The concept has the potential to deliver heat with reasonable efficiency at very low cost. 1. Introduction The majority of existing building stock in Europe and worldwide consists of poor energy performance buildings. Deep renovation to a high energy efficiency standard e.g. according to the EnerPHit standard (heating demand of 25 kwh/(m 2 a)) plays a key role in saving energy and in reducing CO2- emissions (Feist 2012). In order to meet the target of a primary energy demand of maximum 50 kwh/(m 2 a) the aim of the EU project inspire (fp7) is: Deep renovation of the existing buildings through a systemic approach which includes integrated concepts consisting of building and system technologies. Development of energy efficient renovation packages (integrated solutions) and of energy efficient 'kits' such as multifunctional systems, including energy production, distribution and storage technologies, integrated into the envelope system 2. Concept of Micro-heat Pump The micro-heat pump (µhp) is a concept for very efficient buildings - renovated and new buildings with a very low heating demand of 25 kwh/(m² a) or below such as e.g. EnerPHit standard (see corresponding to a specific heat load in the range of 10 W/m². For such high performance buildings reasonable energy efficient and cost effective heating systems are required. A façade integrated µhp (mechanical ventilation with heat recovery and exhaust-air heat pump) is Full papers - NSB 2014 page 1293

124 developed as one renovation kit in the framework of the project inspire for very efficient residential buildings. The µhp has a heating capacity of approx. 1 kw (300 W el speed controlled compressor). Basically, the µhp concept would work for water (radiator, floor heating, radiant ceiling) and air based systems (supply air and principally also recirculated air). As source ambient air and/or exhaust air or brine are possible. The exhaust air-to-supply air has the highest potential to be really micro and thus compact, see FIG. 1., extract supply HRC defros t ambient, extract supply HRC defrost ambient exhaus t exhaust µ-hp µ-hp FIG 1. Scheme µhp with mechanical ventilation with heat recovery (MVHR); extract air-to-air (left) and ambient air-to-water (right) The µhp is a system for decentralized heating. With a reversible heat pump the system could be used also for cooling. Domestic hot water (DHW) preparation has to be solved separately. For multi-family houses domestic hot water can be prepared by a central system using e.g. heat pump and/or solar thermal collectors. In Ochs et al heating of very efficient SFH such as a passive house (heating demand of 15 kwh/(m² a)) with a heat pump in combination with a simple solar DHW preparation with direct electric backup is investigated (see FIG. 2). Reasonable to good performance (PE < 50 kwh/(m² a)) can be obtained if a high solar fraction of about 70 % is obtained (equal to about 10 m² of SC) and the heat pump is operated with a SPF of about 2.5. SC TI BS HXDHW HP CW (10 C) HW (45 C) FH or radiator TI CW: cold water HW: hot water FH: floor heating BH: backup heater SC: Solar collector HP: Heat pump BS: Buffer storage evaporator circulation pump control FIG 2. Scheme of a ambient air-to-water µhp for heating and independent solar DHW preparation Façade integration offers several advantages (see the floor plan of the demo building in Ludwigsburg in FIG 3 as an example, the demo building is discussed below in detail): 1. No additional space for MVHR and heating system is required. 2. Cold ducts (i.e. ambient air and exhaust air) are short and outside the thermal envelope. 3. Extract air ducts are completely placed inside the façade. The inlet of the extract air is placed in the reveal of the window; all ducts are prefabricated and part of the façade. 4. Minimum installation inside the flat (i.e. minimum disturbance of tenants) Full papers - NSB 2014 page 1294

125 µhp FIG 3. (left) Floor plan of GF of demo building in Ludwigsburg (D) with position of the façade integrated MVHR with µhp; (right) integration of extract air ducts in the timber frame façade Filter accessibility from inside is possible via the reveal, however one important issue has to be solved: maintenance of the MVHR and heat pump from outside in case of high rise buildings. 3. Lab Measurements and Prototype A prototype of a timber frame façade with an integrated MVHR is produced and measured in the lab at UIBK, see FIG 4. A second prototype with MVHR and µhp is under construction. supply air extract air: external filter and silencer MVHR Air-to-air heat exchanger ambient air: external filter and silencer exhaust air: silencer and outlet Filterbox and silencer duct s MVHR FIG 4. (left) 3D Sketchup model and (right) photos of the production of the prototype (façade integrated MVHR), photos: Gumpp & Maier GmbH During lab measurements aspects of building physics and the energy performance of the MVHR and the µhp will be measured. Two so-called PASSYS test cells and an acoustic test cell are available. The following parameters will be measured: - building physics (U-Value, g-value, avoidance of moisture accumulation inside the construction and of mould growth, sound emission and sound insulation) - energy performance of the active component (ηhr efficiency of mechanical ventilation with heat recovery, COP (coefficient of performance of heat pump) Full papers - NSB 2014 page 1295

126 For the façade integrated MVHR thermal and hygrothermal simulations have been conducted on component level (2D) and the system has been optimized to allow a secure operation. By means of lab measurements of a prototype these aspects are investigated in detail and simulation results will be proven. First measurement results are expected to be available in spring Demo Building Ludwigsburg The Lubu demo building is a multi-family house is located in a town area in Ludwigsburg. The 1971 residential building consists of four flats: Basement floor with living area approx. 40 m²; Ground floor and first floor: each 1 flat, living area approx. 90 m² (with south loggia); Attic storey: 1 flat, living area approx. 60 m² (with west balcony); (see Ochs et al for details) The east wall is adjacent to another multi-family house, see Figure 1. Some refurbishment measures have been implemented so far. In the 80 s the façade (including the outside basement wall) was insulated with 5 cm external insulation. The timber roof, the ceiling of the attic storey and the basement floor are not insulated. The heating demand of the building is calculated with the PHPP with approx. 125 kwh/(m² a) assuming heated staircase and cellar (A T = m²) or 180 kwh/(m² a) assuming unheated staircase and cellar, see also section Simulation results, below). AF FF GF AF: attic floor FF: first floor, GF: ground floor, CF: cellar floor CF FIG 5. (left) South façade of the multi-family house, which adjacent neighbouring building, (right) section with prefabricated timber frame façade; 5. Feasibility of the Concept - Simulation Study 5.1 Simulation Model TRNSYS is used to simulate the performance of the building and the heat pump. Since TRNSYS 17 a new plug-in is available that allows importing data from a 3D drawing, see FIG 6 for the Sketchupmodel of the Lubu demo building with five zones. attic storey (AS) staircas e Cellar Floor (CF) Ground Floor (GF) First Floor (FF) FIG 6. 3D Sketchup-Model of the Lubu demo building with the five zones (CF, GF, FF, AF and staircase) Full papers - NSB 2014 page 1296

127 5.2 Simulation Results The heating demand can be reduced from 125 kwh/(m² a) (related to the treated area incl. staircase and cellar) to about 22 kwh/(m² a) using a prefabricated timber frame façade element, perimeter insulation, 3 pane windows in PH quality and a MVHR, see TABLE 1. Renovation to 15 kwh/(m² a) is also possible from the technical point of view if enhanced components are used and ground insulation is applied but is (here) not recommended from the economic point of view. For the detailed renovation solution (parametric study of different envelope solutions) incl. details of the timber frame façade and U-values of walls and windows, etc. see Ochs et al REMARK: Here, the staircase is assumed to be heated to the same temperature as the four flats, i.e. to 20 C. The specific heating demand of the existing building is 180 kwh/(m² a) if the staircase and cellar are considered to be unheated. (There is also an absolute difference as a result of differences in geometry (treated area and external surface: cellar, staircase external wall, basement and roof) and boundary conditions (staircase and neighbour temperature, i.e. = 4 K). TABLE 1. Specific heating demand (HD) of the demo building before and after the renovation simulated with the five zone TRNSYS model for the four flats and heated staircase and in comparison with the PHPP results (total heating demand); reference area is treated area Zone Total CF GF FF AT ST Area / [m²] existing building TRNSYS m² PHPP renovated building TRNSYS m² PHPP 20.6 In the renovated case, if the staircase is assumed to be heated the heating demand is about 1500 kwh/a higher than in the case of unheated staircase. The heat load is 500 W higher (maximum of daily average) as can be seen in TABLE 2. With additional controlled shading and night ventilation (for overheating protection) and summer-bypass of the MVHR the heating demand and the heat load increases slightly. TABLE 2. Absolut and specific heating demand (HD) and heat load (HL) simulated with the TRNSYS 5 zone model for different boundary conditions; reference area is treated area Area / [m²] HD / [kwh/a] spec. HD / [kwh/(m² a)] HL / [W] spec. HL / [W/m²] TRNSYS 1Z TRNSYS 5Z *) TRNSYS 5Z fl. adia. #) TRNSYS 5Z fl. perio. +) TRNSYS 5Z fl. perio. summer $) *) heated staircase, #) floating staircase temperature, adiabatic neighbour +) floating staircase temperature and floating periodic temperature of neighbour staircase; $) summer over-heating with protection by means of controlled shading, night ventilation and summer-bypass 5.3 Sensitivity Analysis For the design of the µhp with a heating capacity of 1 kw the heat load of the flats is the important design criterion. To investigate the robustness of such as system a sensitivity analysis is performed. Following aspects are investigated: 1. Boundary conditions staircase and neighbour 2. Non-heated neighbour flat Full papers - NSB 2014 page 1297

128 3. Additional window ventilation 4. Different set point temperatures (single flat and all flats) 5. Occupation profile (internal gains and design air change rate) 6. Heat pump system concept (with/without storage) Building and User Behaviour For the reference case the following heating demands (HD) and heat loads (HL) of the flats have been obtained by means of simulation. The boundary conditions have strong influence. The µhp with air heating is suitable for the GF and FF with the HD below 10 kwh/(m² a) (GF has slightly higher heating demand due to the colder cellar). For the CF a µhp concept with a hydronic heat distribution system would be possible. In the AF with a heat load of about 2 kw the µhp concept is hardly feasible. If one of both neighbouring flats is not heated (the entire winter, e.g. due to long term absence of a tenant) the heat load increases from 880 W to 1120 W. With an additional backup heater the comfort level can be maintained. The increase of the HL with increasing set point temperatures is shown in FIG. 7. If all flats have higher set point, the increase of the HL is less significant. Additional window ventilation (every night! not occasional window ventilation) in the sleeping room would increase the heat load by a factor of 3. TABLE 3. spec. heating demand HD in kwh/(m² a) and heat load HL in W/m² in brackets (TRNSYS 5Z floating staircase, periodic neighbour, summer overheating protection) Area / TRNSYS 5Z TRNSYS 5Z TRNSYS 5Z fl. TRNSYS 5 Z PHPP [m²] fl. adia. #) perio. +) (TABLE 1, PHPP) CF (15.8) 54.8 (17.9) 54.8 (18.1) (18.4) GF (5.4) 8.3 (8.4) 9.0 (8.8) (8.8) FF (5.2) 5.1 (6.7) 5.4 (6.8) (8.8) AT (15.3) 37.4 (21.3) 38.9 (21.8) (22.2) FIG 7. (left) heat load of GF for heated ( CF on FF on ) and unheated neighbour flats (heating off either in CF or FF or in both) and (right) heat load for different set point temperatures System Concept and Sizing In the previous section simulation results for heating demand and heat load apply for the case of ideal heating. In a further step a simulation study using ambient air to water µhp is performed. Different heating systems concepts (with and without storage) and different sizing of heat pump and buffer storage are investigated. Radiators are used with different flow temperatures in case of the system with storage. The heat pump used for the simulation is scaled to match the required power (with constant performance map, COP(35/2) = 3.9, see Konz 2013 for further details). The system with storage is shown in FIG 8 and the one without in FIG 2. Full papers - NSB 2014 page 1298

129 REMARK: As experience shows, small heat pumps are less efficient than large heat pumps for several reasons (e.g. thermal losses, efficiency of compressor, dead volume, parasitic energies). Hence, the performance of larger (i.e. up-scaled) heat pumps is supposed to be slightly higher than calculated and that of the down-scaled heat pumps slightly lower. HP BH pump BS TI circulation pump FH or radiator TI FH: floor heating BH: backup heater HP: Heat Pump BS: Buffer Storage control control TI FIG 8. Heating system with µhp with buffer storage heat pump system FIG 9. (left) SPF as a function of the storage size and (right) SPF as a function of the heat pump size With sufficient storage (> 50 l) the SPF of the heat pump can be slightly increased, however the SPF of the system is significantly reduced due to storage losses and parasitic energies. The SPF without storage is significantly higher. A System SPF of about 2.4 can be obtained. The size of the heat pump (here for the case of a constant speed compressor) is important for a good performance. If the heat pump is over-dimensioned the SPF decreases significantly. The heat load of the building is influenced by the occupation profile on the one hand via the internal gains and on the other hand via the ventilation rate (here, 30 m³/h/person). In addition the ventilation rate limits the maximum heating capacity according to equation 1 in case of air heating. With Q n c V ( max room ) (1) Q heat flow (W), n air change rate (1/h), density, 1.24 (kg/m³), c specific heat capacity, 1004 J/(kg K), V Volume (m³), temperature ( C) With a maximum temperature of 55 C and a room temperature of 20 C the maximum heating capacities shown in TABLE 4 are obtained and compared to the simulated heat load of the GF flat. For a normal occupation of 2 adults and 2 children (2A2C) or 2 adults and 1 child (2A1C) the maximum heating capacity is sufficient. In case of lower occupation level there is a mismatch between heat load and heating capacity which has to be considered during the design. An additional backup heater (e.g. in the bathroom and/or corridor) should be considered for comfort reasons anyway. Full papers - NSB 2014 page 1299

130 TABLE 4. simulated spec. heat load HL and max. heating capacity in case of air heating (acc. equ. 1) Occupation Ventilation rate / [m³/h] HL / [W] max. heating capacity / [W] 2A2C A1C A REMARK: in a very efficient house a slightly under-dimensioned heating system does not lead to a significant temperature drop due to the high inertia (time constant) of the building. Temperatures will hardly drop below 19 C (Bisanz 1998). 6. Conclusions and Outlook It is shown by means of an extensive simulation study that the system concept of a façade integrated µhp is feasible for renovated buildings (EnerPHit standard or better). Reasonable performance with high economic efficiency can be obtained. A more detailed physical heat pump model is under development considering air heating and speed controlled compressor. A multi-zone model of the GF flat has been established and will be used for future investigations and optimization. A prototype with façade integrated MVHR has been developed and will be measured in the lab at UIBK. A second prototype with an exhaust air-to-air heat pump in combination with MVHR will be developed in the near future and optimized for the demo building in Ludwigsburg, where two of the four flats will be renovated with the proposed system in late Monitoring of the performance will be conducted for a period of at least one year. 7. Acknowledgements These results are part of the research and simulation work of the European project inspire funded under the 7th Framework Program (Proposal number: , title: Development of Systematic Packages for Deep Energy Renovation of Residential and Tertiary Buildings including Envelope and Systems, duration: ). References Bisanz, C., Heizlastauslegung im Niedrigenergie- und Passivhaus, 1. Auflage, Darmstadt, Januar 1999 Feist, W. (editor), EnerPHit Planerhandbuch - Altbauten mit Passivhaus Komponenten fit für die Zukunft machen. Autoren: Zeno Bastian, Wolfgang Feist Konz A., Thermische Simulation von Sanierungsvarianten eines Mehrfamilienhauses mit Fokus auf den Einsatz von Wärmepumpen in zentraler und dezentraler Gebäudetechnik, Master Thesis, TUM, UIBK, 2013 Ochs F., Dermentzis G., Siegele D., Konz. A., Feist W Use of Building Simulation Tools for Renovation Strategies - a renovation case study, EnergyForum 2013, Bressanone Ochs F., Dermentzis G., Siegele D., Konz. A., Loose A., Drück H., Feist W. 2014, Thermodynamic analysis of ground coupled heat pumps with solar thermal regeneration, IEA Heat Pump Conference, Montreal 2014 (abstract accepted). Siegele D., Modellierung und Simulation Fassadenintegrierter Aktiver Komponenten, Master Thesis, UIBK, 2013 PASSYS, The PASSYS Test Cells: A Common European Outdoor Test Facility for Thermal and Solar Building Research, Commission of the European Communities, Directorate-General XII for Science, Research and Development, Edited by BBRI - Brussels 1990 Full papers - NSB 2014 page 1300

131 Full paper no: 162 The influence from input data provided by the user on calculated energy savings. Jimmy Vesterberg, M.Sc 1,2, Staffan Andersson, Ph.D 1, Thomas Olofsson, Ph.D 1 1 Applied Physics and Electronics, Umeå University, Sweden 2 Industrial Doctoral School, Umeå University, Sweden KEYWORDS: Simulation, energy savings, retrofit, model calibration, regression SUMMARY: It is generally accepted that the most correct decisions are made when the used support system provides the most accurate description of the starting point as possible. That is, in this case, a detailed initial description of a building, planned to be refurbished and evaluated with the building energy simulation software IDA ICE (v 4.5). In order to assess this statement, we have used two different models to predict energy savings due to different planned energy conservation measures (ECMs): - A basic model based on inputs from currently available standards and as-built drawings. - A calibrated model based on an analysis of measurements from two months, together with measured air handling unit parameters, hourly electricity usage and indoor temperatures. The relative prediction differences between the models are investigated as well as compared with the actual outcome in a neighboring building where the analyzed ECMs have been implemented. The result indicates that a calibrated model should be used, in order to accurately determine the postretrofit energy demand. However, if only investigation of ECMs which aims to decrease a buildings transmission loss is of interest, the findings suggest that BES calibration is of minor importance. 1. Introduction To stimulate investments in different energy conservation measures (ECMs) in existing buildings it is important that these measures meet the expected or by the supplier promised performance. This implies an increasing demand for reliable evaluation methods and accurate input parameters to simulation models, since the question of liability is expected to become more frequent. Typically, implementation of ECMs in existing buildings is preceded by calculations with whole Building Energy Simulation (BES) models to predict the expected savings. Often these BES models are based on standardized building and user templates, even though it is generally accepted that the most reliable analysis of retrofit options are made with calibrated BES models (Reddy, 2006; Zhen, 2013; Heo, Choudhary and Augenbroe 2012; Raftery, Keane and O Donnell, 2011). A hindering factor for the widespread use of calibrated simulations is that it typically demands considerable more resources than a typical design stage simulation. This study was designed to investigate discrepancies between outputs from uncalibrated (standardized) and calibrated BES models, when they are used to calculate energy savings. For this purpose, we have used two different models: The first is based on inputs from currently available standards and as-built drawings (basic model). The second, calibrated model is based on extracted thermal performance parameters from an analysis of measurements from two months as well as measured air handling unit performance parameters, hourly electricity usage and indoor temperatures. Full papers - NSB 2014 page 1301

132 The calibrated model comfortably passes the calibration acceptance criteria s for total energy demand for space heating suggested in ASHRAE guideline 14 (ASHRAE, 2002). In addition to investigating prediction differences between the models, the plausibility of the models predicted post-retrofit energy demand are analyzed through a comparison with the actual outcome in a neighboring building in which the modeled ECMs already have been implemented (with the only difference of a slightly larger addition of an attic room). A verification, possible due to identical building designs as well as very similar energy characteristics, pre-retrofit. The used BES tool in this study is IDA Indoor Climate and Energy (IDA ICE) (EQUA, 2013) which has been used in numerous previous studies e.g. (Pavlovas, 2004; Molin, Rohdin and Moshfegh 2011; Salvalai, 2012; Hesaraki and Holmberg, 2013; Loutzenhiser et al, 2009) 1.1 Buildings and considered ECMs The studied building is a ten apartment multifamily building, constructed during the years 1970/71 for the municipal housing company, AB Bostaden. The building is at the time of writing, part of a large refurbishment project, that includes 21 multifamily buildings located in the city district Ålidhem, Umeå, Sweden. Due to a relative high demand of District Heating (DH) for space heating, a number of ECMs were considered to reduce this energy demand: (ranked in a most likely implementation order) 1. New attic room for placement of a new air handling unit and retrofitting of roof 2. Ventilation system with heat recovery on the exhaust air 3. Improved roof insulation 4. Window upgrade (from two to three glazed) 5. Adjustment of domestic hot water circulation losses 6. Additional interior insulation of exterior walls 2. Modeling The basic model was defined in IDA ICE based on inputs from current available standards and asbuilt drawings including floor plan, dimensions of different parts of the building envelope and heating and ventilation system (one serving all of the thermal zones). The calibrated model was obtained by changing different input data in the basic model to coincide with a performed analysis of measurements. 2.1 Weather file The used weather file contained, an onsite measured outdoor dry-bulb temperature and relative humidity. The global solar radiation was measured at Umeå University weather station (Applied Physics and Electronics, 2013) approximately 1 km from the studied buildings. However no wind measurements were available, instead typical mean year data for wind speed and direction was synthetically generated with the climate software (Meteonorm, 2013). 2.2 Basic model The basic model was heavily based on standardized building and user templates to mimic the situation when detailed measured data does not exists or given time constraints do not allow a typical labor intensive calibration process. Input, regarding: household electricity usage, indoor temperature, window airing, internal shading, heat gain from the occupants and electric heat gain factor were estimated with a widely used guideline in Sweden (SVEBY, 2012). This source was further used to distribute the household electricity usage in the kitchens, living rooms, bathrooms, bedrooms and halls to 51%, 27% and 3 7% of the total household electricity Full papers - NSB 2014 page 1302

133 load, similar to the assumed distribution in a previous study in Sweden (Molin, 2011). The air rates to and from the air handling unit as well as the air leakage were estimated based on the Swedish building regulations (SBN 1975) and (SBN 1980). The most common method in praxis to account for thermal bridges in Sweden has been found to utilize standardized values in the literature (Berggren, 2013). Thus, in this study, the heat loss due to thermal bridges was estimated with standard linear thermal transmittance values, tabulated in the ISO standard (ISO 14683, 2007). This approach applied on the investigated building resulted in a ratio of thermal bridges to total transmission losses of 25%.The definition of U-values were thereafter completely defined with program defaults complemented by design values in literature (Petterson, 2009). In contrast to the above standardized values, the annual measured building electricity was used as input, due to the wide accessibility of this information in Sweden, since it is regulated (SFS 2006:985) to be measured for new buildings. For older buildings as in this study, it was gathered from utility bills. Based on performed site surveys, 85% of the total measured building electricity usage was assigned to the ventilation room (where the majority of the equipment was installed). The remaining 15% was evenly distributed between the stairwells in the model to cover the power need of a few remaining lights. The internal heat gain due to occupancy was estimated from public records and guidelines given in (SVEBY, 2012). This heat gain was modeled as constant heat contributions and evenly distributed over the living areas. An occupancy schedule could have been used but as these heat contributions are small relative to the total heat demand, the increase in model accuracy was assessed to be negligible. A zone strategy that follows the guidelines given in (ASHRAE, 2007) was used to agglomerate actual zones in the building of the same zone type into a single zone, if the actual thermal zones in the building varied less than 45 degrees in orientation from each other. This zone approach led to 21 zones in the final model of which a 3D façade view is shown in fig 1. FIG 1. 3D façade view of the analysed building defined in IDA ICE software, where the location and geometrical model correspond to the real situation before the planned refurbishment. Further model preparation concerned the buildings mechanical supply and exhaust ventilation system with constant air change rate. It was simulated accordingly, as a constant air volume system and with a heating coil in the air handling unit. The heating coil was activated when the supply air was below the set, constant supply temperature. In addition, the radiators were modeled as ideal heaters in the IDA ICE environment, placed in respective zones with a sufficient capacity to sustain the set indoor temperature, during all conditions. Full papers - NSB 2014 page 1303

134 The interior walls were treated as adiabatic, the inside air was allowed to move between the zones in each floor through interior door openings and external infiltration was distributed over the zones according to external surface area. Finally, the heat transfer to the ground was simulated according the standard (EN ISO 13370, 2007) and as the building lacks a air conditioning system, no cooling energy was considered in the simulations. However, when the indoor temperature was above 25 o C, the occupants were assumed to increase the ventilation by airing. The use of the above input values resulted in a simulated energy demand (electricity and DH for space heating) of MWh/yr, this is an overestimation by approximately 19.1% compared to the measured data of MWh/yr. Discrepancies of this magnitude have been reported in other studies (Pedrini, Westphal and Lamberts 2002; Ahmad and Culp, 2006; Danielski, 2012) and are not unusual for BES models heavily based on standardized input values. 2.3 Calibrated model The calibrated model was developed from the basic model by adjusting the previous assumed thermal performance parameters to coincide with a regression analysis of measured data from two months, when the solar gain was the smallest, i.e. a thorough analysis based on the energy signature principles was conducted (pre-retrofit) with a method developed in (Andersson et al, 2011). In addition, the basic model was complemented with hourly measured building and household electricity usage as well as air handling unit performance parameters and temperature data. To summarize, the revised input parameters were: The actual electricity use was loaded into the program at an hourly time step as recommended in (Raftery, Keane and O Donnell, 2011) for the complete year of simulation. The dynamic influence of the indoor temperature was considered. IDA ICE calculation of ground heat loss was adjusted to coincide with the results from a regression analysis of measured data. The buildings loss factor (transmission + leakage) was adjusted to coincide with the results from the regression analysis. Measured air handling unit performance parameters (air rate, supply temperature, efficiency) was loaded into the program as average values during the heating season (Oct-Mar). The use of the above building unique input values resulted in a simulated energy demand (electricity and DH for space heating) of MWh/yr. This corresponds to a modest underestimation (MBE) of -3.7% and a Coefficient of Variation of the root mean squared error (CV) of 5.7% calculated with monthly data. The suggested calibration tolerance limits suggested in (ASHRAE, 2002) is a MBE within ± 5% and CV less than 15% relative to monthly data. Thus the model passes these thresholds comfortably. 2.4 Model predictions The predictions from the basic and the calibrated models were investigated using the six previous mentioned individual ECMs as well as three different sets of ECMs: Set1, Composed of ECM 1 and 2 Set2, Composed of ECM 3, 4 and 6 Set3, All individual measures implemented. The individual and sets of ECMs were evaluated separately so that a decision could be made for each category. This enables the building manager to choose to implement only one ECM at a time or select a group of measures. The energy savings were estimated by adding each individual ECM to the models and the predicted energy demand (Ê ECM,i ) was compared with the forecast of the previous Full papers - NSB 2014 page 1304

135 prediction (Ê ECM,i-1 ). This adding of measures continued until all measures were included and thus represented the post-retrofit situation. The predicted percentage savings for both models was calculated, according to: % E save, i Eˆ = 100 ECM, i 1 E Eˆ measured, DH ECM, i Where E measured DH is the measured DH demand for space heating (radiators and air handling unit), during the analyzed year, pre-retrofit. Eq. 1 was further used to analyze model differences for the sets of ECMs, then with Ê ECM,i-1 substituted against the predicted model energy demand in the present situation (i.e. held fixed) and Ê ECM,i then corresponded to the model forecast for the ith set of ECMs. The implementation of an attic room (ECM 1) is necessary for installment of a new air handling unit and would result in increased energy demand. However, the focus in this study is to investigate prediction differences of simulation models when subject to input parameter changes. In that analysis the direction of the energy change is of minor importance. In table 1, the shared baseline values for both models (basic and calibrated) and the adjusted values in the post-retrofitted models are shown. For consistency, ECM 1 been excluded in table 1 due to the fact that the basic and calibrated models have different overall UA-values. The overall UA-values increased in the basic model from WK -1 to WK -1 and from WK -1 to WK -1 in the calibrated model respectively due to the addition of ECM 1 in the models. The reason for the different changes in the two models overall UA-values originates from that the models assumes different settings of thermal bridges which is also assumed to hold for the installation of ECM 1. TABLE 1. Shared baseline values in the used simulation models and the new (modified) values to simulate the ECMs considered. Design Affected building Pre-retrofit value Post-retrofit value alternative components ECM 2 Air handling unit efficiency 0% 85% ECM 3 U-value roof 0.29 WK -1 m WK -1 m -2 ECM 4 U-value window/shgc* 2.2 WK -1 m -2 /73% 1.1 WK -1 m -2 /55% ECM 5 Heat losses from dhwc** 31 kwh/m kwh/m 2 ECM 6 Heated area/u-value walls 880 m 2 /0.32 WK -1 m m 2 /0.24 WK -1 m -2 *Solar Heat Gain Coefficient. **Constant domestic hot water circulation (dhwc) losses modified to standardized level according to (SVEBY, 2012). 3. Results The predicted savings for both models are shown in fig 2 for all individual as well as for the different sets of ECMs. The smallest discrepancies between the models are seen for the ECMs which affects the buildings transmission losses i.e. individual ECMs (1,3,4,6 max deviation 1.8%) as well as for the combined measure Set2, with a 2.6% difference between the two model predictions. The difference is significant, 14.2% regarding the saving potential for the implementation of heat recovery in the ventilation system (ECM 2). This is mainly due to that different air rates are assumed in the models. In the basic model a higher standardized air rate is assumed compared to the calibrated model, which uses the measured air rate, hence a larger saving potential exists in the basic model for this ECM. This creates subsequent errors by the basic model in the combined measure sets, Set1and Set3 with a 12.4% and 12.2% deviation in comparison with the calibrated model. (1) Full papers - NSB 2014 page 1305

136 Fig 2.Annual predicted percentage savings with the two BES models due to individual as well as sets of ECMs. Minor discrepancies are also seen between the models in the predicted change in energy use associated with the adjustment of the buildings dhwc losses, (ECM 5) of 1.6%. This energy saving potential depends on the length of the heating season in the models. The basic model has a longer heating season and can therefore utilize a larger part of the constant dhwc losses compared to the calibrated model. A reduction of the dhwc loss in the basic model must then be replaced by a larger part of the temperature dependent heat load, which results in a smaller predicted overall heat saving, compared to the calibrated model. 3.1 Analyze of predicted post-retrofit energy demand. Measured post-retrofit energy use, for an entire year, does not yet exist for the studied building. However, the reasonableness of some forecasts can be estimated by comparing the predicted postretrofit energy demand with the actual outcome in a neighboring building, in which the simulated ECMs already have been implemented, with the only difference of a slightly larger attic room. The comparison is possible due to identical building designs as well as heating and ventilation systems, before the refurbishment. The similarities were also confirmed through a comparison of the buildings measured heat demand, during a couple of months prior to the refurbishment. Before the model outputs were compared with the neighboring building, the measured input parameters previously used in the calibrated model were changed to the measured operating conditions in the retrofitted neighboring building. (No modifying was done in the basic model as it assumes standardized operating conditions). This resulted in deviations of the magnitude of +17% for the basic model and less than -1% for the calibrated model in predicted post-retrofit demand of total energy demand for space heating (electricity and DH demand). An additional test was done for the implementation of ECM 2, by comparing the measured supplied DH to the air handling unit in the neighboring building with the same in the studied building, preretrofit, the difference was measured to 26.8 MWh/yr. Thus this magnitude of energy savings is what is reasonable to be expected due to installment of ECM2. The basic model predicted an energy saving of 40.5 MWh/yr, that is +51% more than the comparative value of 26.8 MWh/yr, whereas the calibrated model estimated 28.2 MWh/yr which corresponds to a much more moderate discrepancy of +5%. Full papers - NSB 2014 page 1306

137 4. Conclusions This study was designed to investigate the necessity of using calibrated simulation models in order to get reliable energy saving predictions. We focused on a few selected ECMs implemented in one studied building. Based on the observed agreement between the predicted post-retrofit energy demand with the calibrated model and the actual outcome in the neighboring building, it is indicated that a calibrated model should be used in order to accurately predict the post-retrofit energy demand. In addition, the calibrated model yielded accurate forecast of the energy saving due to heat recovery of the exhaust air. However, in the calculations of the ECMs which affected the transmission losses, small differences were found between the basic and calibrated model. This indicates that BES models based on standardized input parameters can be used to predict energy savings for that purpose. The main conclusion is that, accurate predictions of energy savings can be obtained with standardized BES models, complemented by measured air handling unit performance parameters. However, in order to accurately predict the post-retrofit energy demand, the starting point of the used BES model is crucial, i.e. in such situations the input parameters needs to be thoroughly calibrated. 5. Acknowledgements The project is funded by The Industrial Doctoral School at Umeå University and AB Bostaden, Umeå, Sweden. The authors also thank Umeå Energi for their assistance with data collection. References Ahmad M. & Culp C.H Uncalibrated building energy simulation modeling. HVAC & R Research, 12, Andersson S. et al, Building performance based on measured data. World Renewable Energy Congress Sweden, 8-13 May, 2011, Linköping, ASHRAE ASHRAE Guideline Measurement of Energy and Demand Savings. Atlanta: American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc. ASHRAE ANSI/ ASHRAE Standard Energy Standard for Buildings Except Low-Rise Residential Buildings. Atlanta: American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc. Applied Physics and Electronics. [ ] Berggren B. & Wall M Calculation of thermal bridges in (Nordic) building envelopes Risk of performance failure due to inconsistent use of methodology. Energy and Buildings, Danielski, I Large variations in specific final energy use in Swedish apartment buildings: Causes and solutions. Energy and Buildings. 49, EN ISO Thermal performance of buildings-heat transfer via the ground-calculation methods. EQUA. [ ] Meteonorm. [ ] Full papers - NSB 2014 page 1307

138 Heo Y, Choudhary R. & Augenbroe G.A. (2012). Calibration of building energy models for retrofit analysis under uncertainty. Energy and Buildings. 47, Hesaraki A. & Holmberg S. (2013). Energy performance of low temperature heating systems in five new-built Swedish dwellings: A case study using simulations and on-site measurements. Building and Environment. 64, ISO Thermal bridges in building construction - Linearthermal transmittance - Simplified methods and default values. Molin A., Rohdin P. & Moshfegh B Investigation of energy performance of newly built lowenergy buildings in Sweden. Energy and Buildings. 43(10), Loutzenhiser P. et al, An empirical validation of window solar gain models and the associated interactions. International Journal of Thermal Sciences. 48(1), Pavlovas V Demand controlled ventilation: A case study for existing Swedish multifamily buildings. Energy and Buildings. 36(10), Pedrini A., Westphal F.S & Lamberts R A methodology for building energy modelling and calibration in warm climates, Building and Environment. 37(8 9), Petterson, B.Å TILLÄMPAD BYGGNADSFYSIK. Lund, Studentlitteratur Raftery P., Keane M & O Donnell J Calibrating whole building energy models: An evidencebased methodology. Energy and Buildings. 43(9), Reddy, T. A Literature review on calibration of building energy simulation programs: uses, problems, procedures, uncertainty and tools. ASHRAE Transactions, Salvalai, G Implementation and validation of simplified heat pump model in IDA-ICE energy simulation environment. Energy and Buildings. 49, SBN Statens Planverk. Svensk Byggnorm, Supplement No 1. SBN Statens Planverk. Svensk Byggnorm, Utgåva 2. SFS 2006:985. Lag om energideklaration för byggnader, Näringsdepartementet, Stockholm SVEBY Brukarindata för energiberäkningar i bostäder. The Swedish National Board of Housing Building and Planning Så mår våra hus-redovisning av regeringsuppdrag beträffande byggnaders tekniska utformning m.m. Tian, Z. & Love J.A Energy performance optimization of radiant slab cooling using building simulation and field measurements. Energy and Buildings. 41(3), Full papers - NSB 2014 page 1308

139 Full paper no: 163 An approach for holistic energy retrofitting based on assessment of economic viability and durability of energy saving measures Martin Morelli, Ph.D. Department of Construction and Health, Danish Building Research Institute, Aalborg University Copenhagen, Denmark KEYWORDS: Cost of conserved energy, risk assessment, FMEA, wooden beam, windows, full-scale experiment, measurement SUMMARY The majority of renovation projects are driven by the possibility of reducing energy consumption of buildings. This, however, might result in retrofitting projects that neglects the longevity of the building. Furthermore, many evaluation techniques only consider the profitability of the energy saving measures and forget to consider, whether it is more prudent to demolished the building and erect a new building. An evaluation approach is presented to assess whether to retrofit an existing building or to demolish and replace it. The primary concept of the method is to develop a retrofitting proposal with a profitably combination of energy saving measures. The cost of the combination of energy saving measures is evaluated against the cost of demolishing the existing building and erecting a new building including consideration of maintenance costs and operational costs. The energy price is used as constraint to determine the amount of building retrofitting for implementation. The approach includes also durability assessments of the energy saving measures. An example is carried out to illustrate the application of the approach. The example highlights the importance of including risk assessment and durability evaluation of the energy saving measures when performing holistic energy retrofitting of buildings. 1. Introduction In recent years, major focus is addressed to building renovation given that new buildings add at most 1% a year to the existing building stock. The stimulus for carry out building renovation is reducing the energy consumption of buildings. In Denmark the government has adopted a long-term policy, implying that Denmark should be independent of fossil fuels by 2050, and by 2035 energy supply to buildings should be from renewable energy sources (Danish Government, 2011). To meet this objective, it is of significance to improve the energy efficiency of the existing building stock, but also to invest in and convert the supply network to renewable energy sources. Ideally, a balance must be found between the costs for improving energy efficiency of the existing building stock and the costs of buying energy from heating and power plants based on renewable energy sources. Several approaches exist for optimisation of building renovation, where the commonly used economic techniques are simple payback time and net present value (NPV) (Verbeeck and Hens, 2005; Tommerup and Svendsen, 2006). Both techniques, as well as their limitations, are described by Martinaitis et al. (2004). A method derived from NPV is cost of conserved energy (CCE), which gives the cost to save 1 kwh of energy and is directly comparable to the cost of supplied energy. This makes the CCE technique more transparent and practicable for understanding the profitability of the measures as compared to the monetary result obtained using e.g. the NPV method. The CCE method was applied in building renovation by Martinaitis et al. (2004; 2007) and for design of new buildings by (Petersen and Svendsen, 2012). In common, the methods focus on energy consumption and not the Full papers - NSB 2014 page 1309

140 durability of the energy saving measures (ESMs). The stimulus for saving energy neglects the fact that the longevity of the building could be challenged due to changed hygro-thermal conditions. Therefore, it is important to include both an assessment of the whole building as well as ESMs in the retrofitting approach. The presented approach determines the viability of various ESMs including an assessment whether to renovate the existing building or to replace it with a new building. Furthermore, the approach evaluates the durability of the ESMs. The first part of the method is adapted to building retrofitting from the method presented by Petersen and Svendsen (2012), and the energy price is used as constraint to determine the amount of building retrofitting for implementation. The durability of the measures is evaluated based on hygro-thermal measurements and experiences in a test apartment of a multi-family building. 2. Approach for holistic energy retrofitting The approach for holistic energy retrofitting is shown in Figure 1, which consider both the profitability of the retrofitting project and the durability of the ESMs. First step was to determine the needed retrofitting and whether the retrofitting should be executed or the building should be demolished and rebuild. Second step was to investigate the ESMs regarding their durability. FIG. 1. Holistic energy retrofitting from whole building to energy saving measure (ESM) 2.1 Whole building The profitability of the whole building retrofitting is based on the cost of conserved energy (CCE) concept described by Morelli et al. (2014) and in the following bullets 1-3. This concept allows for a decision-making on whether to invest in ESMs or buying energy, as the CCE results are directly comparable with the energy price. 1. Assessment of ESM and determining the inter-relationship between the CCE R ( /kwh), which is the marginal CCE for the different measures. From Eq. (1) the energy use can be expressed as a function of the CCE R which enable a direct comparison of the ESMs. t a( nr, d ) Imeasure + ΔM year + ΔEoperation, year Penergy type CCER = (1) ΔE year where, t is a reference period that enables a comparison of measures with different service life and is defined as the ratio between the reference period, n r (years), and the useful lifetime, n u (years); a(n r, d) is the capital recovery rate, for which d is the real interest rate (absolute number); I measure is the marginal investment cost ( ), where a(n r, d)* I is the marginal Full papers - NSB 2014 page 1310

141 annualised investment cost( ); ΔM year is the change in annual maintenance cost ( ); ΔE operation,year is the annual change in energy consumption during operation of the measure (kwh); P energy type is the energy price for the energy type used for operational energy ( /kwh); ΔE year is the annual change in annual energy conserved by the measure (kwh). This, however, is easily applicable for continuous ESMs e.g. insulation materials, but not for discrete measures e.g. windows and ventilation. Therefore, a five step algorithm was formulated to rank the discrete ESMs (Morelli et al., 2014). A. A first reference is determined among a number of components based on their investment cost and annual energy use. The components are ranked according to investment cost, and the component with lowest cost is chosen as reference. If the investment cost is identical for two or more components, the component with lowest energy use should be chosen as reference, and the other components should be omitted due to the higher energy use. For existing components the investment cost will be the refurbishment cost, thus the component performs as when it was newly installed. B. The marginal CCE R for each component is calculated applying Eq. (1) using the reference component determined in step A. Components with negative values of CCE R are omitted because they use more energy combined with a higher investment than the reference component determined in step A. C. A new reference is determined based on the marginal CCE R derived in step B. The component with the smallest positive marginal CCE R is chosen as a new reference to form a curve. From the remaining components, those with an energy use equal or higher than the new reference are omitted as they are not ESMs compared to the new reference D. The marginal CCE R for each component is calculated applying Eq. (1) using the reference component determined in step C and its respective investment cost and energy savings. Step C and D are repeated until there are no more components to consider. E. The reference component found in step A and those determined in step C are listed in the order they are determined. These discrete components are thereby transformed in to a continuous CCE R function by calculating the marginal CCE R according to Eq. (1). 2. The method suggest that the determination of a combination of ESMs is defined as the energy weighted average marginal CCE R of the measures (CCE R,average [ /kwh]) equal to the energy price, Eq. (2). First the discrete measures must be chosen as close to the energy price as possible, and thereafter adjusting the CCE R,average by choosing the continuous measures, thus the CCE R,average equals the energy price. CCE R, average ΔE1 CCE = R,1 + ΔE CCE 2 n 1 R,2 ΔE i ΔE CCE n R, n P energy type where, E n is the energy consumption for the ESM n (kwh); CCE R,n is the cost of conserved energy for the ESM n ( /kwh);and, n 1 ΔEi is the sum of the individual energy consumptions of all ESMs (kwh). 3. Calculation of the economic project profit considering whether to renovate the building or replace it with a new building. The profit of any given project is determined on the basis of the market value (MV) for the retrofitted building, or a newly erected building, minus the investment cost (I),which can also including the transaction costs, and the discounted (1/capital recovery rate) maintenance and operational (M&O) costs, as given in Eq. (3). If a new building is erected at the exact same location as the existing building, the cost for demolishing (D) the existing building must also be included. The building project that should be undertaken in economic terms will be the one having the highest profit. (2) Full papers - NSB 2014 page 1311

142 M & O Profit = MV I + D + a ( n d) r, (3) 2.2 Measures Two types of ESMs were investigated, i.e. 1) two types of interior insulation and 2) four different measures to improve the windows. These measures were investigated in full scale in the test apartment. The risk of failures was identified for an interior insulated masonry-wooden beam assembly applying Failure Mode and Effect Analysis (FMEA) (Stamatis, 2003). The critical points, found in the FMEA, were investigated by full scale measurements of temperature and relative humidity behind the interior insulation and in the wooden beam embedded in the masonry. The four window measures were installed in the test apartment and their energy performance was calculated based on the energy balance (Morelli et al., 2012). 3. Multi-family building a case study A typical building in Copenhagen, Denmark, dated from the period was used as case. The building with 30 apartments had six storeys with a floor to floor height of 2.6 m and a gross floor area at each storey of 453 m 2. The solid masonry facades were deemed worthy of preservation and the windows constituted 27% of the overall façade area. The windows were with a single pane; however, the windows on the street façade had a secondary glazing installed. The un-insulated floor divisions were constructed with wooden beams and clay pugging. The building was natural ventilated by opening windows, infiltration and ventilation ducts located in the kitchens and bathrooms. Furthermore, the building employed central heating produced from district heating. A detailed description of the building is given in (Morelli et al., 2012) where the pre-renovation energy consumption for the building was determined to approx. 160 kwh/(m 2 year). 3.1 Description of energy saving measures Two interior insulation materials were tested. i) A combination of aerogel and stone wool fibres with a gypsum board mounted on the surface, hereafter referred to as MiWo-Aero, which has a thermal conductivity of W/(m2 K). ii) Vacuum insulation panel (VIP) having a thermal conductivity of W/(m 2 K) for a thickness of 20 mm under 1 mbar pressure. The four window measures are briefly described in Table 1 where the numbers of panes refer to the total amount of panes in the window (Morelli et al., 2012). TABLE 1. Energy data for window measures # Window type and retrofit measure U W [W/(m 2 K)] g w [-] E ref [kwh/(m 2 year)] Ref. with 1 layer normal pane Ref. with secondary pane (2 panes) Refit with secondary pane (3 panes) Refit with secondary pane (2 panes) Refit with sash on casement (2 panes) New with coupled frames (3 panes) Whole building assessment For the economical assessment of the whole building retrofitting, a reference period of 30 years was considered, which corresponds to a typical loan period for building investments but also the Full papers - NSB 2014 page 1312

143 approximated service life of many building components. The real interest rate was set to 2.5% and expresses the amount that the nominal interest rate is larger than the rate of inflation. The energy price in 2040 for heat based on renewable energy sources was determined to 0.15 /kwh by Morelli et al. (2014), whereas todays energy price was 0.09 /kwh. The graphs in Figure 2 for interior insulation and window measures are based on Eq. (1). Similar figures were developed for each ESM. Applying Eq. (2) and choosing discrete ESMs before continuous ESMs the amount of ESMs in Table 2 were obtained, which correspond to an energy consumption for the building of about kwh/(m 2 year) depending on the energy price (Morelli et al., 2014). The measures implemented were: demand controlled ventilation, new windows, insulation towards basement and attic as well insulation on the inside and outside of the walls. FIG 2. Energy use as a function of CCE R for a) 2 types of interior insulation and b) windows with single pane and single pane with secondary pane (Morelli et al., 2014) TABLE 2. Amount of energy saving measures in relation to energy price and interest rate of 2.5% Energy price 0.09 /kwh 0.15 /kwh Measure # CCE [ /kwh] Type [-] CCE [ /kwh] Type [-] Mechanical ventilation 0.22 Central DCV 0.22 Central DCV Windows yard 0.06 New 3 panes 0.06 New 3 panes Windows street 0.09 New 3 panes 0.09 New 3 panes Floor to basement mm mm Floor to attic mm mm Wall- interior insul mm mm End wall exterior insul mm mm The evaluation of whether to renovate the building or build a new, according to Eq. (3), was based on presumptions of market values, renovation cost of the building and theoretically energy consumption (Morelli et al., 2014). In connection with the building renovation new bathrooms and kitchens would be installed, which was included in the investment cost of the renovation given in Table 3. The retrofitted building was compared to a new building fulfilling the Danish Building Regulations in 2015 corresponding to a total energy consumption of approx. 30 kwh/(m 2 year). TABLE 3. Economic evaluation of retrofitted and new building Energy price 0.09 /kwh 0.15 /kwh Refit [ /m 2 ] New [ /m 2 ] Refit [ /m 2 ] New [ /m 2 ] Market value Full papers - NSB 2014 page 1313

144 Expenses: Investment Demolish Heat Electricity Profit ± Experiences with interior insulation from a test apartment The FMEA conducted on the interior insulated masonry and wooden beam assembly focused on the three main structural parts; masonry, wooden beam, and insulation including vapour barrier. The results were as given below in prioritised order (Morelli & Svendsen, 2013). 1. Collapse of the wooden beam due to moisture penetration into the structure. 2. Loss of adherence between the brick and mortar. 3. Mould growth behind the interior insulation. The first and last failures were investigated by measurements of temperature and relative humidity in the beam and behind the insulation. Figure 3 shows the results for insulation installed to a northeast facing wall and placed between the ground and first floor. After dismantling the MiWo-Aero product on first floor no visible signs of mould growth were present, which was documented through Mycrometer surface tests (measurements on bio-mass from a swap). FIG 3. Temperature and relative humidity a) behind inside insulation and indoor relative humidity; and b) in the wooden beam and temperature in the exterior climate (Morelli et al., 2012) The MiWo-Aero product was reasonably easy to work with, but the product could not take up any deviations on the surface of the wall. Consequently, the preparation of the wall ensuring a relatively smooth surface was very important, thus the applied insulation likewise provided an even, flat surface. However, the mineral fibers in the MiWo-Aero product did not have enough adhesion to keep the gypsum board fixed to the surface of the insulation resulting in a gap between insulation and gypsum board. In comparison to the MiWo-Aero product, the VIP product was a challenge to work with. Specifically, the VIP product needed to be ordered in specific sizes, given that no on site changes could be made, if incorrect sizes were delivered. Furthermore, the VIP product needed special care as the VIP panels were easily punctured; thereby increasing their thermal transmittance from W/(m2 K) to W/(m2 K). Full papers - NSB 2014 page 1314

145 4. Discussion The economical assessment can be used by decision-makers to determine whether to renovate their building or demolish it and thereafter erect a new building. The approach can be applied at different stages of projects, and for a comparison of different ESMs. The method is easily adjustable to different energy prices, what on the one hand is an advantage of the method, because the energy price is difficult to forecast. On the other hand, this implies that care must be taken using the method, because the energy price strongly influences the optimised combination of ESMs. However, the study showed that doubling the energy price did not change the buildings energy usage significant. This could indicate that an upper limit for implementing ESMs is reached regarding insulation measures. Nevertheless, other measures might further reduce the building s energy consumption, e.g. technical installations excluding mechanical ventilation, as technical installations were not considered in this study. A reduction of approx. 70% in energy consumption is achieved after retrofitting, which is close to the expected requirement for nearly zero energy buildings. The economical assessment, in the twofold holistic energy retrofitting approach, relies on two main assumptions, i.e. maintenance cost and operational costs. A verification of these assumptions is needed for the holistic energy retrofitting method before making the final decision whether to retrofit or replace the building. Installing interior insulation on solid masonry walls with embedded wooden beams, the wooden structure becomes critical parts, due to the outer wall s changed moisture balance. This could cause premature deterioration of the beam. However, the measurements are about 5-10% RH higher as compared to the reference measurement. A RH below 75% does not pose a risk for the durability of the beam end. These measurements were performed in a northeast facing wall that received, given its orientation and location on the building (between ground and first floor), a limited amount of wind driven rain and direct exposure to sunlight. Based on Isopleths information provided in (Sedlbauer, 2002) a risk for mould growth could be present when interpreting the measurements. The temperature is around 10 C and the RH is 85%; in such instances under these conditions the germination time is about 2 days. However, there were no visible signs of mould growth on the wall after dismantling the MiWo-Aero product. In case, the measurements indicated risk for wood rot or mould growth, new retrofit measure should have been suggested, installation of monitoring devices or planning maintenance including economically considerations. These new assumptions should then be included in a second round of calculation according to Figure 1 determining an optimised retrofitting. The two critical points related to mould growth and wood rot reveal, in this study, no expected increase in maintenance or operational cost for the ESM. However, the durability of the MiWo-Aero product itself did not show the expected service life, as the gypsum board and insulation material could not stay fixed. This would increase the maintenance cost, if this insulation material was to be used. Relying on the VIP product a significant cost for planning had to be included, thus the needed sizes of VIP panels are ordered and installed in the building. In worst case, this could lead to many poor assemblies and together with a potential increased thermal transmittance by punctured VIP panels this would result in overestimations of the energy savings or increases in operational cost. Through these two simple examples of insulation materials the significant of the durability is shown when approaching a holistic energy retrofitting in an early project stage. Especially, when considering new materials such as the MiWo-Aero and VIP products. In instances, where either the MiWo-Aero or VIP products are the only materials considered as interior insulation, one could have entailed a significant increase in cost for maintenance, operation or even installation. A more difficult parameter to appraise is the loss of living space due to installed interior insulation, which also should be accounted for in the holistic energy retrofitting approach. 5. Conclusion An approach for holistic energy retrofitting is developed that consider both the economical profitability of the energy saving measures (ESMs), whether to retrofit the building or demolish it and Full papers - NSB 2014 page 1315

146 build a new building, and the durability of the ESMs. The economic assessment integrates methods of component-based optimisation and evaluation of the project economy for building renovation measures. A trade-off between investing in ESMs and buying energy is established entirely on the predicted future renewable energy price. The method uses the marginal cost of conserved energy (CCE) to identify an optimised combination of ESMs having the energy weighted average marginal CCE R equal to the energy price. The profit of the project is determined as the market value deducting the cost for renovation/new building (including demolishing), maintenance and operation. The building with highest profit must be chosen. In cases where replacement of buildings is not an option because of preservation value on facades, not heritage buildings, the method can be used to evaluate, whether it is reasonably to preserve these buildings. Furthermore, the holistic method includes and risk assessment and durability evaluation of the ESMs, thus the energy savings are not the only stimulus for executing the building renovation. The holistic energy retrofitting approach for building renovation developed is highly relevant to and useful for the many future retrofitting projects. References Danish Government Our Future Energy. Danish Ministry of Climate, Energy and Buildings. Available from (accessed ). Martinaitis, V., Kazakevičius, E. & Vitkauskas, A A two-factor method for appraising building renovation and energy efficiency improvement projects. Energy Policy 35, Martinaitis, V., Rogoža, A. & Bikmaniene, I Criterion to evaluate the twofold benefit of the renovation of buildings and their elements. Energy & Buildings 36, 3 8. Morelli, M., Harrestrup, M. & Svendsen, S Method for a component-based economic optimisation in design of whole building renovation versus demolishing and rebuilding. Energy Policy 65, Morelli, M., Rønby, L., Mikkelsen, S.E., Minzari, M.G., Kildemoes, T. & Tommerup, H.M Energy retrofitting of a typical old Danish multi-family building to a nearly-zero energy building based on experiences from a test apartment. Energy & Buildings 54, Morelli, M. & Svendsen, S Investigation of interior post-insulated masonry walls with wooden beam ends. Journal of Building Physics 36(3), Petersen, S. & Svendsen, S Method for component-based economical optimisation for use in design of new low-energy buildings. Renewable Energy 38, Sedlbauer, K Prediction of mould growth by hygrothermal calculation. Journal of Building Physics 25, Stamatis, D.H Failure Mode and Effect Analysis: FMEA From Theory to Execution. 2 nd ed. Milwaukee, WI: ASQ Press. Tommerup, H. & Svendsen, S Energy savings in Danish residential building stock. Energy & Buildings 38, Verbeeck, G. & Hens, H Energy savings in retrofitted dwellings: economically viable? Energy & Buildings 37, Full papers - NSB 2014 page 1316

147 Full paper no: 164 Destructive testing in buildings building performance investigation and comparisons with non-destructive testing Dennis Johansson, Ph.D. 1 Jesper Arfvidsson, Professor 2 Hans Bagge, Ph.D. 2 Lars-Erik Harderup, Ph.D. 2 Johan Stein, Ph.D. student 2 Petter Wallentén, Ph.D. 2 1 Building Services, Lund University, Sweden 2 Building Physics, Lund University, Sweden KEYWORDS: Destructive, measurements, investigation, inspection, health, status, renovation SUMMARY: It is important to investigate building performance and relate it to health and satisfaction of the occupants in order to enable sustainable and healthy buildings. Studies have been made on performance of buildings and relationships between indoor environment and occupants health and satisfaction. However, to really obtain the building performance and construction conditions, such as moisture conditions or material degeneration, it is necessary to look into the construction behind the surface layers, which means that the building must be taken apart. This is usually not possible because it is not allowed or economically reasonable in inhabited dwellings. This paper presents a pilot study of a project that owns the unique possibility of investigating building performance where the buildings will be taken apart to reach behind the surface layers. In this pilot study, the aim was to provide an image of the normal state of normal detached houses after normal aging to assist in the design of energy-saving measures, and to compare with non-destructive testing. Non-destructive and destructive tests were carried out in 14 houses in Malmberget in Sweden. Examples of results are damages found in attics and in wet rooms, but also working houses and good correlation between non-destructive and destructive test results. The pilot study also resulted in a number of useful methods for destructive testing behind the surface layers to enable efficient destructive testing in the main project where 148 apartments will be investigated in Kiruna starting Introduction Much of the housing stock built since the late 1940s, not to forget the Swedish Miljonprogram, where one million apartments were built, is currently in need of major energy improvements and renovations. The design of such actions is today often based on assumptions and estimates from the original drawings and descriptions. The materials and properties of the design is assumed to be as new in absence of better information. In practice, the building does not even need to match the drawings geometrically and materially. Beside possible changes between drawings and actual construction, aging, movements and normal use adds to influence on design thermal performance, airtightness, moisture protection and other building technical performance. The uncertainties are thus embedded in the structure and become even more pronounced over time. Normally it is not considered to be an option to study the current status of renovation projects in detail, which leads to unsafe conditions and hence uncertain results. A good knowledge of the condition of a building is necessary to assure the quality of the final result with redevelopment. At design of the redevelopment, and if there are no Full papers - NSB 2014 page 1317

148 reported or presumed damages, the condition of an assumed "standard building" is highly relevant for the controlled construction process and subsequent operation and maintenance. The condition of the building and its components are essential starting points for energy and moisture design and also for the assessment of estimated remaining life time. Early in the planning phase of renovation it must be analysed and identified whether it is reasonable and profitable to replace certain parts of structures or not. The existing components often have unknown properties due to aging, unknown material data, large variations from manufacturing, unknown version, etc. These factors are very important for calculations of moisture levels and energy use. Interaction of these factors will in turn also affect the renovated building's energy use, indoor environment and life time. Today cases of damaged buildings are inspected and documented. Typical damages are connections between exterior walls and foundation, and exterior walls and windows. Unfortunately, these reports are not easily accessible and also describe damage houses and not the normality of the building stock. Purchase inspections could provide an overall picture of a building's normal state, but they include only inspection of the visible surfaces and are not normally at the hand of the actors of the building process. The Swedish National Board of Housing, Building and Planning has recently presented preliminary results from the BETSI study containing surveys and some measurements, but due to type of the study - no destroying testings. Previous studies of non-damaged buildings have been made on a few detached houses (Örtengren, 1988) but otherwise there is very little available material. There are also studies comparing occupants with building conditions but not with the option to get behind the surface materials (Folkhälsoguiden, 2010; Hägerhed, 2006). 1.1 Prerequisites Ore mining, once founding the cities of Kiruna and Malmberget, is now undermining part of these cities which means that new built areas must be created. A large amounts of buildings will be affected and need to be evacuated and then demolished. Access to some of these buildings has been provided prior to demolition to provide an opportunity to develop methods for studies using destructive testing. In the future, a larger amount of buildings will be studied based on the methods found in this study. The focus is on areas that affect energy efficiency, moisture control and indoor environment and the users of the buildings. The projects was carried out in collaboration with the real estate owner LKAB Fastigheter and Kirunabostäder, which own and run a large number of the buildings to be demolished. 1.2 Aim The overall aim of developing methods that use destructive testing is to reduce uncertainty in the implementation of energy improvement measures in existing buildings with regard to energy, indoor environment and sustainability. Objectives are to reduce uncertainty in the implementation of measures in existing buildings develop better models and input data for efficient and optimized new construction develop methods for efficient and optimized renovation describe relationships between residents' health and experience of their indoor environment and the physical conditions of the building develop methods for non-destructive testing The part of the study reported in this paper aims to develop methods and efficiency for destructive testing and to connect the results to indoor environmental questionnaires in a few, still habited, houses in Malmberget. This will result in a picture of the condition of a number of cross sectional buildings after normal aging Full papers - NSB 2014 page 1318

149 methods for combining destructive testing of buildings with epidemiological surveys of the buildings occupants as a basis for studies on a larger scale in Kiruna condition of selected structural elements beyond what normal inspections provide a comparison between destructive and non-destructive testing methods an overview of the experienced indoor environment in the houses Only a few houses were investigated which means that the result will not be statistically founded and conclusions must be drawn with caution. 2. Method Suitable buildings to study was provided in the city of Malmberget, lat N67.2, lon E20.6, 69 km north from the Arctic circle. The area is called Elevhemsområdet and was be completely erased during These buildings were from the 1960s. Methods of destructive testing were developed in this area by Arfidsson et al (2011) (Johansson et al (2012). On previous visits an overview of the area was made together with the mining company LKAB who has the responsibility to find new homes for people in the area. The selection was also made in cooperation with GMG Bygg och Maskintjänst who carried out the demolitions in the area. The remaining houses were remaining due to different aspects such as to demolish entire neighbourhoods at a time. In total the study contained 19 houses in Elevhemsområdet, see Figure 1, noted house 1-19, of which 7 were also investigated by nondestructive methods, houses 1-7. Questionnaires based on the Swedish BETSI study with some additional questions concerning the life span of the house were distributed to the 13 houses that were supposed to be demolished last during the summer of Heating season would have been preferred but since the demolition was not scheduled, that would be a risk of losing the data. FIG 1. Elevhemsområdet in Malmberget that was the base for selecting the tested 19 houses and is now erased. The implementation followed the following steps Full papers - NSB 2014 page 1319

150 Selection of questionnaire for the investigation of the indoor environment and distribution together with loggers for temperature and relative humidity in some of the houses Non-destructive testing, which means normal sales inspection Planning of specific sampling in the houses. Sampling of the houses in place in Malmberget Analysis of samples in the laboratory and compilation of results. 2.1 Mould index analyses Mould is quantified by the method of Johansson (2012) described in Figure 2. FIG 2. Mould quantification by Johansson (2012). FIG 3. Moisture content in % for house 1.Paragraphed values mean that there was resin indicating a too high value in the measurements. Full papers - NSB 2014 page 1320

151 3. Results Houses 1-7 were examined by non-destructive inspections and destructive testing including a hole between two studs in a systematically chosen outer wall. Houses 8-14 were examined by smaller holes in the walls, and in all, 1-14, samples were taken from the attic. Houses were having answers on the questionnaires together with three of the houses Arfvidsson et al (2013) give a comprehensive presentation of findings from each house. Physical examinations were made in May FIG 4. The wall investigation of house 1. It is dirty in the outer parts from the outside asphalt impregnated board but without apparent moisture damages. 3.1 Sill tests A part of the sill at the ground level was taken out from houses 1-7. Figure 3 shows the moisture content in % from house 1. The methods of measurements are specified by Arfvidsson et al (2013). This sample had a clear mould odour and there was a moisture damage. All the other samples showed lower values. It is not clear if the normally lower values depend on houses not habited for a while, dry outdoor climate, high quality wood or something else. Full papers - NSB 2014 page 1321

152 3.2 Walls All 7 examined outer walls between studs except 1 had brick facade. These brick facades are all lacking ventilation in the intended air gap and in some cases air gap is missing. Most commonly, an intended air gap behind the bricks was followed by asphalt impregnated soft fibre board. The supporting structure varies between standing studs with mineral wool between to standing boards with outside insulation. If there are standing studs, generally there is an inner surface of boards. The inner surface material is in all cases a soft porous wooden fibre sheet with wallpaper. Asphalt impregnated paper is usually used as air and vapour barrier on the inside. In many cases the insulation material is mineral wool, and in some cases paper or saw dust. Figure 4 gives examples from the wall investigation between the studs in house Attics All houses, 1-14, had cold attics which is a known risk construction. Boverket showed that 18% of houses built before 1976 with cold attics had mould odour or mould. Viitanen (1996) described mould growth. Table 1 gives the mould index for all houses, All attics were affected by mould to very varying extents. Figure 5 shows the most extreme example where the occupant had installed a ventilation duct between the washroom and the attic. The joist between living volume and attic was most problematic, insulation was not apparently correlated to moisture, and there was some cases of mould even if visual inspections did not show it. TABLE 1.Mould index in the attics of house Paral roof means parallel roof. If there is no metal roof, there is brick roof. Insul and Saw dust means number of cm of insulation material. Then there is mould indexes for the different construction elements of the attic. Nr Metal roof Paral roof Poor vent Leak Insul Saw dust Tot Board Upp beam Truss Joist Other index index index index index 1 yes yes yes yes yes yes yes yes yes yes yes yes yes yes yes yes yes yes Full papers - NSB 2014 page 1322

153 FIG 5. Mould growth in the cold attic of house Questionnaires In 4 of the 8 houses with responses of the 13 questionnaires distributed, the occupants had lived in the house for more than 10 years, in 1 house between 6 and 10 years, and in 3 houses in between 3 and 5 years. In 3 of the houses they were very happy with the dwelling, while in the other houses they were quite happy with the dwelling. In the houses there are different types of window opening patterns from not at all to several hours a day. There were complaints about high indoor temperatures during summers and low during winters and that it was difficult to control indoor temperature..in 6 houses there was complaint on food smell spread, and in 5 there was sometimes bad odour. Symptoms from eyes or nose are reported in 6 houses but not believed by the occupant to be caused by the indoor environment. 4. Discussion and conclusions The opportunities with destructive testing in buildings have been shown and methods have been developed to enable future research to deepen understanding of how buildings act as systems over their life cycle. This provides an opportunity to knowledge feedback to the building industry and ultimately it will lead to both better and more optimal buildings and better competitiveness in the construction industry. The study has also given results from some houses that have been normally aged. Comparisons between non-destructive and destructive tests can be done at two different levels. Firstly, one can compare if damages are consistent, secondly, the construction can be compared. Inspections, which are normally non-destructive, should give fewer and more uncertain results of a building performance because the destructive testing enables access to more of the construction. On the other hand, the time that has elapsed since the occupants have moved out affect moisture damages and conditions. In this and related studies, there has usually elapsed some time between move out and investigation. What was seen in this study was that a lot coincide between inspections and destructive investigations, but there are differences both regarding construction details and actual damages. Looking at the exterior of the building should imply almost similar constructions, but the destructive testing shows discrepancies in for example air and vapour barriers and air gaps. The clear damage that was found in house 1 was shown in both inspection and investigation. On the other hand, the destructive testing in this study is also based on random checks and not total. There can be damages everywhere even if it is reasonable to believe that the construction details are the same within the same house. In the study, the choice of check points has been systematic in a random way and not based on suspected risk of damage. Full papers - NSB 2014 page 1323

154 Moisture measurements that reflect the habited conditions are a problem when there is the time elapsed between move out and investigations. An alternative could be to keep the dwelling heated and with a synthetic moisture production to that preferable should have been measured in the same dwelling. Arfvidsson et al (2011) showed that out of 3 outdoor wash room walls 2 were mouldy. In the present study there was no systematic way to test outdoor washroom walls. The demolition schedule together with the relatively low number of houses in the study unfortunately resulted on only a few investigated houses with questionnaire response. Therefore it is difficult to compare questionnaire outcome with physical findings. An example of an indicative correlation is house 5 where a leaky window frame connection coincided with complaints on cold draught. A more extensive project is funded and will start during 2013 dealing with 148 apartments in Kiruna where hopefully there will be more options to find significance. Mould in cold attics is a non-solved problem and the study points at a need of better constructions for new buildings, renovation projects and existing buildings in operation. This problem is also referring to attics with low amount of insulation, and inspections did not show all the actual problems. 5. Acknowledgements Thanks to SBUF, NCC, LKAB and GMG Bygg och Maskintjänst for funding and helping out with the study. References Arfvidsson, J, Bagge, H, Harderup, LE, Johansson, D, Stein, J, Wallentén, P, 2011, Tillståndsbedömning av naturligt åldrade byggnadskomponenter inför energieffektiviserande åtgärder en förstudie med fält och laboratorieundersökningar, Department of Building Physics, Lund University, Sweden, TVBH-7234, in Swedish Arfvidsson, J, Bagge, H, Harderup, LE, Johansson, D, Stein, J, Wallentén, P, 2013, Tillståndsbedömning av byggnader med hjälp av förstörande provning av byggnads-komponenter kopplingar till brukarnas hälsa och upplevd innemiljö, Department of Building Physics, Lund University, Sweden, TVBH-7236, in Swedish Folkhälsoguiden, 2010, Bamse en studie om barn och allergi, Institutet för miljömedicin, Stockholm, in Swedish Hägerhed Engman, L. 2006, Indoor Environmental Factors and its Association with Asthma and Allergy Among Swedish Pre-School Children, doctoral thesis, Department of Building Physics, Lund University, Sweden, TVBH-1015, Johansson, D, 2012, Building Performance Investigation in the Arctic by Help of Destructive Testing in Buildings a Pilot Study, proceedings of Cold Climate HVAC conference, Calgary, Canada Johansson, P, 2012, Critical Moisture Conditions for Mould Growth on Building Materials, Rapport TVBH-3051 Lund, Department of Building Physics, Lund University, Sweden Viitanen, H, Vinha, J, Salminen, K, Ojanen, T, Peuhkuri, R, Paajanen, L, Lähdesnäki, K, 2010, Moisture and biodeterioration risk of building materials and structures, Journal of Building Physics, Vol. 33, No 3, pp Örtengren (Sikander), E, 1988, Mögelpåväxt i friska hus, ISBN: , in Swedish Full papers - NSB 2014 page 1324

155 Full paper no: 165 Measurement of heat loss from hot water supply system in a hotel and insulation retrofit Shuichi Hokoi, Professor 1, Yoshinori Masuda, M. Eng. 1, Tsutomu Goi, Engineer 2 1 Kyoto University, Japan 2, Kansai Electric Power Co. Inc. Japan KEYWORDS: Hot water supply, Hotel, Heat loss, Storage tank, Heat bridge, Insulation retrofit SUMMARY: In Japanese hotels, 30% or more of the total energy is consumed by the hot water supply. Significant heat loss from such a supply system can be expected because the hot water is continuously circulated throughout the building. In this study, the heat loss was measured from one such hot water supply system for a hotel in the Kansai area. This hotel is 11 stories high with a floor area of 4708 m 2. It is aimed towards businessmen and has a shower in each guest room. The supply water is heated in a gas boiler, stored in a hot water tank, and then circulated via a piping system throughout the building. The temperature and flow rate of the hot water in the piping system were measured in order to calculate the heat loss from both the piping system and storage tank. More than 40% of the heat of the water was lost during circulation via the piping system. Additional thermal insulation was applied to the pipes and storage tank to examine the effect; the heat loss was significantly reduced. 1. Introduction As of 2012, carbon dioxide emissions in Japan have increased by 6.3% over 1990 levels (National Institute for Environmental Studies 2013). Thus, enormous efforts are required to reduce these emissions, even though the rate of increase in carbon dioxide emissions is decreasing. In the five sectors of industry, housing, commerce, transportation, and power generation, the growth rate of carbon dioxide emissions has remained high in the housing and commerce sectors. A breakdown of energy consumption by hotels shows that 31% is for domestic hot water use (Institute for Building Environment and Energy Conservation. 2001). Thus, reducing the energy consumption of the hot water supply is very important. In order to reduce the energy consumption of a hot water supply, the whole hot water supply system (HWSS) needs to be optimized. Different types of high-efficiency hot water boilers and heat pumps have been introduced to optimize HWSS operation and reduce heat loss. Since the insulation method and the characteristics of the hot water supply and demand have a great influence on the heat loss, the actual situation should be surveyed. With respect to heat loss through hot water pipes, a wide range of studies have been carried out: from fundamental studies such as investigating the thermal properties of pipes (Iwamoto et al. 2003) and modeling the temperature change in water plugs (Mizuno et al. 1995) to practical examinations of actual systems (Wang et al. 2004, Mae et al. 2004). While there have also been many studies on HWSS and heat loss in houses (Kondo et al. 2011), few studies have considered the heat loss through piping in actual use and minimization of the heat loss of HWSS to reduce the overall energy consumption with regard to hotels. This study measured the heat loss of the HWSS for a hotel where each guest room is equipped with a shower. Full papers - NSB 2014 page 1325

156 2. Outline of measurement 2.1 Measured hotel and system Measured hotel The hotel under study is a business hotel in the Kansai area of Japan with a floor area of 4708 m 2. It has 11 stories with one underground floor for 168 guest rooms. It is made of reinforced concrete with a partial steel structure. The hotel has a restaurant and store with two large baths for common use that are partly exposed to the outside on the top floor. The check-in and check-out times are 15:00 and 11:00, respectively Hot water system Figure 1 shows the schematics of the hotel s hot water supply system (HWSS). The hot water is produced by six gas furnaces on the roof (heating capacity: kw, gas consumption: 45.0 m 3 /h (13 A)), stored in one storage tank (volume: 4454 L, dimensions: 1400 mm diameter 2500 mm length), and then sent to the large common-use baths and guest rooms, as well as a cooking room and staff rooms. The unused water returns to the storage tank. Running water is supplied to the tank. When the water temperature drops below 65 C, the water is sent from the lower part of the tank to the gas furnace and heated before returning to the upper part of the tank. This HWSS supplies the following facilities with hot water: a washbasin and shower in each guest room, the showers in the large common-use baths, the hot water tap in the cooking room, and the staff rooms. The guest rooms have no bathtub. Since the common-use baths are supplied with hot water from a separate hot water supply system, it was not measured in the present study. FIG 1. Schematics of measured hot water supply system 2.2 Measurement of temperature and water flow rate The temperature and flow rate of the hot water in the pipes and the air temperatures at several points in the pipe shafts were measured. Table 1 lists the details of the measured items, and Figure 1 shows the measured points. The letters T and F denote the temperature and flow rate, respectively. The flow rate of the hot water and running water were measured by the flow meter set in the pipe in units of 0.1 L every 10 s. The hot water temperatures were measured by thermocouples set in the pipes every 10 s. The measurement was started in May Full papers - NSB 2014 page 1326

157 TABLE 1. Measured items and locations Location Measured item T1/F1 water temperature and flow rate at branch to guest rooms T2/F2 return water temperature and flow rate at branch from guest rooms T3/F3, T5/F5 water temperature and flow rate to guest rooms T4/F4, T6/F6 return water temperature and flow rate from guest rooms T7/F7 water temperature and flow rate at inlet of gas boiler T8/F8 water temperature and flow rate to guest rooms, T9/F9 return water temperature and flow rate from guest rooms T10 return water temperature from gas boiler T11 temperature of supplied running water T12 outdoor temperature T13 air temperature in pipe shaft at branch to guest rooms T14/T16, T15, T17 air temperature in pipe shaft on 11 th, 3 rd, and 2 nd floors 3. Measured results 3.1 Water temperature and flow rate between gas boiler and storage tank As an example, Figures 2 and 3 shows temperature T7 and flow rate F7 of the hot water flowing from the storage tank to the boiler and hot water temperature T10 from the boiler to the tank on October 1. When the water temperature dropped below 65 C, the water in the lower part of the storage tank was sent to the boiler and returned to the tank after being heated. On this day, the boiler was turned on 13 times. The water was sent from the tank to the boiler at a flow rate of 26 L/10s when the boiler was in operation (Fig. 3). Flow rate [L/10s] F 0 0:00 4:00 8:00 12:00 16:00 20:00 Time (October, 1) FIG. 2 Hot water temperatures (T7 and T10): FIG. 3 Hot water flow rate (F7): October 1 October Water temperature and flow rate at exit of storage tank Figures 4 and 5 show temperature T8 and flow rate F8 of the hot water flowing from the tank to the guest rooms and T9 and F9 of the water returning from the guest rooms to the tank on October 1. The temperature peaks in Fig. 4 occurred at almost the same times as those in Fig. 2. Since return water flow rate F9 was nearly constant (Fig. 5), the difference between the flow rates from the tank to the guest rooms (F9 and F8) can be regarded as the amount of used water (showers at common-use baths, water basin and shower in each guest room, and cooking room and staff rooms) at that time. The water consumption was 0 5 L/10 s. 3.3 Guest rooms Figures 6 and 7 shows temperatures T3 and T4 and flow rates F3 and F4, respectively, for one vertical Full papers - NSB 2014 page 1327

158 pipeline from the eleventh floor to the second floor. Fig. 6 shows that the difference between inlet temperature T3 (eleventh floor) and exit temperature T4 (second floor) was 2 3 C. Since such a large temperature drop occurred during the flow from the eleventh floor to the second floor, the heat loss was not negligible. Temperature [ ] 70 T9 68 T :00 4:00 8:00 12:00 16:00 20:00 Time (October, 1) Flow rate [L/10s] 25 F :00 4:00 8:00 12:00 16:00 20:00 Time (October, 1) Flow rate [L/10s] F :00 4:00 8:00 12:00 16:00 20:00 Time (October, 1) FIG. 4 Hot water temperatures (T8 and T9): FIG. 5 Hot water flow rates (F8 and F9): October 1 October 1 Temperature [ ] T3 T4 54 0:00 4:00 8:00 12:00 16:00 20:00 Time (October, 1) Flow rate [L/10s] F :00 4:00 8:00 12:00 16:00 20:00 Time (October, 1) Flow rate [L/10s] F :00 4:00 8:00 12:00 16:00 20:00 Time (October, 1) FIG. 6 Hot water temperatures (T 3 and T4): FIG. 7 Hot water flow rate (F3 and F4): October 1 October 1 4. Heat loss through hot water system 4.1 Measurement of surface temperatures of HWSS The infrared radiation temperature was measured around the boiler and storage tank on the roof and at several points in the pipe shafts. The hot water pipes are insulated by glass wool insulation with aluminum craft and steel mesh, while the storage tank is insulated by a glass wool insulating mold, polyethylene film, and Galvalume plate. Since the reflectivity of the pipe surface was so high that measuring the surface temperature was very difficult, the surface temperature was approximated by attaching tape to the pipes, which lowered the surface reflectivity. Figure 9 shows the surface temperatures of the hot water pipe in the pipe shaft on the eleventh floor. Since the air temperature in the pipe shaft was about 25 C, the pipe surface temperature was more than 10 C higher than the air temperature despite the insulation. Figure 10 shows the temperature of the metal plate supporting the pipe: contact point A was at 45.1 C. Figure 11 shows the bulb and flange connecting the upper and lower pipes, which had surface temperatures of 56.7 C. The bulb temperature was also high at 36.9 C. Fig. 12 shows the storage tank and supporting leg on the roof floor. All (four) legs showed an area with high temperature. As the outdoor temperature decreased, the amount of outgoing heat became significant despite the surface area not being very large. Full papers - NSB 2014 page 1328

159 50.0 C FIG. 9 Pipe surface temperature (T3-T4 pipeline) FIG. 10 Temperature of metal support 50.0 C C C FIG. 11 Temperatures of valve and flange FIG. 12 Temperature of hot water storage tank 4.2 Estimate of heat loss Method of calculation The heat losses in the guest rooms and storage tank were evaluated. Heat loss in guest rooms Heat loss Q1 [MJ/day] in the guest rooms was determined by the following equation: Q1 [ T8 ( t) T9 ( t)] F9 ( t ρ) c t t (1) Where ρ[kg/l]:density of water, c [MJ/ kg]:specific heat of water. Here, the heat dissipated by cooling of the water remaining in the branch pipes from the main piping to the guest room is neglected and regarded as used heat Q2. Heat loss from storage tank (Fig. 13) Used heat in the guest rooms Q2 [MJ/day], heat supplied to storage tank Q3 [MJ/day], and heat lost from storage tank Q4 [MJ/day] were determined by the following equations: Q T ( t) T ( t)][ F ( t) F ( t)] ρc t Q3 [ T7 ( t) T10( t)] F7 ( t ρ) c (2)(3) Q [ t t t 4 Q3 Q1 Q2 (4) FIG. 13 Thermal calculation of storage tank FIG. 14 Supplied, lost, and used heat [June 2011 May 2012] Full papers - NSB 2014 page 1329

160 4.2.2 Calculated results Figure 14 shows the calculated results from June 2011 to May The daily values of Q1 Q4 [MJ/day] are shown along with the monthly average outdoor temperature. The supplied heat to the storage tank (Q3) and used heat at the guest rooms (Q2) were low in the summer season while high in the winter season. They were negatively correlated with the outdoor temperature. The heat lost from the guest rooms (Q1) showed a similar seasonal change ranging from 650 MJ (August) to 830 MJ (February). The heat lost from the storage tank Q4 had the same seasonal trend ranging from 380 MJ (August) to 590 MJ (March). This seems to have been influenced by the temperature difference between the outdoor air and storage tank. 5. Insulation retrofit for reducing heat loss 5.1 Outline of insulation retrofit Based on the measured surface temperatures by the infrared thermo-camera (section 4.1), a significant amount of heat loss occurred. Therefore, the insulation of the hot water pipes and storage tank was retrofitted. The insulation retrofit area is shown by broken lines in Figure 1, and the outline of the retrofit is shown in Figures 15 and 16. The retrofit to the hot water pipes took place over October 1 3, 2012 and replaced 25 mm glass wool insulation with pipe covers made of 20 mm thick expanded insulation. The retrofit was done for only one pipeline from the eleventh floor to the second floor. The glass wool insulation was replaced with a pipe cover because it seemed to be too tight. Additional insulating cover was attached to exposed parts such as the bulbs and flanges (25 th and 26 th October 2012). Glass wool insulation with aluminum craft was added to the exposed parts of the connecting flanges and bulbs for the same insulated vertical pipeline. Figure 16 shows the schematics of insulation retrofit for the legs of the storage tank; this was carried out on October 18 20, The entire leg was covered by 50 mm thick glass wool insulation and finished by Galvalume plates. FIG. 15 Insulation on pipe, bulb, and flange FIG. 16 Thermal insulation to storage tank 5.2 Effect of insulation retrofit The heat losses before and after the insulation retrofit were compared; heat loss Q3-4 of the retrofitted vertical line (second to eleventh floors) and heat loss Q5-6 of the non-retrofitted line (third to eleventh Full papers - NSB 2014 page 1330

161 floors) were evaluated. Figures 17 and 18 show the results. Heat losses Q3-4 and Q5-6 were calculated from the following equations. Q3 4 [ T3( t) T4 ( t)] F3 ( t ρ) c t Q5 6 [ T5 ( t) T6 ( t)] F6 ( t ρ) c t (5) (6) Lost heat [MJ/day] t Replaced insulation around hot water pipe (October 1 3, 2012) The heat loss temporarily increased on October 1 and 2 due to the replacement work. The heat loss of the retrofitted line decreased from 60 MJ/day before the retrofit to 50 MJ/day afterward. Since the heat loss of the non-retrofitted line did not change from 50 MJ/day, the retrofit can be concluded to have reduced the heat loss by 10 MJ/day Addition of insulation cover to exposed parts (October 25 26, 2012) After the insulation was added, the heat loss of the retrofitted line decreased from 50 MJ/day to 40 MJ/day. Since the heat loss of the non-retrofitted line did not change from 50 MJ/day, the retrofit can be concluded to have reduced the heat loss by about 10 MJ/day Addition of insulation to storage tank (October 18 20, 2012) Figure 19 compares heat losses Q4 in March 2012 to those in January 2013 when the mean outdoor temperature was almost the same as that in the previous year (8.6 and 7.7 C, respectively). The heat loss was 590 MJ/day before the retrofit and 340 MJ/day afterward; thus, the reduction was more than 250 MJ/day. lost heat (line without retrofit) lost heat (line with retrofit) date (September to October, 2012) Lost heat [MJ/day] t lost heat (line without retrofit) lost heat (line with retrofit) Date (October, 2012) FIG. 17 Heat loss from hot water piping FIG. 18 Heat loss from hot water piping [September 24 October 10, 2012] [October 15 31, 2012] FIG. 19 Heat loss from hot water storage tank Full papers - NSB 2014 page 1331

162 6. Conclusions In this study, the heat loss was measured from a hot water supply system of a hotel aimed at businessmen in the Kansai area of Japan. The insulation was retrofitted in order to reduce the heat loss. The main results are as follows. 1. During the flow process in a vertical pipeline from the eleventh floor to second floor, the hot water temperature decreased by 2 3 C, which is a significant heat loss. 2. Thermal bridges were identified at several areas, such as flanges connecting upper and lower pipes, bulbs, and the concrete slab in contact with the pipe. Since the surface temperature of the insulated pipes was rather high, improving the insulation seemed necessary. 3. The heat loss due to hot water circulation throughout the building was more than 600 MJ/day. 4. The supplied heat to the storage tank and guest rooms had a strong negative correlation with the outdoor temperature. 5. Replacing the usual 25 mm thick glass wool insulation with a 20 mm expanded insulation pipe cover reduced the heat loss of one vertical pipeline by about 10 MJ/day (20%). 6. Adding an insulation cover on exposed parts such as the connecting flanges and bulbs of the hot water pipe reduced the heat loss of one vertical pipeline by about 10 MJ/day (20%). 7. Retrofitting the insulation of the storage tank reduced the heat loss by more than 250 MJ/day (42%). References Institute for Building Environment and Energy Conservation Handbook for Energy Conservation of Residential and Non-residential Buildings. Iwamoto S., et al Study on Evaluation and Design Method of Hot Water Supply System for Dwelling Part 11 A Study on Heat Loss from Hot Water Pipe and Tap, Summaries of Technical Papers of Annual Meeting. The Society of Heating, Air-conditioning Sanitary Engineers of Japan.HASE. PP Kondo S. & Hokoi S Heat Loss from Hot Water Supply Line in a Residential Building. J. of Environmental Engineering (Transactions of AIJ). No PP Mae M., et al Experimental Study on Low Energy and Resource Saving Technologies for Autonomous Housing. Part 9 Experiment on energy efficiency of hot water boiler. Summaries of Technical Papers of Annual Meeting Architectural Institute Japan. PP Mizuno M. et al Heat Loss from Hot Water Drain Pipe. Part 2 Heat Loss from Vertical Drainage Stack. The Society of Heating, Air-conditioning Sanitary Engineers of Japan. No.58. PP National Institute for Environmental Studies Wang X., Mae M., Iwamoto S., & Kamata M. Examination on a Thermal Efficiency and Heat Loss of Hot Water Supply System in Dwellings Part 1. J. of Environmental Engineering (Transactions of AIJ). No PP Full papers - NSB 2014 page 1332

163 Full paper no: 166 The role of thermal investigations and user involvement in energy renovation Marlene Hagen Eriksen, PhD Student 1 Anne Sofie Lorentzen, M.Sc. Civil Engineering 1 Carsten Rode, Professor 1 Søren Peter Bjarløv, Associate Professor 1 1 Technical Univeristy of Denmark, Civil Engineering, Denmark KEYWORDS: Energy renovation, holistic approach, thermal indoor climate, user involvement, multistory residential buildings. SUMMARY: Around 75 % of the Danish building stock is built before the early 1970s where the national building regulations started addressing energy. Most of the buildings have not undergone thorough renovation, however the buildings have not exhausted their lifetime either and therefore the potential for energy renovation of the buildings is large. A wide range of technical solutions already exists for energy renovation of buildings, but full benefit of their potential can only be achieved by a holistic approach and by including the users of the buildings. This paper presents results of technical investigations performed as part of a case study of a holistic energy renovation of a typical pre-war multi-storey residential building in Copenhagen, Denmark. Findings from measurements of the thermal indoor climate are supported by indications from a questionnaire survey among the users of the building. By involving the users early in the renovation process, it is possible to create ownership of the suggested solutions, and thus to enhance the realised performance of the implemented technical solutions. The presented results have been part of a thorough investigation that forms basis for a new concept for holistic energy renovation. 1. Introduction In the municipality of Copenhagen 90 % of housing stock are multi-storey residential buildings and 85 % of these buildings are built before 1970 (Danmarks Statistik, 2013), where new regulations regarding energy was implemented in the building code. Today 43 % of the Danish energy consumption is for electricity and heating of buildings (Havelund, 2011). Since the regulations concerning energy before 1972 were much less demanding than today there is a huge energy saving potential by renovating old and highly energy consuming buildings. And since quite similar multistorey buildings represent a large part of the building stock, there is a potential to find generic solutions which can be duplicated, adjusted and applied. This paper presents the role of pre-renovation investigations of the thermal indoor climate for a multistorey residential building, covering user perception based on workshops, user surveys and thermal measurements. Based on the results of the survey, it was chosen to do the thermal measurements to investigate if the tendencies seen would be confirmed or disconfirmed by a traditional measure. The main idea is to involve the residents in the renovation process so the knowledge of the inhabitants, which can be difficult to obtain otherwise, will be taken into account early in process. It is believed that a thorough involvement of users in the building renovation process can increase the value of the project and lead to more holistic projects. This can be obtained through ongoing communication, understanding of the proposed solutions and the sense of being part of the project and being heard. By Full papers - NSB 2014 page 1333

164 giving the end users a share in the early stages of a renovation the chance of them feeling ownership of solutions adopted in the building will increase. 2. Method As part of the Danish development project Holistic Energy Renovation of Buildings, different investigations were conducted on a typical pre-war multi-storey building in Copenhagen. The case building is a 5 storey building from 1935 and it consists of massive brick walls, wooden decks, older windows and an unused attic. The building is naturally ventilated and the flats are heated by radiators in the bedroom and in the living room. The flats are in the size 39 m m 2 with the main part of the flats about 50 m 2 (OIS, 2013). No thorough renovation of the building has been performed. The investigations presented in this paper are based on a questionnaire survey conducted on the residents, followed by a workshop including owner, caretaker and residents and finally followed by measurements conducted in different flats. The purpose of combining the activities is to get as much information as possible about both the experienced and the actual thermal indoor climate and to find possible correlations between the different investigations. The questionnaire consisted of 36 questions concerning general information about the participants, economy, architecture and quality of the building, the building conditions and the perceived indoor climate. The response rate of the 136 possible replies was approximately 18 %. The questionnaire was sent out in May The workshop was facilitated by an industrial psychologist from a larger Danish consultant company, focusing on wishes of the participants and the need of the building. Around 15 people participated in the workshop and the output was a sketch of what according to the participants would be the perfect solution for the renovation of the building. The thermal measurements were used to determine the thermal indoor climate in the building. The measurements conducted took place in the last week of February 2013 in four flats. The flats were found representative for the building, being located different places in the building and containing bedroom, living room, hallway, bathroom and kitchen. In all the flats temperature, relative humidity and CO 2 measurements were done, in order to determine the thermal indoor climate. Temperatures and relative humidity was measured in the bedroom, kitchen and bathroom and living room. Additionally the CO 2 concentration was measured in the living room. The loggers were in the living rooms placed in a height cm and in the rest of the rooms in heights in the range cm. Flat 1 and 2 have kitchen and bedroom oriented NNE and living room oriented SSE. Flat 2 and 4 have kitchen and bedroom oriented WSW and living room oriented ENE. There were two permanent residents in flat 2 and one permanent resident in the rest of the flats. 3. Results The results presented in this paper are based mainly on the questionnaire survey and the thermal measurements. The results of the questionnaire survey have influenced the type of measurements chosen to be performed in the flats. 3.1 Temperatures A part of the questionnaire referred to the residents perception of the economy also related to energy use. In FIG. 1 it is seen that more than 50 % of the residents perceive their energy use on heating as Full papers - NSB 2014 page 1334

165 above average or high. This indicates that there perhaps is an issue regarding the use of heating in the building, however it does not imply the reasons for the problems. According to the replies only 48 % respond positively that they use heating in the bedroom whereas 92 % respond positively on the question regarding the living room. Furthermore, the residents were then asked if they experienced problems with adjusting the heating in the apartment. The result that can be seen in FIG. 2 indicates that approximately 50% or more experience problems adjusting the heat sometimes or weekly, most often in the living room. 100% 16% 8% 80% 16% 60% 40% 60% 20% Below average/low Above average/high Agerage Don't know/no replay 0% Bedroom Living room Weekly/sometimes Never Don't know FIG. 1 Perception of energy use for heating FIG. 2 Problems by regulating the heating As the last question regarding heating, the residents were asked if it was typically too cold in their apartment during the winter time. The result is to be found in FIG. 3 and shows that around 50 % or more of the resident s experience that it is often or sometimes too cold in the main part of the apartment during wither time. 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Bedroom Living room Kitchen Bathroom Other rooms Yes, often/sometimes No, never Don't know/no reply FIG. 3 Perception of too cold in the winter period The indications found in the questionnaire points towards problems with the heating systems, leaving the flats too cold during wintertime, without the residents being able to influence. Full papers - NSB 2014 page 1335

166 Furthermore, the workshop indicated that a new heating system was a priority for the residents, when asked what they wanted from a renovation. Therefore temperatures were measured in 4 flats, distributed across most of the building investigated. The results of the measurements in the kitchen can be seen in FIG. 4. This graph shows that the measured temperatures in general are in the lower end or outside the comfort zone (20 C-24 C) (Dansk Standard, 1993) in the kitchens, where there is no heating, with an average temperature in three of the kitchens is about 18 C. Temp [ C] Flat 1 Flat 2 Flat 3 Flat 4 FIG. 4 Measured temperatures in kitchens Both in the living rooms and the bedrooms, where the radiators are placed, the temperature is most often 18 C 20 C as seen in FIG. 5 and FIG. 6. In one bedroom the temperature is very low between C. Temp [ C] Flat 1 Flat 2 Flat 3 Flat 4 FIG. 5 Measured temperatures in living rooms Full papers - NSB 2014 page 1336

167 Temp [ C] Flat Flat Flat 3 Flat FIG. 6 Measured temperatures in bedrooms The results of the questionair are supported by the temperature measurements performed, indicating that the heating of the building should be a focus area in renovation of the building. 3.2 Relative humidity As part of the questionnaire the residents were asked questions regarding the indoor climate, among others about their perception of moisture spots and mould growth. 12 % of the residents responded that they have experienced moisture spots in the living room or the kitchen or in the bathroom and 20 % have according to the questionnaire experienced mould growth on floor, wall or ceiling in the bathroom. The measurements afterwards performed of the relative humidity in the bathroom underpin the problems with high moisture content and possible problems with mould growth. In peak periods with assumed longer use of hot water, the humidity increases from the normal level between % RH up to 95 %, FIG. 7. This is a wet though, but for one of the flats the relative humidity first drops to a level under 45 % RH after almost three hours after the peak. This gives longer periods with humidity above 75 % and an increased risk of mould growth on organic materials. Furthermore, this is much higher than the recommended humidity for living rooms below % (SBi, 2009). RH [%] Flat Flat Flat 3 Flat FIG. 7 Measured relative humidity in bathrooms Full papers - NSB 2014 page 1337

168 Also in the kitchen the relative humidity was measured and even without mechanical ventilation, the relative humidity only increased to a level above 45 % RH in one of the flats in peak periods, as seen in FIG. 8. RH [%] Flat 1 Flat 2 Flat 3 Flat 4 FIG. 8 Measured relative humidity in kitchens 3.3 Carbon dioxide air exchange with outdoor air As a measure of the indoor climate, also the CO 2 concentration in the living room was measured. The measurements show that the normal level of CO 2 does not rise much above the assumed outdoor CO 2 - concentration on 380 ppm (Indeklimaportalen, 2013), FIG. 9. For flat 2 with two permanent residents, the CO 2 -concentration in peak load periods does not exceed the level on 1000 ppm recommended for workspaces and for the rest of the flats, the CO 2 -concentration in peak load periods is between 450 and 600 ppm. CO 2 -concentration [ppm] Flat 1 Flat 2 Flat 3 Flat 4 FIG. 9 Measured CO2-consentration in living rooms These values indicate that there is a large uncontrolled exchange of air with the outside. As the last part of the questionnaire the residents were asked to evaluate if they would be willing to pay more in rent if the indoor climate would be improved. The result can be seen in FIG. 10 below, showing that 52 % are not willing to pay a higher rent for an improvement of the indoor climate and another 32 % have not made up their mind. Full papers - NSB 2014 page 1338

169 32% 16% 52% Yes No Don't know FIG. 10 Willingness to pay a higher rent if the indoor climate is improved This indicates that most of the residents can point out areas where the indoor climate can be improved, but the willingness to pay for it is low. This is a challenge for the building owner and underpins a problem in the Danish legalisation often referred to as the owner-tenant paradox. 4. Discussion The temperature levels, the level of relative humidity and the leakiness in the flats are problem areas that are all applicable for improvements in a renovation project of the building. This is indicated in the questionnaire and is in general supported by the measurements performed in four of the respondents flats. This underlines the problem areas of the case building and it is reasonable to believe that the problem areas and solutions could be the same in buildings of the same type and therefore the results are useable in similar buildings. It is clear that the building needs thorough renovation to improve the thermal indoor climate, however economical challenges such as who has to pay for the increased comfort is still present. 4.1 Data basis By comparing user survey and measurements it is possible to show a connection between perceived and actual thermal indoor climate. The reliability of both the user survey and the thermal measurements are questionable regarding the low response rate in the questionnaire and the very few flats used for thermal investigations. With the relative low response rate of 18 % for the questionnaire, the quality of the composition of the replies can be questioned compared to the actual composition of residents in the building. Furthermore, also the choice of flats for the thermal measurements could be questioned as representative for the average resident. Nevertheless, the findings in the survey along with the workshop for the residents are supported by the measurements. Another uncertainty is the difference in time between the user survey and thermal measurements. However no improvements were done to the building in the period in-between, but the year-to-year difference in the winter may have influenced the replies. Furthermore, the measurements took place only in the winter period and data from the other seasons are thus not available for comparison. 4.2 Further investigations It is evident that more investigations would be preferable in order to give a more precise picture of the thermal conditions of the building. For example it would be relevant to measure the inside wall temperatures of the cold un-insulated exterior walls and compare these with the relative humidity in order to examine the probability of moisture and mould growth on the walls. Full papers - NSB 2014 page 1339

170 The infiltration level and hence the draught problems should also be further investigated. This could be done by a Blower Door test that could be used for both investigation of leakiness of doors and windows. As the building has not yet been renovated, it is difficult to evaluate the influence of the user involvement, however it can already now be seen that the early involvement has given a broader and more holistic perspective of the renovation and caused more detailed investigations of the building. The challenge will be through the renovation process to keep up the user involvement and thereby creating ownership of the solutions proposed and implemented in the renovation. 5. Conclusions Based on the user survey and the technical investigations it is clear that there is a potential for renovation of the building. Especially the heating system seems insufficient and the ventilation that is uncontrolled and comes in places where it is not desired. The measurements underline the thermal problems identified in the user survey and the conclusions are more reliable when there is a match between measured and perceived conditions. Even though further investigations would have clarified the problem areas even more, the performed measurements gives technical evidence of the main problems perceived by the residents. These investigations and their results once again indicates the importance of early user involvement, as picking up on some of the areas proposed by the users have broadened the perspective of the renovation and added knowledge to a more holistic approach. It is likely that the found problem areas can be used generally for buildings in the same category since the results were consistent and in agreement with the expected. It will be preferable to do more investigations for similar buildings to document the findings. References Danmarks Statistik. (2013). Statistikbanken. Retrieved November 11, 2013, from Dansk Standard. (1993). DS 474 Norm for specifikation af termisk indeklima, 1. udgave. Havelund, M. (2011). Hvidbog om bygningsrenovering - Et overblik over den eksisterende viden og de væsentligste studier af renoveringseffekter. Bygherreforeningen og Grundejernes Investeringsfond. Indeklimaportalen. (2013). Indeklimaportalen - Alt om indeklima - Indendørs CO2. Retrieved November 11, 2013, from OIS. (2013). Ejendomsdata for Kretahus. Ministeriet for By, Bolig og Landdistrikter. SBi. (2009). Skimmelblog - Relativ fugtighed. Retrieved November 11, 2013, from Full papers - NSB 2014 page 1340

171 Full paper no: 167 Mieke Deurinck, Ph.D. 1 Dirk Saelens, Professor 1 Staf Roels, Professor 1 Predicting energy savings at district level: representative vs. individual dwelling approach 1 KU Leuven, Department of Civil Engineering, Building Physics Section KEYWORDS: residential buildings, energy savings, aggregated level, Monte-Carlo analysis SUMMARY: When predicting energy savings at aggregated level, a common simplification is the representation of a large group of similar houses by one single representative dwelling, occupied by one specific inhabitant. The calculated energy savings for this representative dwelling are then multiplied with the number of houses to obtain the expected aggregated energy savings. In this paper, this representative dwelling approach is compared with the individual dwelling approach where multiple different dwellings are modelled separately and their energy savings are added. When combined with probabilistic user behaviour, it is found that the representative dwelling approach predicts similar mean aggregated savings, but underestimates the actual spread due to the lack of variety in building characteristics. 1. Introduction Policy makers often rely on aggregated building stock models to estimate the energy saving potential of future policy measures. Due to the aggregated scale, assumptions and simplifications have to be made to keep the building stock models manageable. A simplification often used is the representation of similar housing groups by a single dwelling model with most probable characteristics like size, orientation, insulation level, equipment etc. This is called the representative dwelling approach (Cyx 2011). A single user behaviour profile which best reflects the average user is then chosen and the calculated energy savings for this single dwelling are multiplied with the number of houses to obtain the expected aggregated energy savings. The main advantage of this approach is the limited modelling work and reduced calculation time. However, there are some important disadvantages. The existing variability in building use and characteristics, even for houses belonging to the same district, cannot be reflected by one single building model and a single user, which limits its applicability for policy makers. Also, the specific choice and combination of both the representative dwelling and average inhabitant have an important impact on the predicted energy savings. If one would select another dwelling and /or user, the aggregated outcome could be heavily influenced. In this paper, two different approaches in modelling energy savings at district level will be compared: (i) the representative dwelling approach where a fictitious dwelling is modelled, based on average characteristics from the district, and where its energy savings are scaled up to district level and (ii) the individual dwelling approach where 10 individual dwellings, sampled from the district, are modelled in detail and where their individual energy savings are added up to compose the district savings. In both approaches, the user behaviour will be implemented in a probabilistic way, meaning that heating patterns and temperature setpoints are given by probability distributions instead of fixed values. A Monte Carlo analysis based on the maximin Latin-hypercube sampling is performed to obtain the overall spread on the energy savings due to this user behaviour. In the next section, the case district and the generic building model are described and the composition of the representative dwelling is discussed. The third section shows how the probabilistic user Full papers - NSB 2014 page 1341

172 behaviour is modelled and how the Monte Carlo analysis is performed. The final section discusses the results of the simulations. 2. Case study 2.1 Description The case study in this paper is a small district in Leuven, Belgium, consisting of 52 identical dwellings built by the same building company around They are relatively large 2-storey dwellings with uninhabited attic, both in detached and in semi-detached typology. Some pictures and the original floor plan of the dwellings are given in Figure 1. The total volume V is 432 m³ and gross floor area (including garage) is 162 m². Due to the limited floor area of the ground floor, many owners have enlarged the dwellings by adding a ground floor extension at the backside. Outer walls are cavity walls in brick. Both slab-on-ground and internal floors are concrete structures, while the pitched and flat roofs are wooden structures. FIG 1. Casestudy dwelling (open and semi-terraced) and floor plan of ground floor (dimensions in mm). The detailed survey information of 10 randomly sampled dwellings can be found in Table 1. Although all dwellings were originally uninsulated, roof insulation (mineral wool) and cavity wall insulation (blown-in foam) is recently installed in most of them and original windows sometimes have been replaced by better performing ones. The overall mean U-value, U m [W/(m²K)], varies between 0.76 and TABLE 1. Survey data of 10 individual dwellings Dwelling Typology (Semi)- Detached D D D D D S-D S-D S-D S-D S-D orientation front facade NW SE SE SE NE SE SE SE SE SW depth extension [m] d wall,pur [m] d roof,mw [m] d floor,pur [m] U window [W/(m²K)] v 50 [m³/(h.m²)] Condensing boiler? [Yes/No] Y Y Y N Y Y Y N N N U m [W/(m²K)] BES-model A generic building model, easy adaptable to simulate the different building variants, is developed in TRNSYS, a dynamic building energy simulation (BES) software package. The dwelling is divided in 2 zones: a dayzone at the ground floor (V day = 243 m³, A fl,day = 80 m²) and a nightzone at the second Full papers - NSB 2014 page 1342

173 floor (V night = 189 m³, A fl,night = 82 m²). The depth of the extension (if present - see Figure 1 and Table 1) is treated as a parameter in the BES-model, leading to an enlarged volume and heat loss area of the dayzone. Different temperature settings are applied in both zones. Each zone is considered as one node for which heat balances are solved every time step. The time step is set at 30 minutes. Heat transfer between the different zones is assumed to occur only by heat conduction through the internal walls and floors, thereby neglecting possible heat transfer via interzonal air flows. Hourly outside conditions are taken from the Meteonorm weather data file of Ukkel, Belgium. Air infiltration rates are expressed as a function of the heat loss surface area A i and the measured air permeability at 50 Pa, v 50 [m³/(h.m²)], given in Table 1: inf = 0.04 v 50 A i [m³/h]. Since none of the dwellings are equipped with a ventilation system, no additional ventilation rates are incorporated in the BES-model. Hence, the heat loss due to the occasional air flows from opening windows and doors is not included in the calculated heat loss, leading to a slight underestimation of total energy use. Internal gains are assumed only function of the heated volume and set constant throughout the year (Φ int = ( /V)*V [W]), which is consistent with the Flemish implementation of the EPBD. 70% of this value is attributed to the dayzone, the remaining part to the nightzone. To reduce the calculation time, the heating system is not explicitly modelled in TRNSYS. Instead, a monthly overall efficiency of the heating system η TOT,m [-] is used to obtain the monthly total energy use E use,m [kwh] = E net,m / η TOT,m with E net,m [kwh] the monthly net energy demand. E net,m is obtained by using an ideal heater (no production, distribution or emission losses and no thermal inertia) in the TRNSYS model and is defined as the energy the ideal heater would need to deliver to reach the zone set point temperatures at any time. 30% of the heat is emitted by radiation and the remaining part by convection, which corresponds with convecto-radiators, an emission system commonly used in Flanders. The heating power for each zone is limited by the maximum heating power as determined by the European standard EN The monthly overall heating efficiency η TOT,m for different systems and control parameters is obtained from Peeters et al. (2008) in function of the monthly heat balance ratio, being the ratio of the occurring heat gains (internal and solar gains) and the occurring heat losses (ventilation, infiltration and transmissions losses). Here, two systems are chosen: (i) an on/off noncondensing high efficiency boiler and (ii) a modulating condensing boiler, both with central room thermostat and no thermostatic valves on the convecto-radiators. 2.3 Composition of the representative dwelling Given the survey information in Table 1, a representative dwelling could be composed in different ways. One could choose to search for a dwelling likely to occur in reality and as close to the average values as possible, or one could choose to compose a fictive dwelling that equals the average values, even if these values do not occur in reality. In this paper, the last approach is chosen: a fictive dwelling is made by averaging all dwelling parameters (see Table 2). This is also done for the typology and geometry, resulting in a semi-terraced typology with a common wall area half the common wall area of the semi-terraced typology and an extra outer wall area equal to half the outer wall area of the open typology. For the averaged heating system, the weighted average efficiency of both monthly system efficiencies is applied. Yet, the modelling software imposes an important limitation in averaging the U-value of the window, since a predefined window has to be chosen in the simulation software. The mean U-value of the representative dwelling should equal 1.88 W/(m²K), a U-value which is not commercially available and thus, not readily available in the software. Therefore, the representative dwelling model is duplicated: once with a window type with U=1.4 W/(m²K) and once with a window type with U=2.83 W/(m²K). The energy uses of both dwelling models are then weighted averaged with weighing factor f 2.83 = ( )/( ) and f 1.4 =1-f 2.83 to obtain the final energy use of the representative dwelling. Yet, one has to be aware of the limitations of the latter procedure. The window type does not only influence the transmission losses by its U-value, but also influences the amount of solar gains by its solar transmission factor (g-value). Or thus, the representative dwelling should in fact also have the Full papers - NSB 2014 page 1343

174 average g-value of all 10 dwellings. Since the predefined window types come with a fixed combination of U-value and g-value, a choice has to be made which of both values will be averaged in the representative dwelling. For moderately insulated dwellings as in this paper, the heating season energy use is proven to be more sensitive to the exact U-value than to the solar gains (see Brohus et al. 2009, Firth et al. 2010), so the U-value is chosen here. TABLE 2. Composition of the representative dwelling Typology Free/Semi- Terraced Representative dwelling orientation front facade 216 (S=0 /W =90 ) depth extension [m] 2.1 d wall,pur [m] d roof,mw [m] d floor,pur [m] U window [W/(m²K)] (1.88) (1-f)*1.4 + f*2.83 v 50 [m³/(h.m²)] 9.33 Heating system 60% eff condensing + 40 % eff non-cond 3. Incorporating probabilistic user behaviour Instead of using a fixed heating schedule and/or temperature setpoints, all dwelling models are subjected to different possible combinations of heating schedules and setpoints. As such, the expected energy consumption of every dwelling will be formulated as a probability distribution rather than a fixed deterministic value. The possible heating patterns and temperature setpoints and their respective probability distributions are defined in 3.1. The procedure to compose stochastic user behaviour from these distributions is explained in Heating patterns and setpoints By lack of reliable and extended datasets, realistic user behaviour is defined based on the approach of Deurinck et al. (2012). Based on mainstream employment status (full-time out to work, halftime out to work, continuously home), the different time schedules from Table 3 are imposed in both day- and nightzone. TABLE 3 - Overview of the different deterministic time schedules in the dayzone and nightzone. All 7 days of the week are identical, except for the dayzone where during the weekend dayzone pattern 4 is always used. X = set temperature presence, -- = set temperature absence, = no heating. dayzone nightzone :00 06: X X X 06:00 09:00 X X X X X 09:00 12:30 -- X -- X 12:30 17: X X 17:00 22:00 X X X X X 22:00 00: X X X PROBABILITY Table 4 summarizes the probability distributions used. In total, 11 parameters are to be altered per simulation run. The set temperature in the dayzone during presence is picked from a uniform distribution between [19-21] C. During absence and during night, the set temperature in the dayzone is picked from [15-18] C. The nightzone is never heated during the day. During the night, a probability of 0.3 is attributed to the chance that the nightzone is heated to a temperature of S-T Full papers - NSB 2014 page 1344

175 [13-18] C; the remaining 0.7 probability is attributed to the nightzone being unheated. Probabilities of occurrence are arbitrary attributed to each of the time schedules. After a time schedule is chosen, each of the start and end times of every heating period is altered with a random value picked from a uniform distribution between [+0.5h,-0.5h]. Finally, the internal gains are uniformly changed by [-20%; +20%] of their initial value of section 2.2. Remark how all parameters are assumed to be uncorrelated, which is unlike reality. For example, elderly persons are likely to be at home all day (see schedule 4) and tend to choose higher temperature settings. However, for the pragmatic modelling of user behaviour in this paper, correlations are not considered. TABLE 4 Probability distributions for the 11 user behaviour parameters (p = probability; U(a,b) = uniform continuous distribution between a and b; Bern(p) = Bernoulli distribution with p = chance at success) nr parameter distribution 1 T day,presence U(19 C, 22 C) 2 T day,absence U (15 C, 18 C) 3 4 T night,presence U(13 C, 18 C) * Bern(0.3) 5 8 start and end times (max #4) initial start/end time + U(-0.5 h,+0.5 h) 9 Heating pattern dayzone p(1)=0.5 ; p(2) = p(3) = ; p(4) = Heating pattern nightzone p(1) = p(2) = p(3) = 1/3 11 Internal Gains initial value * U(0.8,1.2) 3.2 Monte-Carlo analysis using maximin Latin-Hypercube sampling scheme The Monte Carlo technique is used to vary all 11 user behaviour parameters simultaneously in multiple simulation runs, leading to a large range of possible output values per dwelling. The parameter sampling is done with a distance-based space-filling maximin sampling scheme that maximizes the minimal distance between Latin Hypercube sampling points and that proves to be more efficient than a random or Latin Hypercube sampling (Janssen 2013). Due to this efficient sampling scheme the number of simulation runs per dwelling can be limited to 100 runs. Note that only one single sampling scheme (with 100 user profiles) is generated and re-used for all dwellings, since this is the only way one can be assured that the observed differences in output are to be attributed to the different user characteristics and not to different sampling schemes. Or, this means that the same set of 100 stochastically defined inhabitants is used for all dwelling simulations. For the calculated energy savings in this paper, this also implies that the user and its heating habits remain the same before and after retrofit. However, it is generally known that inhabitants tend to take back part of the potential energy savings in enhanced indoor comfort by increasing the set temperature, heating more rooms more often etc. This effect, known as the rebound or temperature takeback effect, is not incorporated here. 4. Predicting energy savings To illustrate the methodology, only one retrofit measure is discussed here: all pitched (MW) and flat roofs (PUR) and the ceiling (MW) between nightzone and unheated attic are insulated to reach a total insulation thickness of 0.2 m. Note that this might not be an economically viable retrofit measure for every single dwelling, since some cases already have (partly) insulated roofs. However, for this study the economical viability of a retrofit measure is not assessed, but the applicability of aggregated models evaluated. To calculate the energy savings, the BES-model of every dwelling with every sampled user profile thus needs to be simulated twice, both for the original and retrofitted situation. Per dwelling, this leads to 100 calculated energy use values before and after retrofit and thus, to 100 values of net energy savings. Full papers - NSB 2014 page 1345

176 4.1 At dwelling level In this section, the simulation results of the 11 separate dwelling models (10 individual dwellings and 1 representative dwelling) are discussed. Figure 2 shows the empirical cumulative distributions of the total heating season energy use, both before and after retrofit. These cumulative curves show both the influence of the insulation levels on the energy use (compare the lateral position along the x-axis between left and right curves) and the impact of user behaviour on the calculated energy use (the steeper the cumulative curves, the lower the impact of the user behaviour on the energy use). The mean energy use can differ by a factor two, with the representative dwelling situated in the middle of all curves. The curves before retrofit are slightly flatter than the ones after retrofits. This indicates that the energy use in pre-retrofit dwellings is more sensitive to inhabitants than post-retrofit dwellings. Cumulative frequency individual representative x 10 4 Energy use before [kwh] 0.2 individual representative x 10 4 FIG 2. Cumulative plots of total heating season energy use for every individual and the representative dwelling model, both before (left) and after (right) retrofit. Figure 3 shows the empirical cumulative distributions of the resulting energy savings at dwelling level. Here, user behaviour heavily impacts the distributions. Around the cumulative frequency of 0.7, a clear shift in distribution is seen. This shift divides the users who do not heat the nightzone (70% of them, see Table 4) and those who do heat the nightzone. Although it is an artificial division due to the rigid application of the proposed user behaviour in section 3.1 and thus, unlikely to occur in reality in this extent, it does show how heating patterns can have a great impact on calculated energy savings. If only part of the dwelling is heated, the energy savings will be lower and less influenced by temperature setpoints and time schedules (see first steep part of curves). If one chooses to heat the nightzone during the night, the energy savings will be higher and a larger spread is found around the mean value (see flatter second part of curves). Since the representative dwelling already has some roof insulation before renovation, as is the case for the 4 dwellings at the left of it, the impact on the energy savings of heating the nightzone is less pronounced. Cumulative frequency individual representative Energy savings [kwh] FIG 3. Cumulative plot of total heating season energy savings of every dwelling model. Cumulative frequency Energy use after [kwh] Full papers - NSB 2014 page 1346

177 4.2 At district level Composing the district data The district level is defined here on a small scale, the sum of 10 dwellings. To compose the data at this level, single datapoints are randomly picked from every dwelling and added up. Due to the small calculation time of this procedure, this can easily be repeated times, resulting in aggregated values for each approach. For the representative dwelling approach, 10 values are picked only from the representative dwelling values. For the individual dwelling approach, one value is picked from each of the 10 dwellings, so every dwelling is always represented once in the aggregated sample. Note that the composition of the district data by (randomly) sampling 10 values, adding them and repeating this multiple times, is the appropriate procedure to obtain a reliable distribution of the aggregated outcome. Another procedure would be to obtain 100 aggregated energy values by multiplying the 100 representative dwelling values by 10 (representative dwelling approach) or by adding all 10 energy uses under the first user to obtain a first aggregated energy use value, adding all energy uses under the second user for a second value etc. (individual dwelling approach). However, a housing group is then composed in which all 10 dwellings are each time inhabited by the same type of user, which is very unlikely in reality and which leads to an overestimation of the actual spread in aggregated energy use Results Figure 4 (left) shows the empirical probability distributions of the total heating season energy use at the district level. All are best fitted with normal distributions (see Table 5). The mean values of the fitted distributions before retrofit differ by about 1%, while after retrofit, the mean values differ by only 2%. Or, both approaches predict almost equal mean aggregated energy use, both before and after retrofit. The spread for the individual dwelling approach is slightly higher before retrofit (due to the variation in building characteristics), but the difference with the representative dwelling approach remains quite small. This means almost all variation is defined by the user behaviour. This is an important finding in favour of the representative dwelling approach: if the user behaviour is indeed as variable as assumed in section 3, the spread in aggregated energy use might be predicted equally well by a single dwelling model and stochastic user behaviour than by 10 separate dwelling models with the same stochastic user behaviour. Density 1.5 x individual before individual after representative before representative after Density 2 1 x 10-4 individual representative lognormal fit - individual lognormal fit - representative x 10 5 Aggregated energy use [kwh] x 10 4 Aggregated energy savings[kwh] FIG 4. Probability distribution of total heating season aggregated energy use (left) and aggregated energy savings (right). The district energy savings are also shown in Figure 4 (right). Both proved to be best fitted by lognormal distributions. The mean energy savings values practically equal the difference between the before and after values of Table 5 and also, the difference in mean value between the 2 approaches is Full papers - NSB 2014 page 1347

178 very small. Due to the probabilistic approach however, additional information is available about the possible spread in energy savings, given the user behaviour from section 3. Now, the two approaches do differ from each other. The standard deviation of the individual approach is almost twice as large as the standard deviation from the representative approach. Based on the fitted distributions in Table 5, the probability that the predicted aggregated energy savings are lower than 33 MWh, is only 5% for the representative dwelling approach but still more than 25% for the individual dwelling approach. So, the representative approach could easily overpredict the amount of district energy savings. This might be important when also costs are to be involved in the analysis, because lower energy savings than expected lead to larger payback times and lower return on investment rates. TABLE 5. Fitted probability distributions: normal ~N(μ ; σ) and lognormal ln(μ ; σ ) with μ = mean and σ = standard deviation in kwh. individual dwelling approach representative dwelling approach before ~N( ; 5055) ~N( ; 3906) after ~N( ; 2974) ~N( ; 2960) savings ~ln (35320 ; 3186) ~ln (36536 ; 1834) 5. Conclusion Using a representative dwelling to represent a larger group of similar houses does not automatically lead to bad energy saving predictions. If rigorously composed to match the average building characteristics and when combined with probabilistic user behaviour, the mean predicted energy savings of the representative dwelling approach are almost equal to the mean energy savings predicted by the individual dwelling approach. However, if one is also interested in the calculated spread on the energy savings, the representative dwelling approach performs less, since no spread due to differences in building characteristics can be taken into account. For districts with a uniform housing population, e.g. low renovation rates in the past, the representative dwelling approach thus could be an option. For districts where a considerable part of the houses already has been renovated to a small or large extent, it could be important to gain more information about the spread on building characteristics and to include more dwelling types as is done in the individual dwelling approach. References Brohus, H., Heiselberg, P., Hesselholt, A., & Rasmussen, H. (2009). Application of partial safety factors in building energy performance assessment. Eleventh International IBPSA Conference, July , Glasgow, Schotland. Cyx, W., Renders, N., Van Holm, M., & Verbeke, S. (2011). Report IEE TABULA - Typology Approach for Building Stock Energy Assessment. Deurinck, M., Saelens, D., & Roels, S. (2012). Assessment of the physical part of the temperature takeback for residential retrofits. Energy and Buildings, 52, Firth, S. K., Lomas, K. J., & Wright, A. J. (2010). Targeting household energy-efficiency measures using sensitivity analysis. Building Research & Information, 38(1), Janssen, H. (2013). Monte-Carlo based uncertainty analysis: Sampling efficiency and sampling convergence. Reliability Engineering & System Safety, 109, Peeters, L., Van der Veken, J., Hens, H., Helsen, L., & D haeseleer, W. (2008). Control of heating systems in residential buildings: Current practice. Energy and Buildings, 40(8), Full papers - NSB 2014 page 1348

179 Full paper no: 168 Testing a new method for VIP interior insulation for heritage buildings Stefan Bichlmair 1 Martin Krus 1 Ralf Kilian 1 1 Fraunhofer Institute for Building Physics IBP Holzkirchen KEYWORDS: VIP interior insulation, reversible mounting, retrofitting historic buildings, listed buildings SUMMARY: In old traditional buildings and even more in listed, historic buildings energetic refurbishment has to be planned thoroughly. For small rooms it may be of advantage to use thin and high efficient internal wall insulation systems like vacuum insulation panels (VIP). These systems have the best ratio of thickness to insulation, but they are also absolutely diffusion tight. This tightness makes a VIP system sensible to air leakage and air flows behind the panels. Also the mounting of interior insulation mostly affects disadvantageously the original surfaces and plasters. In case of valuable buildings this can lead to restrictions in retrofitting energy saving measures such as interior insulation. A special focus is therefore put on reversible application in historic buildings. Refurbishments in old and valuable buildings should be carried out without or at least minimal damage to original surfaces and plasters compared to typical mounting systems. The presented mounting system reduces possible damage to original surfaces and plasters if a removal is necessary. The new system uses an additional layer between interior insulation and original surfaces to protect the surface and enable a save fixing. The interior insulation system also has been tested experimentally to assess the influence to the original surface and possible damages to the plaster. Therefore the original surface was evaluated before mounting and after removal of the VIP interior insulation. The performance of the interior insulation was measured for one heating period. This paper highlights the concepts of reversible application of the system. Also the measured data of the interior insulation and comparison of the simulation results with the measured data are shown. 1. Introduction The use of internal wall insulation with vacuum insulation panels (VIP) provides a way for energetic building stock refurbishments, where special consideration to the external appearance of a building has to be taken into account and thin insulation thicknesses are required. The installation of VIP internal wall insulation in building stock often faces design problems and issues of removal and reversibility. Conventionally, fully adhered assemblies could not be dismantled without damage to the original surface. 1.1 Mounting of VIP One problem of vapor-proof systems such as VIPs is a possible backside air flows between VIP- Panels and wall if there are cracks in the surface layer and cavities behind the Panels. These backside air flows can transport moisture and mould spores from the indoor air to the cold wall surface Full papers - NSB 2014 page 1349

180 underneath the interior insulation and may lead to mould growth. To avoid backside air flows is full bonding of the panels to the wall, but this cannot always be guaranteed under the conditions of typical construction sites. The best method is to put the adhesive on both sides. But unevenness of old wall constructions, stiffness of panels and a thin use of adhesive makes failures possible. Not durable tight joints to adjacent building components in conjunction with cavities may lead to back side air flows. Existing wall surfaces with loose plaster or painting could make an addition doweling of the panels necessary. Doweling of VIP panels is mostly only in additional edge zones made of e.g. polystyrene possible. This edge zones or dowel zones are additional thermal bridges. 1.2 Conservation of original surfaces and plaster in cultural heritage preservation In the view of cultural heritage preservation a mostly comprehensive tradition of the historical building substance is aspired (Charta of Venice 1964). Also plasters or paint layers are informative about the ancient way of live and therefore worth to conserve. The use of typical adhesives based on cement with additional plastic additives may damage or destroy these near surface layers. Furthermore substances of the adhesive may migrate into deeper material layer. With this procedure an irreversible damage may occur to these layers. A good adhering system may destroy additionally the plaster if removed. One example of near surface layers in a historic building shows FIG 1. The exposed layers are documents of the historic of the building and give an impression of the taste of the epochs. FIG 1: Exposed historic paint layers of a rectory oft he 16th century in Haimhausen in greater Munich (Picture: Klaus Klarner, conservator, Munich) 2. Experimental setup and measurements The general aim of the project is the innovative application of an exemplary wall construction with vacuum insulation panels (VIPs) in combination with adhesive mats in the field of internal wall insulation of the building stock. New solutions and approaches should be developed, tested and demonstrated by the investigations. In conjunction with an adhesive mat as a separating layer, it is possible to design the internal wall insulation removable, as an important aspect of the reversibility for renovation and repair work in old buildings and historic preservation areas. The mats are pinned with only a few dowels to the wall. The adhesive mats are made of thin mesh with single-lined fleece. The fleece protects the original surface from the adhesive mortar used for fixing the VIPs to the wall, and thus enables an almost completely reversible attachment. Similarly, these mats enable a better adaptation of the dowel position given on the ground, which offers the possibility of placement in voids and thus help to protect valuable wall areas, e.g. with decorative historic paintings, and therefore can be beneficial for conservation reasons. Full papers - NSB 2014 page 1350

181 In this project, a combination of measurements in a case building object and computational simulation is performed, which serve to check a prototype wall construction for the economic and safe use of VIPs for existing buildings. In addition, a possible change of condition of the masonry surface will be examined and the measured data will be processed. In a first step different adhesive matt systems that are available on the market, have been looked for and one of them is selected for use. Prior to installation, the construction was checked by calculation with the hygrothermal building simulation software WUFI (Bichlmair et al 2012) developed by Fraunhofer IBP (Künzel 1994). The experiments took place in a building at the test site of the Fraunhofer IBP Holzkirchen. Here suitable test buildings, laboratories and workshops for the implementation of the project exist and also the required climate data for the site are known. FIG 2 shows the interior of the test building with test setup during mounting. FIG 2. Interior view of the experimental building with the east and south walls with color swatches and some already mounted adhesive mat. FIG 3. Component opening with layer indication of the interior insulation at the east wall. If the interior insulation is not fixed accurate backside air flows with infiltrated indoor air may occur and then mold growth is possible. For specific measurements of backside air flow, the application of adhesive mat is suitable, since a defined layer of air of approximately 1 cm is present in the mesh (FIG 2 and 3). To achieve a higher level of security against backside air flow in the adhesive mat special horizontal seal joint was formed dividing the masonry in three sectors. To implement this sealing without thermal bridging a double layer VIP system with ca. 10 cm thickness were used for this purpose, originally developed for exterior insulation (Kolbe 2012). To assess the impact of the seal joint on the original wall surface different sealing methods were developed. Four different systems were selected therefrom and applied (Bichlmair et al 2012). In addition, some specific open joints as a gap with ca. 1 mm width were produced to the adjacent bottom and ceiling as a reference for not tight seals to adjacent building parts. The test setup for measuring backside air flows was planned and carried out (Bichlmair et al 2012) with the homogenous tracer gas emission method (NTVVS 118, 1997) due to good experiences with this method in historic buildings (Kilian et al Full papers - NSB 2014 page 1351

182 2011). The measuring period were too long, therefore the results cannot be used for interpretation. The further investigations on backside air flow were made computationally, based on the measured data for temperature and used test set up for the experiment. The surface of the wall was painted with defined colors with different historical binders (FIG 2 and 4) to assess changes in consequence of the insulation (Bichlmair et al 2012). The color values were measured prior to mounting, using the standard method of CIELAB (EN ISO 11664, 2012). After removing the interior insulation, a further check on the original wall was made in order to assess the surface concerning degree of damage-freeness and change of the color values. The measured color lightness and color space are shown in FIG 4 in the right graph. There are only small deviations of the two measurements before mounting of the internal insulation and after removing almost 6 month later. The left picture in FIG 4 shows the removed internal insulation with some leavings of the adhesive mat and adhesive. The first image shows almost complete conservation of the original surface. Only a small strip (rectangle 1) has high losses of substances. This strip was made as a reference sealing with adhesive directly mounted on the surface. Removing this directly applied adhesive led to the typical loss of the upper layers of the original wall, i.e. in this case the newly applied test colors. FIG 4. Left picture: South wall after removal of the interior insulation. For the areas of the numbered rectangles visual macro photos exist. Right graph: Color reflection on the binder system lime-casein on the south wall with the colors red, yellow and blue before and after application of interior insulation. With examination of the surface visually mold growth could be observed on several areas. Within the different colors and binder systems no obvious pattern was recognizable. As mentioned several cracks were built for some test fields. The bottom fields with cracks had only a few and small mold spots, the fields at the ceiling had the most intense mold growth. The least mold growth but still visually recognizable has been observed in fields with no deliberately installed gaps in the middle of the wall. To assess the effect of the internal insulation monitoring measurements were made with temperature sensors, relative humidity sensors and heat flow meters at the boundary layer and additionally with an infrared camera In addition, a combination of measurement and calculation is used for checking the developed component structure. In FIG 5 the north façade of the test building is shown with visual image and infrared (IR) image. The different thermal behavior is made visible with the IR image. In the center of the IR Image the insulated wall partition is shown with blue colors. The right wall partition remained with its original wall brick structure and is shown yellow-red colored. Two rectangular arranged measurement fields refer to the measured IR temperature within the insulated and untouched wall partition. The PT 100 temperature measurement is also located within the rectangular IR measurement field. One part is insulated with the internal insulation. The IR camera has an absolute accuracy of ± 1.5 C and a thermal sensitivity of 0.07 K at 23 C. Full papers - NSB 2014 page 1352

183 FIG 5. Exterior view of the north side of the test building with IR image, before sunrise, 14 th Feb 2013.The temperature measured with IR correspond to the measurement with calibrated PT 100 temperature sensor within the accuracy of the measurements. The left graph in FIG 6 shows the course of temperature of the insulated and not insulated wall partition from to Comparing the both wall partitions the former original surface behind the interior insulation is cooling down almost to the level of the outside wall surface temperature. The temperature drops below 0 C on the former original surface. The right graph in FIG 6 shows the temperature course of the insulated wall partition with additionally measured relative humidity on the former original wall surface. During the complete measuring period of almost 6 month the relative humidity was at 100 % RH, measured with a capacitive sensor with an accuracy of ± 2% RH and ±0.3 C. The PT 100 temperature sensor was calibrated to ± 0.1 K. The temperature sensor of the humidity sensor was checked with an additional PT 100 sensor at the same interstitial layer whereas the PT 100 was fixed on the original wall surface. The humidity sensor measured the air in the small air cavity within the adhesive mat. FIG 6. The left graph shows the temperature course of the north wall from 28th Jan to 6th May T AOFT names the outer surface temperature, T GS the temperature of the original surface behind the internal insulation, T IOFT the inuslated wall surface temperature inside. The right graph shows the period form 1st April to 17th April 2013 with relative Humidity and temperatur courses, whereas TGS Kombi and RH GS Kombi names the temperature and relative Humidity on the orgininal surface unterneath interal insulation.the level of RH is constant at 100 % RH. 3. Simulation Further computational studies were carried out on the basis of the first simulation which were calculated with the measured data as boundary condition and compared to measured data in the Full papers - NSB 2014 page 1353

184 interstitial layer. The influence of the assumed backside air flow was taken into account and the results of the simulation were also compared to the measured data. With this calibrated simulation long term calculations were performed with and without backside air flow and their impact calculated on the moisture balance. For the boundary conditions for the long term calculation a typical indoor moister load of residential housing between 40 % RH at 20 C in winter and 60 % RH at 22 C in summer were used. For the outdoor climate the Holzkirchen climate of 1991 were implemented and repeated for 10 years. The heat transmission coefficient was assumed outside with [m²*k/w] and indoors with [m²*k/w]. The results on the water content of the plaster and brick wall are shown in (FIG 7) on the left side and the temperature and relative humidity course on the surface of the original plaster underneath the adhesive mat on the right side in (FIG 7). The leaky construction with backside air flow shows for the plaster with the red line a certain increase in the water content within the annual cycle and also a long term increase over ten years. The water content of the brick wall with backflow air current is only increasing slightly compared to the tight construction. The relative humidity shows a similar behavior to the plaster with an annual cycle and a continuous small long term increase. FIG 7: Sequence of water content in the existing original wall plaster and brick wall of the north side with and without backside air flow of the VIP interior insultion over a time span of ten years. The right graphs shows the temperature and humidity on the surface of the original plaster underneath the interstitial layer between original plaster and new applied adhesive mat calculated with tight (black line) and leaky (red line) construction on a north oriented facade over a time span of ten years. To assess the influence of the increase of moister in the construction an additional simulation with WUFI Bio (Sedlbauer 2001) was performed on the results of the previous simulation. With this tool a predicted mould growth can be calculated based on relative humidity, temperature and substrate with transient boundary conditions. With combining the results of the bio hygrothermal model with the mould index of the Viitanen model (Viitanen H. & Ritschkoff A.1991) it is possible to use an accepted and demonstrative measure for WUFI Bio (Krus et al 2011). The mould index reaches from 0 (no mould growth) to 6 (100 % mould coverage of the surface). With a mould index of 3 a mould growth is clearly visible. For Mould Index below 1 only low risk of mould growth exists. In (FIG 8) the Mould-Index is calculated for the shown data in (FIG 7) of the 10 th year with and without backside air flow. The simulation with backside air flow shows a visible mould growth. If the construction is tight only a very low risk of mould growth can be calculated at that position of the wall construction. Full papers - NSB 2014 page 1354

185 FIG 8. Sequence of the mould-indexes in the interstitial layer between original plaster and new applied adhesive mat calculated with tight and leaky construction on a north oriented facade of the last calculated year. 4. Conclusions The investigations for removable assembly contribute to a significant development of reversible internal wall insulation. It was possible to dismantle the VIP largely non-destructive to the orginal surface. The results in terms of the conservation status of the colored surfaces and changes of the colors are encouraging. The mold growth by backside air flow could not be resolved despite the effort for sealing. With simulation the effect of the background current was reproduced and the long-term performance was calculated. From the perspective of conservation of historic surfaces the separation with lamination of the original surface from the cement adhesive is a promising option for internal insulation. 5. Acknowledgements These examinations were funded by the national public funding body BBSR, Aktenzeichen SF / II 3-F , and conducted in cooperation with St. Gobain Weber GmbH. References Bichlmair S. Krus M. & Kilian R Eine neue Mehtode zur VIP-Innendämmung im Bereich der Denkmalpflege. Konzepte Aufbau Erste Ergebnisse. In: Tagungsunterlage. 2. Internationale Innendämmkongress. TU Dresden. Dresden. Charta of Venice download of the homepage of ICOMOS, EN ISO Farbmetrik - Teil 4: CIE 1976 L*a*b* Farbenraum. Beuth-Verlag.Berlin. Kilian R. Bichlmair S. Wehle B. & Holm A Passive sampling as a method for air exchange measurements for whole building simulation of historic buildings. 9th Nordic Symposium on. Building Physics. Proceedings V3, p Tampere University of Technology, Tampere, Finland. Kolbe G Fit mit Vakuumdämmung. Ausbau+Fassade 02/2012, p Geislingen. Künzel H.M Verfahren zur ein- und zweidimensionalen Berechnung des gekoppelten Wärmeund Feuchtetransports in Bauteilen mit einfachen Kennwerten. Dissertatiton. Universität Stuttgart. Full papers - NSB 2014 page 1355

186 Krus M. Seidler C.M. & Sedlbauer, K.: Übertragung des Mould-Indexes auf das biohygrothermisches Modell zur Schimmelpilzvorhersage. IBP-Mitteilung 38, Valley NTVVS Ventilation: Local Mean Age of Air-Homogeneous Emission Technique; Nordtest Method, Finland. Sedlbauer K Vorhersage von Schimmelpilzbildung auf Bauteilen. Dissertation, Stuttgart. Viitanen H. & Ritschkoff A Mould growth in pine and spruce sapewood in relation to air humidity and temmperature. Uppsala: Swedish University of Agriculture Sciences, Department of Forest Products. Full papers - NSB 2014 page 1356