Optimal Drying of Flooded Brickwork Masonry John Grunewald, Dr.-Ing., Research Professor, Dresden University of Technology, Institute of Building Climatology, Germany grunewald@ibk.arch.tu-dresden.de, http://www.tu-dresden.de/aribk Rudolf Plagge, Dr.-Ing., Head of the Laboratory, Dresden University of Technology, Institute of Building Climatology, Germany plagge@ibk.arch.tu-dresden.de, http://www.tu-dresden.de/aribk KEYWORDS: flooded buildings, insured properties, moisture simulation, brickwork masonry, drying costs, air dehumidifiers SUMMARY: Flooding is a global phenomenon as recently highlighted by the major catastrophic events in Central and Northern Europe. When flooding occurs in an area populated by humans, it can cause substantial damage to property and threaten human life. When insured properties are damaged by flooding, insurers and repairers are called upon to provide services to homeowners in order to return the dwellings back to a habitable state. As one prerequisite to minimize costs for rehabilitation, such services should include optimal drying of flooded buildings. Those measures comprise a drying strategy tailored to local particularities as climatic conditions and building material properties, monitoring and supervising of drying and repairs as well as a comprehensive damage assessment. The paper introduces a new methodology to schedule drying measures and to evaluate their success for a large variety of buildings. The capabilities of computer simulation to predict the costs of alternative drying strategies are demonstrated by means of flooded masonry for different types of brickwork frequently used in historical buildings. The numerical simulation of moisture transport in flooded brickwork masonry allows statements about the moisture release to the indoor air, the recommended duration of usage and number of air dehumidifiers, the current drying state, the total time to return the brickwork back to a dry state and the drying costs as function of time. 1. Aims of the study The historical town of the city of Dresden was largely flooded by the river Elbe in summer. Well-known historical buildings, among them the Church of our Lady, the Dresdner Zwinger and the Semper Opera, but also many residential houses and homes were affected by this catastrophic event. FIG.1: View of the historical city of Dresden with river Elbe in foreground (left) and the flood of Dresden (right).
The total damage due to the flood was estimated to 1 Billion. A huge amount of money was spent to dry out the basement walls, to remedy possible contaminations and to restore buildings back to their original state. Questions arose about optimal drying measures and their costs: what is a dry state; how long does a drying process take; do thick walls differently behave than thin ones; can air dehumidifiers be used with advantage; what kind of ventilation should be provided and can we combine those measures? Especially insurance companies were interested in this kind of questions, in particular, the drying costs for thick historical brick walls were asked. The questions above gave reason to investigate the drying behaviour of brick walls as function of different parameters: initial state; surface finishing; duration of flooding; wall thickness and material properties; indoor climate (relative humidity, temperature, ventilation). Since the drying effort of flooded basement walls depends on many parameters, the problem was investigated by numerical simulation of coupled heat and moisture transport processes in porous media. The authors have access to appropriate simulation software DELPHIN (/1/ Grunewald, 1997; // Funk and Grunewald, ; // Grunewald, 3), developed at the Institute of Building Climatology. In this simulation study, a relation is established between the drying effort in terms of time and money and the above parameters wall thickness, material properties and indoor climate at a given initial state, surface finishing and duration of flooding.. Analysed construction Exemplarily, an unrendered brickwork masonry section of different thickness (.3m and 1.m) and.1m height was used to carry out the simulation study. FIG. shows a two-dimensional discretized construction with separate modelling of bricks and mortar joints. To simulate days flooding, a boundary condition water contact was assigned to the internal wall surface (right side in FIG.). The outside (left side) of the construction was assumed to be vapour and water tight. The drying started immediately after the flooding by using two boundary conditions: Vapour diffusion and Heat conduction. This requires specification of the indoor relative humidity and temperature, denoted as cellar climate below. internal side First days flooding then drying FIG.: Analysed construction: Brickwork masonry section of.3m and 1.m thickness and.1m height.
.1 Brick quality Three different types of brick were used to consider the influence of material parameters: a very dense stone, a normal brick and a very porous (possibly previously damaged) brick. The material parameters for normal brick (Brickwall) and mortar were taken from the materials database of the DELPHIN program (see /3/ Grunewald, ; // Plagge et al., ). The parameters of the very dense stone (Stonewall) and the porous brick (Porowall) were estimated (see TAB.1). TAB.1: Analysed brick types BRICK QUALITY Stonewall Brickwall Porowall Mortar Porosity in Vol.% 3 Water uptake coefficient in kg/m s 1/.3.9.11.. Indoor Climate The second parameter influencing the drying behaviour is the indoor climate (relative humidity, temperature). In particular, it was of interest to study the influence of constant climate to be provided by air dehumidifiers versus varying climate to be provided by good ventilation. The first case is subdivided into three levels of relative humidity, all at 1 C, and the second case is subdivided into ventilation in wintertime and summertime, just by assuming that the flooding event with consecutive drying happened in winter or summer, respectively. TAB. gives an overview of the cellar climate types taken into account. TAB.: Used indoor climate types CELLAR CLIMATES Constant 1/ Constant 1/ Constant 1/3 Ventilation Summer Ventilation Winter Temperature 1 C 1 C 1 C cellar climate cellar climate Relative Humidity % % 3% cellar climate cellar climate Remark no ventilation no ventilation no ventilation ventilation starts in summer time ventilation starts in winter time..1 Drying by air dehumidifiers A constant climate achieved by air dehumidification causes costs that depend on the type and technical data of the drying machines. Commonly available types have been collected in a list (FIG.3) and the TROTEC TTK 3 was selected for the cost calculation. For this type of drying machine, a drying performance of.7 l/h at 1 C / % causes a specific energy demand of. kwh/l. The specific energy demand contains an evaporation enthalpy of.3 kwh/l, the difference to. kwh/l is caused by energy loss due to heat production, friction and other effects that cause an efficiency factor lower than 1. Dryer Electrical Drying Specific Typ power performance* energy demand W l/h kwh/l WILMS KT 7.7 1.17** TROTEC TTK 3 73.7.** EBAC MK 1.9.7** **evaporation enthalpy.3 * at 1 C / % FIG.3: Types and technical data of air dehumifiers.
.. Drying by ventilation It is expected that the varying indoor climate provided by ventilation is similar to the outdoor climate. In principle, it follows the course of the outdoor temperature and relative humidity but with a certain attenuation in temperature due to the thermal storage capacity of thick cellar walls. In the cellar, it will be warmer and dryer than outdoors in wintertime, in summertime, it will be more cold and more moist in the cellar. The cellar climate (see FIG.) is calculated according to Eq. (1) and Eq. (), a sinusoidal attenuation of the outdoor temperature with an amplitude of 3K, a phase shift of 1d and an oscillation period of 1a has been assumed. The data of the outdoor climate was taken on an hourly basis from the climate zone Middle Germany. T cellar () t = T outdoor () t + 3K sin π t ---------------- + 1d 1a p sat ( T outdoor ) ϕ cellar () t = ϕ outdoor () t ---------------------------------- p sat ( T cellar ) Cellar temperature (1) Cellar relative humidity () Temperature - Cellar Climate 9 Relative Humidity - Cellar Climate Temperature in C 1 - Relative Humidity in % 7 3 1 Time in days 3 3 1 Time in days 3 3 FIG.: Cellar climate - attenuated temperature and relative humidity..3 Investigated variants The variants numerically simulated are shown in FIG.. Reference case is the Brickwall at Constant climate of 1 C / %. In a 1 st calculation series, the influence of brick quality is analysed, in a nd calculation series, the influence of indoor climate is investigated. Finally the cost analysis is done for the reference case. 1 st calculation series (brick quality) reference case (cost analysis) nd calculation series (indoor climate) FIG.: Investigated variants: Brick quality versus indoor climate.
3. Results To give an impression of the water uptake and drying behaviour of the analysed wall section, the moisture fields of the three brick types (3 cm wall only) are shown in FIG.; left after days flooding, right after 1 month drying. The simulation of the drying process has been continued until years for all cases, shown by the drying curves in the next subsection. In the Stonewall case, the mortar joints dominate strongly the stone bricks in terms of liquid water transport. The water uptake of the dense stone is very little. In the Brickwall case and the Porowall case, the mortar joints dominate the bricks only slightly, indicated by the penetration depth in the left pictures. In the Porowall case, the overall water uptake is much higher. There is also an interaction between the bricks and mortar joints. Even though the mortar quality was not varied, there is a different penetration depth in the mortar joints between the three cases. StoneWall\3cm [d] StoneWall\3cm [d] 3 3 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 3 3 1 3 3 BrickWall\3cm [d] BrickWall\3cm [d] 3 3 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 3 3 1 3 3 PoroWall\3cm [d] PoroWall\3cm [d] 3 3 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 3 3 1 3 3 FIG.: Moisture fields in the wall section after days flooding (left) and after 1 month drying (right). After stop of the water uptake and with start of the drying process, a redistribution process of water to the inner parts of the wall takes place. This redistribution process is more distinctive for the wall thickness of 1.m (not shown here) in comparison with the.3m thickness. The right graphics in FIG. show that the highest remaining water content can be found in -3 mm depth from the surface after 1 month drying. A shift of maximum remaining water content to higher depth is observed with higher water uptake (lower brick quality).
3.1 Influence of brick quality The integral water mass content (in kg/m length of the wall section shown in FIG.) versus time is a measure of the drying behaviour (drying curve) of the wall. The graphics in FIG.7 show the short term drying behaviour left and the long term drying behaviour of the variants analysed in the 1 st calculation series right. A significant influence of the brick quality - much higher than the influence of the wall thickness - can be observed for the water uptake process. The water uptake after days is about kg/m (11 kg/m ), 11 kg/m (1 kg/m ) and 1 kg/m (9 kg/m ) for the Stonewall, Brickwall and Porowall, respectively. The short term drying curves show an almost parallel course between both wall thicknesses. Opposite to that, the long term drying behaviour is much more influenced by the wall thickness. Particularly the Porowall shows a distinct divergence between drying curves of the.3m and 1.m cases caused by a faster redistribution process towards inner regions after flooding. Water mass in kg/m Wall length 1 1 Integral water mass Stonewall 3cm Stonewall 1m Brickwall 3cm Brickwall 1m Porowall 3cm Porowall 1m Water mass in kg/m Wall length 1 1 Integral water mass Stonewall 3cm Stonewall 1m Brickwall 3cm Brickwall 1m Porowall 3cm Porowall 1m 1 Time in [d] 3 3. 1 1.. Time in [a] 3 3.. FIG.7: Drying curve as function of brick quality over 1 month (left) and over years (right). The initial (dry) states are. -.9 kg/m and 1. - 1.9 kg/m for the.3m and 1.m walls. From the right graphics in FIG.7 it can be seen that the Brickwall (1.m) and the Porowall (1.m) do not reach a dry state after years drying time. The surface may be dry but there is a remaining moisture mass between 3- kg/m (17- kg/m ) due to redistribution of water after flooding. As a conclusion it can be stated that the material properties have a great influence on the drying behaviour of thick brick masonry. A realistic prediction of drying effort and duration must involve material measurements. 3. Influence of cellar climate The cellar climate as influencing parameter has been investigated for the.3m and 1.m thick walls; the results for the.3m wall are shown in the graphics in FIG. below. A significant difference in short term drying can be observed between ventilation in summertime and the other variants which show almost the same course. It must be stated that usage of air dehumidifiers has nearly the same effect as good ventilation in wintertime. Moreover, an increased number of drying machines has a very little effect. Ventilation in summertime is much less effective because of the high relative humidity in the cellar. 1 Brickwall\3cm 1 Brickwall\3cm Water mass in kg/m wall lenght Dryer, const 1 C / % Ventilation, Flooding in Wintertime Ventilation, Flooding in Summertime Dryer, const 1 C / % Dryer, const 1 C / 3 % Water mass in kg/m wall lenght Dryer, const 1 C / % Ventilation, Flooding in Wintertime Ventilation, Flooding in Summertime Dryer, const 1 C / % Dryer, const 1 C / 3 % 1 time in [d] 3 3. 1 1.. time in [a] 3 3.. FIG.: Drying curve as function of indoor climate over 1 month (left) and over years (right).
A permanent usage of an optimum number of drying machines in order to keep the relative humidity as constant as assumed in our simulation is advantageous only in long term view (see FIG. right). At a constant relative humidity of 3% (unrealistic case) one would return a.3m Brickwall back to its initial state after. years (shown by the horizontal black line), at % it would last 3 years and at % years. Air drying machines are recommended generally in summertime only; in cases when as fast as possible drying is necessary. In wintertime, good ventilation is the best choice. 3.3 Cost estimation Prerequisite for a cost estimation of drying measures is knowledge of the moisture load to be removed from the indoor air by the drying machines. The vapour diffusion flux density in g/mh is used for this purpose (again per length of the wall section given in FIG.). Per definition, a negative flux leaves the wall to the indoor air; the curves represented in the graphics in FIG.9 are output from the numerical simulation. Reference case is here the Brickwall at 1 C / %. The walls were flooded with C cold water, the drying was assumed to take place at 1 C. The rapid change of the vapour flux in the beginning (between days -1) happens due to a warming up of the wall. After the 1 th day, a continuously decreasing vapour flux over years can be reported. Vapour diffusion in g/mh - - - - - -1 Vapour flux from the wall into the room Brickwall 3cm Brickwall 1m Vapour diffusion in g/mh -1 - -3 - Vapour flux from the wall into the room Brickwall 3cm Brickwall 1m -1 1 1 1 1 Time in [d] 3 3 3 3 3 1 3 Time in [a] FIG.9: Vapour diffusion flux over 1 month (left) and over years (right). The vapour flux can be easily converted into a moisture load (condensate) with the height of the simulated wall section and assuming a total wall area to be dried out (here 3 m, accounting for a small castle, see upper and lower left tables in FIG.). Knowing the drying performance of the chosen air dehumidifier model (TTK 3), an optimal number of dryers can be calculated. This number of drying machines must be leased to keep the relative humidity constant. From that follow the energy costs and the leasing costs for the respective number of dryers, all as function of time due to time-dependent nature of the moisture load (see right table in FIG.). Wall data Calculation height m.1 Wall area m 3 Data dryer (TTK 3) Drying performance 1 C / % l/h.7 Energy demand kwh/l. Leasing costs (per working day) /d 1 working days /Mon 1 /d. Energy data Evaporation enthalpy kj/l Energy costs /kwh.13 FIG.: Cost calculation tables.
From the tables in FIG. it follows that the total drying costs for this particular case under the assumptions mentioned above will amount to after 1 month and to after years. It is interesting to see that the leasing costs are higher than the energy costs; a statement that depends on the specific energy costs of.13 /kwh in Germany. The graphics in FIG.11 show the short term development as well as the long term development of costs. Costs in keur 1 Total costs Leasing costs Energy costs Costs vs. time, Brickwall 3 cm Costs in keur 3 3 1 Total costs Leasing costs Energy costs Costs vs. time, Brickwall 3 cm 1 Time in [d] 3 3. 1 1.. Time in [a] 3 3.. FIG.11: Drying costs over 1 month (left) and over years (right).. Conclusions The authors demonstrate that numerical simulation can help to estimate duration and costs of drying measures in flooded buildings. A cost optimization is possible taking into account some simple rules: 1. Good ventilation in wintertime, in summertime no ventilation. Usage of dryers for accelerated short term drying in summertime (close all openings) 3. Determination of duration of drying effort and optimum number of dryers by simulation as function of Material properties; Climatic zone; Wall thickness.. References /1/ Grunewald J., 1997: Konvektiver und diffusiver Stoff- und Energietransport in kapillarporösen Baustoffen. PhD at the TU Dresden, Faculty of Civil Engineering // Funk M., J. Grunewald, : Physical assumptions and derivation of transport coefficients in the thermodynamical model of the Heat and Moisture Transport Simulation Program DELPHIN. Building Research Journal Vol., Number 3-, Slovak Academy of Sciences, Bratislava /3/ Grunewald J., : Implementation of the Material Database of the Moisture Transport Simulation Program DELPHIN. Symposium on: Measurement of water flow in porous material and its numerical simulation, National Research Institute of Cultural Properties, Tokyo, Japan. 1-3 // Grunewald J., 3: Heat-energy-air-moisture performance: COND and DELPHIN. nd International Conference on Building Physics, Workshop on computerized design tools, Leuven, Belgium, Sept. 1-1 // Plagge, R., J. Grunewald, H. Fechner, P. Häupl, : Water retention transfer functions of ceramic bricks of the Dresden building stock. ASHRAE Publisher, Publications