Modeling methods for building-integrated and mixed-mode HVAC systems Timothy Moore ApacheHVAC Product Manager timothy.moore@iesve.com Building Energy Simulation Forum Portland, OR October 2013
Building-integrated and mixed-mode HVAC topics Natural ventilation for fresh air and cooling in mixed-mode systems Thermal displacement and stack-effect ventilation Vented double-skin facades Hydronic radiant cooling slabs Ground-coupled thermal labyrinths Underfloor Air Distribution (UFAD) Thermally stratified high-solar-gain spaces Passive downdraft cool towers
General approach to illustrate methods for presentation Examples given from mix of fictitious and actual building models. Single geometry model or subset it used to perform all analyses within the IES Virtual Environment for each particular topic or project. Thermal, solar pre-sim data, bulk-airflow, HVAC, controls at each time step CFD using boundary conditions from single time step of thermal model
Modeling natural ventilation and mixed-mode operation Bulk-airflow modeling Cross-ventilation Single-sided ventilation Stack-effect ventilation Controlled openings Mixed-mode operation Interaction with mechanical HVAC Displacement ventilation Vented double-skin facades Solar stack pre-heat of intake air Buoyancy-driven earth tubes Passive downdraft cool towers Augment with CFD as needed (but no more) Initial boundary conditions for CFD module
Natural Ventilation: Opening selection and assignment Opening types assigned to fenestration 1 st -floor awning windows + atrium First floor plan Atrium stairwell open to 1 st floor WC Open-plan offices Upper floors have laboratory spaces that cannot have operable windows, and use active chilled beams to address high loads
Natural ventilation: visualization of wind-driven potential
Natural ventilation: visualization of wind-driven potential
Natural ventilation: visualization of wind-driven potential
Natural ventilation: opening selection and assignment Operable opening types are selected or defined by user Exposure Operable area Aerodynamics of opening type Crack flow (when closed) Control of openings schedules, sensors, and formulae Window/door: side hung Window: Center-hung Window: Top-hung Window: Bottom-hung Window: Parallel Window: Sash Louvre Grille Duct (smooth) Acoustic Duct Special opening types needed for intake grills, diffusers, and openings between segments in ducted flow, double-skin facades, earth tubes, etc. Algorithms differ from those for punched openings between larger zones.
Natural ventilation: pre-defined exposure types Wind pressure coefficients
Natural ventilation: opening selection and assignment Opening types assigned to fenestration 1 st -floor awning windows + atrium Atrium stairwell Open-plan offices WC
Natural ventilation: control of opening operation Formula profile for control of window opening Describe BAS control sequence Mimic occupant behavior for manually operated windows Open if and to the extent gt(ta,72,4) & gt(to,56,8) & (to<(ta-2)) Open if room air temperature is greater than 72 F. Ramp opening from closed to fully open over 4 F control band (from 70 74 F). AND Open if outside air temperature is greater than 56 F. Ramp opening from closed to fully open over 8 F control band (from 52 60 F) AND Open if outside air temperature is at least 2 F below room air temperature (BAS version) Occupant version: Open if outside air is not more than 3 F warmer than room air (subject to fuzzy > ramp)
Natural ventilation: mixed-mode HVAC system operation Mixed-mode controls within a prototype VAV system Set up controls to mirror BAS version of operable opening formulae CO 2 and High-temperature overrides re-engage VAV flow Enabling control Atrium Differential temperature sensor
VAV Controls mixed-mode operation
Natural ventilation vs. mechanical HVAC On cooler days, natural ventilation maintains indoor temperatures quite well. As outdoor temperature exceeds indoor cooling setpoint, mechanical HVAC takes over.
Natural ventilation vs. mechanical HVAC Stack flow up through and out of the atrium under ideal conditions is significant. Ideal conditions, however, are not consistently available.
Natural ventilation vs. mechanical HVAC Flow rate and air temperature entering and leaving the cooling coil indicates: Fair number of hours with reduced or eliminated HVAC operation Numerous airside economizer hours Entering cooling coil Leaving cooling coil
Natural ventilation vs. mechanical HVAC Nat vent over initial shoulder-season test period for AHU serving first-floor & atrium reduced chiller energy ~ 13 % reduced fan energy ~ 15% Mixed-mode operation in this test describes means of providing operable windows for users while moderately reducing cooling system energy. Potential improvements More complete evaluation Refined control strategies Climates more conducive than Atlanta, GA!
Vented double-skin façade & mixed HVAC systems Bulk-airflow network used for vented façade. Coupled to HVAC only via heat transfer DV hospital patient rooms on HVAC network Excess air from corridor: Flows into patient room, then WC Extracted via WC exhaust. Bottom of stratified zone Corridor VAV Fully mixed, positively pressurized Adiabatic isolation shell Patient rooms VAV thermal displacement (DV) Air introduced in Occupied zone Air not transferred to WC flows to Stratified zone Patient water closets Transfer and exhaust (no direct supply air) Double-skin façade vented by thermal stack effect, but no air exchange
HVAC system network for airflows other than vented facade Fully mixed corridors, stratified patient rooms, transfer air, and exhausted WC. Occupied zone Corridors WC Stratified zone RA
Vented double-skin façade All surfaces should have thermal properties (absorption, radiation, etc.) All opening should have aerodynamic properties (drag, turbulence, etc.) Façade openings controllable doors Catwalk horizontal shading devices opaque surfaces with doors Solar control and thermal properties of interstitial louver as window + louver Combined solar, thermal, & bulk-airflow simulation of façade interacts with HVAC network via inner façade (wall and fixed solar-control-low-e glass).
Vented double-skin façade exploring results Solar gain for outer vs. inner façade cavities Temperatures resulting from gain and convective heat transfer to each subsequent cavity with vertical airflow
Vented double-skin façade exploring results Solar gain vs. external gain via air from external and internal vent openings
Vented double-skin façade exploring results Flow rates driven by solar gain and sequential façade cavity temperatures Peak buoyancy-driven flow at peak solar gain (below left)
Vented double-skin façade exploring results Outdoor temperature Inside glass surface temperatures Facing into occupied zone Facing into façade cavity Room temperature
Radiant cooling hydronic thermo-active slabs Hydronic radiant cooling slabs can be accurately modeled as a thermal zone, much like a concrete UFAD plenum with negligible air volume, as heat transfer through slab is the main constraint. Tube spacing and depth THERM or similar 2D heat transfer model provides combined U-value for all paths between water and slab surface...
Radiant cooling slabs U-value from THERM is converted to adjusted conductivity for slab material (specific to the tubing type, depth, and spacing).
Radiant cooling slabs Adjusted conductivity is then applied directly to the slab construction. Values for slab top and bottom will differ when tubes are above or below the slab center and when surface boundary conditions differ.
Radiant slab: surface temperature sensors One surface per zone should be tagged for a surface temperature sensor.
Radiant cooling: waterside plant loop Consider Cooling plant, such as water-source heat exchanger, matched to warmer SWT used for radiant cooling slab. Maximize waterside economizer operation via SWT reset and coupling HX to secondary loop return. Cascading loop temp control (not shown)
Radiant cooling: waterside zone hydronic loops In ApacheHVAC the chilled ceiling panel dialog doubles as zone loop. So long as ample capacity at a very small reference temperature difference is provided in this dialog, the chilled ceiling model stays out of the way and allows the hydronic slab model to be constrained only by water flow rate, temperature, and heat transfer through the slab. The number of units can also be multiplied in the zone-level controller. VE users, see Appendix F of the ApacheHVAC for details. EnergyPlus offers a dedicated hydronic slab model, but has limitations related to application and control in combination with airside systems. The concrete slab properties plus loop temperature and flow control are the critical elements of the model
Radiant cooling: airside and assignment of hydronic loops Add chilled slab to pre-defined DOAS + heated & cooled panel system Radiant panels or heated/cooled slabs can serve a space also conditioned by any airside system. Radiant slab zone on each multiplex layer (one for each occupied zone)
Radiant cooling: controls
Radiant cooling: controls Zone loop temperature Track secondary CHWL supply temperature, including reset. Use temperature control to represent zone-level mixing value. Zone loop flow rate, along with temperature, is the critical determination of capacity at the core of the slab. Set this value with some care, and beware that autosizing this can be sketchy, as the thermal capacitance of the slab plays a significant role.
Radiant cooling: controls Time switch controls and midbands for sensor-based on/off control, proportional flow control, and temperature resets can use formula profiles for timing.
Radiant cooling: controls Proportional flow control can use separate sensor e.g., slab surface or in occupied space above/below the slab sensed variable, radiant fraction, etc. Modulating flow rate in the slab according to surface temperature has been shown to work particularly well in spaces with high solar gain, leaving the room temperature sensor to determine when the cooling loop should be off. Logical AND/OR connections in loop controls can reference any other controller.
Slab core and surface temperatures Radiant slabs results include surface temperatures, influence on MRT, thermal comfort analyses, boundary conditions for CFD,
Radiant cooling: thermal comfort Thermal comfort analysis Predicted mean vote (PMV) Room dry resultant temperature operative temperature with still air ( F) Predicted mean vote (PMV) Slab surface temperature chilled slab ( F) Slab core temperature chilled slab ( F)
Radiant cooling: thermal comfort Thermal comfort analysis Percent People Dissatisfied (PPD) PPD (%) 10 8 6 4 >2
Radiant cooling: thermal performance Dry resultant space temperature as simple basis for matching performance.
Radiant cooling: system-level results Comparing plant equipment and energy performance Important to account for difference in pre-cooling potential Cooling Season HVAC System Energy May September Denver, Colorado (TMY climate data) Radiant+DOAS VAV+WSFC-precool VAV with WSFC VAV Baseline 1.93 0.64 0.26 0.21 Estimated savings = 62 to 71% 3.64 4.19 0.49 4.51 4.13 0.36 6.41 3.85 Chillers (VAV only), cooling towers, and chilled water pumps (MWh) Hydronic system pumps and evaporative cooling spray pump (MWh) Fans (including cooling tower fan for waterside free cooling) (MWh) Boilers, natual gas (MWh) 0.36
Ground-coupled thermal labyrinth Labyrinth geometry should be explicitly modeled. Segments provide progressive delta-t. Soil (12 36 ) and then monthly ground temperatures according to depth. Convective heat transfer coefficients are set as appropriate if fan-forced flow. Bulk-airflow model is needed if buoyancy-driven flow, but only HVAC network and thermal model otherwise needed for dynamic simulations. CFD takes boundary conditions from initial dynamic simulation and provides feedback for adjusting heat transfer coefficients until agreement is strong. Labyrinth is then ready for use in thermal comfort and energy modeling.
Ground-coupled thermal labyrinth Labyrinth model provides both tempering and capacitance Temperatures at segments or zones within the labyrinth OA 80.0 F Zone 9 71.7 F exhibits temporal shift relative to OAT profile. Zone 18 66.2 F shifts further relative to OAT profile. Zone 37 62.0 F profile is shifted ~11 hours relative to OAT. Outlet 63.5 F slight heat gain in riser to outlet, but still 16.5 F below OAT. Outlet temperature profile at is nearly flat and is shifted ~12 hours relative to OAT.
Ground-coupled thermal labyrinth Labyrinth zones can be added added to inlet of any HVAC system network. Bulk-airflow model is needed only if flow is driven by wind or buoyancy, as fan-forced flow will typically overwhelm natural ventilation. Labyrinth bypass damper is modulated according to RA and zone temperature sensor-based votes for heating or cooling.
Ground-coupled thermal labyrinth Inlet temperature 78 F at design condition Outlet temperature 60 68 F, depending on conditions for preceding 12 hours Inlet Labyrinth segments or thermal zones plotted
Ground-coupled thermal labyrinth Thermal labyrinth for regional hospital Bi-directional flow (day vs. night back-flush) Alternate seasonal configurations for inlet, bypass, flow direction, etc.
UFAD plenum model Heat transfer in UFAD systems Heat loss from return plenum (gain into supply plenum above ) 10-15% From Total system heat gain 100% Room air extraction 60-70% (RCLR) Ceiling-slab radiation Ceiling-floor radiation Through ceiling To To 65 F Through slab Through floor Heat gain into supply plenum 35-45% through floor deck and raised floor UC Berkeley Center for the Build Environment (CBE) 2007
UFAD plenum and stratified zones
UFAD airside network
UFAD temperature and RH control
UFAD results for peak cooling days Afternoon cooling peak Peak OA temp: 95 F SAT leaving AHU: 60 F UFAD SA temp: 66 F Occ. Space temp: 76 F Stratified zone: 79.2 F RA plenum: 77.5 F Weekend run with raised temp setpoint as hot soak before weekday cooling operation.
Façade-integrated passive conditioned nat-vent UFAD system Cascading airflow Ventilation driven by pressure differentials Wind-driven Stack-effect +/- HVAC pressure BAS control of facade ventilation openings Mixed-mode operation MacroFlo runs at every simulation time step, as does ApacheHVAC Thermal stratification UFAD and thermal displacement ventilation Secondary air outlet Primary ventilation air intake duct to underfloor (UFAD) plenum User controlled operable vent Primary air outlet to stack vent in building core
Façade-integrated passive conditioned nat-vent UFAD system MacroFlo results can inform design and control of facade elements
Passive downdraft cool towers ApacheHVAC for cooling coils or evaporative cooling and heating coils Fans are optional for assist, but not required Example model is fully buoyancy driven (no fans) Buoyancy-driven air movement through building modeled via MacroFlo
Passive downdraft cool towers Two downdraft towers: both have controllable inlets; one has wind baffles. UFAD supply air plenums 1 st floor provides example of controlled vent/cooling diffusers; heat by room units. 2 nd floor uses controlled inlets from towers to plenum; heat by coils in plenum. Occupied & stratified zones with typical internal gains RA plenums discharge to atrium Outlets placed high on atrium façades controlled to open only on downwind side
Passive downdraft cool towers Complete network with all zones and nodes included. Facilitated adding damper sets and heating coils where needed.
Passive downdraft cool towers Minimal network reduced to just the necessary zones and nodes.
Passive downdraft cool towers Confirming downdraft flow between cells in cool towers Passive downdraft outflow from cooling coil cell of the tower (North and South towers shown) to the next cell or tower segment below.
Passive downdraft cool towers Confirming flow into the towers and the majority of the flow down from there 964 l/s 2042 cfm Passive downdraft outflow from cooling coil cell of the tower (North and South towers shown) to the next cell or tower segment below.
Passive downdraft cool towers Confirming flow at controlled floor diffusers Flow at two controlled discharge dampers in the UFAD supply plenum (controlled diffusers in MacroFlo are doors with formula profiles).
Passive downdraft cool towers Confirming flow through inlet dampers on 2 nd -floor UFAD plenum Flow into the 2 nd -floor UFAD plenum is controlled by dampers at the connection to the cool towers so that plenum can be heated in winter.
Passive downdraft cool towers Maintaining cooling setpoint (23 C +/- 1 C) in summer Temperature in occupied zones (green, blue, and red lines) vs. outdoor temperature (light teal Brisbane, AU, Mar 4-6).
Passive downdraft cool towers Shoulder season performance of 2 nd -floor occupied zone Occupied zone temperature (green) is maintained by reducing flow from cool tower (light blue) to vent only and engaging floor plenum heating coil
Passive downdraft cool towers Thermal stratification and temperatures on flow path: 1 st -floor occupied zone Control of both coil LAT in towers and passive flow via dampers results in very tightly controlled space temperature and avoid excessive cooling.
Passive downdraft cool towers Thermal stratification and temperatures on flow path: 2 nd -floor occupied zone Temperature in 2 nd -floor occupied zone is slightly more variable as a result of heat transfer from RA below through floor deck into UFAD plenum.
Passive downdraft cool towers Thermal stratification and temperatures on flow path: atrium Atrium occupied zone temperature is very consistent in spite of substantial solar gain as well as associated gain (1 2 C) in atrium floor plenum.
Bulk airflow vs. CFD for air movement, buoyancy, and comfort CFD can be very useful where bulk airflow modeling leaves off Bulk-airflow model driven by temperature difference for adjacent fully mixed zones. No wind-driven or stack-effect flows within a zone (only at openings between). No thermal plumes adding to overhead pool of hot air. Use CFD to predict air movement and associated local thermal comfort conditions. Where temp gradients are significant, use zonal method for energy model. Use CFD to calibrate or tune the zonal bulk-airflow model
Atrium analysis example: CFD model used to tune zonal model Single volume (no zonal subdivisions) Initial boundary conditions for each test case taken from selected time step in zonal thermal and HVAC model Initial DV airflow and temperature at large side-wall diffuser set to the same value used in the zonal thermal and HVAC model Airflow from DV diffuser on back wall Occupied zone Temperature at center of occupied zone was basis for comparison of airflow requirements at a given supply air temperature.
Zonal model of atrium for bulk-airflow and HVAC network Subdivision of atrium into 10 zones: 6-in deep pool of cooling air (SA introduced here) 7.5-ft Occupied zone above the cooling air pool Upper stratified zone Lower stratified zone 28-ft Lower stratified zone 4-ft Upper stratified zone against the ceiling Occupied zone 6-in deep Façade zones for each orientation Matching height of occupied and lower stratified zones Concentrated zone of convective heating at glass surfaces Façade zones Floor-level pool of DV cooling air
Atrium HVAC system model Thermal displacement ventilation system airside network (ApacheHVAC)
Atrium zone temperatures: Zonal method (dynamic HVAC model) Upper stratified zone Lower stratified zone Occupied zone Floor-level SA pool Supply air at diffuser Zone temperatures for the zonal modeling method: 10-zone model with bulk-airflow network.
Modeling methods for building-integrated and mixed-mode HVAC systems Questions? Timothy Moore Senior Product Manager ApacheHVAC Integrated Environmental Solutions timothy.moore@iesve.com