HW 3: Energy-Based Systems Modeling in Modelica Design of an Exterior Shading System for Multistory Buildings Yuming Sun Brent Weigel ME 6105 Modeling & Simulation Design Dr. Chris Paredis October 28, 2010 Task 1: Define your goals and problem domain The simulation is designed for the following problem domain: designing an outrigger exterior shading system for minimizing total annual heating and cooling energy consumption of a perimeter space in a multistory building. Our selected design context is a south-facing exterior wall of a perimeter space in a multistory building located in Atlanta, GA. Figure 1 and Figure 2 below show and a rendering and a layout drawing (respectively) of the perimeter space design context. Figure 1: Rendering of the perimeter space design context. 1 Figure 2: Elevation view (top) and plan view (bottom) of the perimeter space design context. The context of the perimeter space is considered to be representative of a modern exterior office. The exterior wall is comprised primarily of glazing, for which the outrigger shading system is designed to control the passage of solar irradiance (solar heat gain). An exterior shading system minimizes heating and cooling energy consumption by maximizing solar heat gain during heating and minimizing solar heat gain during cooling. Figure 3 shows a three-dimensional perspective of the perimeter space served by the shading device (ceiling plenum not shown). 2 Figure 3: Three-dimensional perspective of space served by shading device (dimensions in mm). The overall design context of our simulation is similar to our original proposal. One significant change to our design context is a reduction in the number of design variable for our outrigger shading device. Since our design context contains many inherent (and uncertain) variables (T-stat setpoints, room occupancy, lighting schedule, window Uvalue, etc.) we have reduced the number of shading device design parameters from 5 to 2 (outrigger depth and outrigger height above window). This revision does not oversimplify the design context or the simulation effort, since many dimensions of uncertainty remain. Additionaly, our proposed simulation design has been revised to include not only the shading and envelope systems, but also the heating and cooling (HVAC) system serving the perimeter space (or zone). The HVAC system is now included in the simulation design for the purpose of estimating space cooling and heating energy consumption (as opposed to estimating merely the solar heat gain entering the perimeter space). 3 The goal of our simulation model is to determine an outrigger design (outrigger design parameter values) that minimizes total annual energy consumption for heating and cooling a perimeter space in a multistory building. To do so, our simulation must be able to address several questions: What is the total HVAC energy consumption and cost? (inclusion of cost based on comments on initial proposal) To minimize space energy (or cost): o What should the outrigger depth be? o What should the outrigger distance above the top of the window be? How does uncertainty in space conditions (temperature setpoints and internal heat gain) and envelope performance (wall thermal resistance, glazing thermal resistance, glazing area, glazing SHGC, etc.) affect the outrigger design? To answer these questions, our simulation must provide a reasonably accurate representation of the thermal energy dynamics of a perimeter space in a multistory building. Thus, our simulation must represent the relevant components and component interactions of the thermal energy system in our design context. Task 2: System and Simulation Specification Our Modelica model is based on a characterization of the thermal energy system, which contains many interacting components. In the interest of answering the design questions listed in the previous section, the simulation model must at least represent the design parameters of the shading device and the operation of the HVAC system serving the perimeter space. In order to relate the shading device design parameters to the operation (energy consumption) of the HVAC system, the simulation model must also represent the relevant thermal energy components (external energy loads, material properties, and geometry). Figure 4 below shows a schematic of the heat flows and temperature nodes of the system components included in the system simulation model. 4 PLAN VIEW Q_solar_rad Q_infil/ exfil Shading Device Q_solar_rad _shade_wall T_room_surface Q_solar_ glaz_rad T_amb Q_glaz_ext _conv T_glaz_ext Q_solar_room _rad T_glaz_int Q_glaz_cond Q_surface_conv Q_glaz_int_conv T_room_air Q_cooling Q_wall_int_conv Q_wall_ext_conv Q_internal_conv T_wall_ext T_wall_int Q_heating Q_wall_cond Figure 4: Simplified schematic of heat flows and temperature nodes of system components in the simulation model. In the above figure, the energy consumed by the HVAC system is represented by the heat flows Q_cooling and Q_heating. Although actual HVAC energy performance is dependent upon many building mechanical system parameters (fan efficiency, duct heat loss/gain, air mixing/stratification, vapor-compression efficiency, etc.), the cooling and heating flows may be modeled as prescribed heat flows to and from the room air, T_room_air. For our cooling and heating system, we assume a conservative COP and heating efficiency of 2.8 and 85% respectively. Furthermore, in our simulation of the thermal dynamics of the perimeter space, we assume a uniform room air temperature. The room air temperature is affected by heat transfer with several other system components. One of these components is internal heat gain from equipment, lighting, and occupants of the space, Q_internal_conv. This total convective heat load is estimated to be 36.7 W/m2 and follows a building occupancy schedule of 8:00 am – 6:00 pm, Monday – Friday. The room air temperature is also affected by the flow of heat through the building envelope, to/from T_amb. This heat flow is modeled through several parallel pathways: 1) Conduction through the opaque envelope (Q_wall_ext_conv Q_wall_cond Q_wall_int_conv); 5 2) Conduction through the envelope glazing (Q_glaz_ext_conv Q_glaz_cond Q_glaz_int_conv); 3) Long wave radiation through the envelope glazing to the room interior, and convection to the room air (Q_solar_rad Q_solar_room_rad Q_surface_conv); 4) Infiltration/exfiltration through the envelope, Q_infil/exfil; Heat pathways 1, 2, and 3 are affected by the shading performance of the outrigger shading device. The shading device influences the area of solar irradiance on the exterior glazing surface, which in-turn influences the solar irradiance on the interior room surfaces. The short wave radiation on these surfaces ultimately affects the convective heat flow to/from the room air, which in turn affects the heating and cooling energy needed to serve the space. In addition to those previously stated, the energy system simulation includes many assumptions and abstractions which are necessary for developing a model that captures the relevant phenomena/dynamics in a feasible manner (within the limits of available time and available expertise for the modeling effort). One assumption is that the solar radiation is absorbed by the interior room surfaces in the following proportions: 30 percent on the floor, 30 percent on the furniture, 20 percent on the east partition, and 20 percent on the west partition. This assumption simplifies the complex estimation of room surface temperatures, and the assumption is consistent with most commercial building energy simulation models. The simulation model includes heat flows through the floor, ceiling, and wall partitions (not shown in Figure 4). The perimeter space context is assumed to be symmetrical (with respect to materials, geometry, and space conditions) with all adjoining spaces). Thus, whatever heat flows from a partition surface to an adjacent space is assumed to flow through the opposite partition. This heat flow is included to model the energy balance of the room surface temperature nodes. Another assumption is that all heat flow within the room is either conductive or convective (no long wave radiation between surfaces). With respect to the weather conditions, it is assumed that the convection coefficient on the envelope exterior is constant (does not vary with wind speed or direction). Furthermore, we assume that the solar heat gain coefficient (SHGC) of the glazing is constant (does not change with solar angle). With respect to building material components, the estimated thermal resistances and heat capacitances are assumed to be consistent with typical, code-compliant commercial building construction (see Appendix A). 6 Task 3: Create your models in Dymola Our simulation of the energy performance of an outrigger shading device is accomplished by incorporating the aforementioned components and relationships into Dymola. The simulation consists of several connected models that represent unique portions of the energy system. One portion of the system is the opaque exterior wall, represented in Figure 5. U_? k=8? k=3? U_? U_? k=0? thermalCond? Exterior_? Interior co? co? G=G heatCapaci? heatCapacitor 1231? 1231? Exterior_? Figure 5: Dymola schematic of exterior wall model. The exterior wall model represents the thermal conductance (U-value), heat capacitance, surface convection, and solar radiation on the exterior surface. The building exterior ports exchange heat with the ambient air temperature and the solar radiation, respectively, and the interior port exchanges heat with the room air. 7 The window model (see Figure 6) is similar to the exterior wall model, except that it utilizes a different U-value and heat capacitance. The SHGC of the window is applied outside of the window model, along the path of solar, short wave radiation heat transfer. U_? k=8? k=3? U_? U_? k=2? thermalCond? WinExteri? co? WinInte? co? G=G heatCapacitor heatCapaci? 81648 81648 WinExteri? Figure 6: Dymola schematic of window model. The SHGC is applied outside of the window model, along the path of short wave radiation heat transfer (see solar radiation model, Figure 14). The building envelope consists of not only the window area and opaque wall area in contact with the room air, but also the opaque wall in contact with the ceiling plenum. A ceiling model was developed to account for heat transfer between the exterior and the ceiling plenum, and between the ceiling plenum and the room air (see Figure 7). 8 T_amb Exterior_? const2 k=3.6336 k=210 const1 Slab FixRoom? G=89.376 heatCapaci? heatCapacitor 2728? 2728? convection3 const3 Plenum V=20 k=210 Ceiling exterior_? G=840 heatCapaci? heatCapaci? 684902 684902 convection4 Figure 7: Dymola schematic of ceiling model. The ceiling model contains the wall model for heat exchange between the plenum air and the building exterior (both conduction and solar radiation). The floor model is the same is identical to the ceiling model. To properly account for the wall partition surface temperatures and the subsequent convective heat transfer to the room air temperature, models of heat conduction through the wall partitions are included (see Figure 8). 9 Partition port_a1 port_b1 G=48.2 heatCapaci? heatCapacitor 492591 492591 Figure 8: Dymola schematic of partition model. In the wall partition models, the convection coefficients are external to the model, so that the short wave radiation heat flow, convection heat flow, and symmetrical heat flow (heat flowing out of west partition connected to heat flowing into east partition) can be connected separately. The furniture model shown below in Figure 9 is included to account for the heat capacitance of the room interior objects. Furniture Exterior Interior G=90 heatCapaci? heatCapaci? 3000? 3000? Figure 9: Dymola schematic of furniture model. The room interior, partition, and envelope components are combined into a composite envelope and interior model (see Figure 10). 10 pr? Wal? T_Ambient CEI? k=8? pr? WA? K tem? K con? pr? Win? k=1? k=210 tem? W? co? Total_SR_w in_South K K pre? K con? pr? ROOM_AIR k=147 co? k=210 k=420 Total_SR_intoRoom PAR? con? PAR? co? co? co? tem? Fur? K co? pre? k=147 K con? Total_SR_Plen_Wall tem? R? k=? R? Total_SR_Wall_South pre? K pre? K pr? pr? k=? FL? Figure 10: Dymola schematic of envelope and interior model. The envelope and interior model represents the flow of conductive, convective, and radiative heat across the room boundaries and into the room air. The room air is also connected to heat from the internal heat gain model, shown in Figure 11. The internal heat gain model represents the generation of heat from lighting, people, and equipment, and the heat is generated according to a common occupancy schedule. 11 Lighting_load period=86400 Occupancy add3_1 +1 +1 prescribed? Internal_? + +1 period=86400 Equipment period=86400 Figure 11: Dymola schematic of internal heat gain model. Additionally, a model of the air infiltration and ventilation load has been included (see Figure 12). The model adds heating or cooling directly to the room air based on the difference between the room air and outdoor air temperature, an ASHRAE 62.1 ventilation rate, and an assumed air infiltration rate. 12 Ventilation add1 +1 period=86400 Infiltration + +1 product Ambient_Tem? k=5.72 prescribed? Ventilatio? add +1 + offset=0 -1 Room_Air K tem? Figure 12: Infiltration and ventilation model Estimation of the heating and cooling energy is accomplished in an HVAC system model (see Figure 13). 13 Heating_Energy division Heating_cost I booleanToReal1 B prescribed? con? k=1 k=6? R Cooling_Energy divisi? Cooling_Cost I booleanToReal B prescribed? con? k=-1 k=1? R port1 2? K 2? temperatu? T_Stat_Cool u k=294.15 refe? T_Stat_Heat const Figure 13: Dymola schematic of HVAC system (heating and cooling cost). The HVAC system model includes temperature control setpoints for both heating and cooling. The cooling capacity is set at 1,000 W (approximately 0.3 tons of cooling) and the heating capacity is set at 1,000 W. The division constants account for the assumed efficiencies, and the energy rates to calculate the heating and cooling cost. The solar load on the room is estimated by the solar radiation model shown in Figure 14. 14 shadin? solar_P? COS_T? product1 add3_1 +1 product +1 Absor? +1 timeTable Direct_SR offset=0 offset=0 k=0.05 SHGC_? Sky_Diffuse_? offset=0 Horizontal_To? SR_WIn + SR_IntoRoom k=0.636 VF Area_w in k=0.35 k=13.5 Albedo Area_w in1 Area_plen_w all Abso? SR_plen_w all k=0.7 offset=0 k=0.2 VF2 k=13.5 add3_2 +1 +1 k=3.6336 Area_w all Absor? SR_w all + +1 k=0.7 k=8.0664 k=0.5 Figure 14: Dymola schematic of solar radiation. The details of the solar radiation model are illustrated in the verification portion of this report (Task 4). The solar radiation model contains the shading device model and it connects to the envelop and interior model (see Figure 10). The complete arrangement of component models is shown below in Figure 15. From this complete mode we may estimate the total heating and cooling cost associated with the shading device design. 15 Total_Cost add +1 + +1 Ambient_Tem? volume hVAC offset=0 V=80 Enve? Solar? internal? ventilati? Figure 15: Dymola schematic of combined energy-based system models. Task 4: Verification During the development of the simulation and model components, the model components were tested individually and in combination, so that the proper function of the components could be verified. 1. Solar Position Our simulation is designed to use a large set of input data (hourly weather data for a typical meteorological year) to calculate the thermal loads on a perimeter office space. In order to evaluate the performance of the outrigger shading device, it is essential for the simulation to convert the solar radiation weather data into radiation data that accounts for the direction of the solar radiation (and the subsequent projected shaded area). The conversion requires calculation of the solar position, a process that is performed by a component of the solar radiation model (see upper left-hand corner of Figure 14). The sun’s position in the sky is expressed in terms of the solar altitude angle, β, above the horizontal and the solar azimuth angle, , measured from the south pole. The angle of incidence θ for any surface is defined as the angle between the incoming solar rays and a 16 line normal to that surface, and θ changes throughout the day as the sun follows its ecliptic through the sky (see Figure 16). These angles are used to determine sunrise and sunset time, and solar radiation intensity on surface. Figure 16: Solar angles for vertical and horizontal surfaces. Source: 2005 ASHRAE Handbook—Fundamentals, SI Edition, Chapter 31. Our solar position model in Dymola is programmed to calculate the solar position (solar attitude angle β and azimuth ) from city latitude, longitude and time zone parameters, and a time table of the local time. The expected result for solar position is shown in Figure 17, which is derived from a previous research project involving solar loading and shading. 17 Figure 17: Hourly solar altitude and azimuth angle for Atlanta, GA on January 1st. The solar position model was tested to verify that it produces the expected solar position. Figure 18 below shows the solar position results produced from the Dymola model. 18 Figure 18: Continuous solar altitude and azimuth angle for Atlanta, GA from January 1st to 6th. The test results indicate that the solar position model provides an appropriate calculation of solar position. According to the software documentation, Dymola performs a linear interpolation of the hourly input data to calculate a more continuous solar position (horizontal axis of Figure 18 is displayed in seconds) 2. Solar Radiation Solar radiation is a major part of our shading device simulation, so it is important that the intensity of solar radiation is modeled correctly. The total short-wavelength irradiance Et reaching a surface is the sum of the direct solar radiation ED, the diffuse sky radiation Ed, and the solar radiation Er reflected from surrounding surfaces. The irradiance on the fenestration aperture (window area) of the direct beam component ED is the product of the direct normal irradiation EDN and the cosine of the angle of incidence θ between the incoming solar rays and a line normal (perpendicular) to the surface: Et= EDN*cos θ + VFsky-surface * Ed + Er Solar radiation EDN and Ed are read from TMY data. The view factor (VF) between sky and window is a fixed value (0.35) in the solar radiation model. 19 Similar to the verification of the solar position model, the solar radiation model (see Figure 14) was tested to see if the results match expectations. Figure 19: Test of solar radiation model component from January 1st to January 6th. The results show a daily fluctuation in the total solar irradiance, which is the result of the estimated variation in weather conditions in the TMY data. The relative proportion of solar radiation allocated to the window, room, and wall is determined by the relative areas of the surfaces, the absorptance of the window and wall material, and the SHGC of the window. The proportions shown in the test results appear reasonable since the wall has a significantly higher absorptance than the window. The solar radiation into the room is considerable since the window glazing area dominates the envelope area in our simulation model. The above results were further verified with hand calculations, which confirmed the values shown. 3. Shading Device The outrigger shading device, which is the focus of this design simulation, is designed to be installed above the window exterior, so that it may decrease the solar radiation entering the room by forming a shaded area on the window. The sunlit area (shaded area) calculation is the function of this shading device component. The input parameters for this model are the geometry of the shading device, which will eventually be optimized to minimize heating and cooling energy consumption. Our shading device model 20 component was verified by comparing our model results to hourly calculation results produced from a shading device design tool developed for KAWNEER. The design tool results and Dymola model results are shown in Figure 20 and respectively. Figure 20: KAWNEER design tool sunlit area results. 21 Figure 21: Shading device model sunlit area test results. The test results for sunlit area are very similar to the results produced by the commercial design tool. The variation in the vertical portions of the line plots can be explained by variations in linear interpolation of the solar radiation TMY data. 4. HVAC Component Estimation of the heating and cooling energy consumption associated with the shading device design is based on the operation of an HVAC system (see Figure 13). Thus, it is important that the HVAC system model be tested and verified. Figure 22 shows the schematic of a simple test of the HVAC system model. 22 Figure 22: Simple test of HVAC system model. A prescribed heat flux connected to a sinusoidal input was used to mimic the heating and cooling load. To maintain room air temperature in the comfortable range, the HVAC system is supposed to provide cooling and heating to the room to offset heat flux into and out of the space (in response to room air temperature). We can observe this phenomena in the following test results (see Figure 23). 23 Figure 23: HVAC system model test results The heating and cooling energy consumption alternates (never occur simultaneously), and there exists some time delay compared to the input signal, which is due to the heat capacity of the air. Task 5: Experimentation and Interpretation Following the verification of the model components, we have run our complete simulation model under multiple design scenarios (see Table 1). Each of these scenarios represent a change in various input parameter values. The purpose of these scenario experiments is to gain a sense for how much each of the input parameters influences the total energy consumption of the system. Due to the high time cost of running each of the scenarios for a complete TMY year, the scenarios were evaluated for the first 3 months of the year (7,883,100 seconds). Our results are therefore based on the heating season, but due to the high internal heat gain, there still exists significant cooling during this period. 24 Table 1: Model Experimentation Scenarios Scenarios Parameters Changes Exterior Wall Insulation Thickness 90.2 mm --> 20mm Thermal Conductivity (U Value) (W/m2-K) 0.371 --> 0.932 Glazing Type Double Low-E --> Double Clear Thermal Conductivity (U Value) (W/m2-K) 2.47 --> 3.226 SHGC 0.636 --> 0.758 3 Lighting Schedual 8 am - 6pm ---> 8am-10am & 4pm-6pm 4 Ventilation (L/S/Person) 8 ---> 12 5 HVAC Setpoint Temperature (︒C) 1 2 Cooling: 24 --> 26 Heating: 21 --> 19 6 Shading Device Outrigger Depth (M) 1 --> 1.5 7 Shading Device Outrigger Depth (M) 1 --> 2 8 Distance between Shading Device and window top (M) 0 --> 0.5 9 Distance between Shading Device and window top (M) 0 --> 1 The experimentation scenarios are discusses in each of the sections that follow. Following these sections is a summary of the impact on cooling and heating cost. 1. Exterior Wall Insulation Thickness In our first experiment scenario, we explore how exterior wall insulation will affect envelop cooling and heating load and HVAC energy consumption. Figure 24 and Figure 25 below show a comparison of the simulation results for the two wall insulation thicknesses. 25 Figure 24: Heat flow through exterior wall. 26 90.2 mm Insulation 12 10 8 6 4 2 0 -2 0E0 50 20 mm Insulation hVAC.Heating_cost 1E6 2E6 hVAC.Cooling_Cost hVAC.Heating_cost 3E6 4E6 5E6 6E6 7E6 8E6 5E6 6E6 7E6 8E6 hVAC.Cooling_Cost 40 30 20 10 0 -10 0E0 1E6 2E6 3E6 4E6 Figure 25: Heating and cooling cost Figure 24 shows heat flow through south wall. As is expected, the better insulated wall will decrease heat flow both for heating and cooling season. HVAC energy cost results (Figure 25) tell us that using higher insulation will decrease heating energy consumption, but for cooling energy consumption, it may either decrease or increase which depending on other cooling load sources. A room will need cooling because of higher internal load or solar radiation when outside is actually cooler, this sometimes happens in commercial buildings (known as exceeding the “balance point”). For this room, we can see better insulation actually results in slightly higher cooling energy consumption. 2. Glazing Type Next, we compare the simulation results in terms of solar radiation, conduction and HVAC energy consumption by using double clear glazing instead of double low-e glazing. 27 solar_radiation.SR_IntoRoom solar_radiation.SR_IntoRoom 3.0E5 7.0E5 9000 8000 7000 6000 5000 4000 3000 2000 1000 0 2.0E5 4.0E5 5.0E5 6.0E5 8.0E5 9.0E5 1.0E6 1.1E6 1.2E6 Figure 26: Solar radiation into room. Figure 26 shows that solar radiation into the room increases by using double clear glazing instead of double low-e glazing, because double low-e glazing has a lower SHGC (Solar Heat Gain Coefficient) by coating on the surface. 200 envelop_Load.window.thermalConductor.Q_flow [W] envelop_Load.window.thermalConductor.Q_flow [W] 0 -200 -400 -600 -800 -1000 0E0 1E6 2E6 3E6 4E6 5E6 6E6 7E6 8E6 Figure 27: Heat flow through window. Figure 27 indicates that by using low-e glazing, the heat flow is lower than when using clear glazing both for heating and cooling season, because low-e glazing has a higher thermal resistance. 28 3. Lighting Schedule Lighting is a major part of not only for building electricity consumption but space cooling load. We want to explore how much energy will be saved by turning off the light from 10 Am to 3Pm, when natural light is most likely to meet indoor lighting requirement. internalLoad.prescribedHeatFlow.Q_flow internalLoad.prescribedHeatFlow.Q_flow 1600 1400 1200 1000 800 600 400 200 0 0E0 1E5 2E5 3E5 4E5 5E5 6E5 7E5 8E5 Figure 28: Internal load. Figure 28 shows that the internal load decreases during the time when the lights are turned off, and when peak cooling load is also most likely to occur for a south facing room. 4. Ventilation Space ventilation is essential for creating an inhabitable indoor air quality by bringing fresh air into the room. It can cause either cooling or heating load which depends on air entropy difference between inside and outside. During our simulation period from January to April, ventilation will always bring cooling to the room. It will increase heating energy consumption but also save some cooling energy. 29 ventilation_Infiltration.product.y ventilation_Infiltration.product.y 400 0 -400 -800 -1200 -1600 -2000 -2400 -2800 0E0 1E6 2E6 3E6 4E6 5E6 6E6 7E6 8E6 Figure 29: Ventilation and infiltration load. The results in Figure 29 show that an increase in the ventilation load significantly increases the cooling effect of the ventilation component of the simulation model. 5. HVAC Setpoint Temperature In this scenario, we evaluate the impact of widening the gap between the heating and cooling HVAC temperature control (T-stat) setpoints. Figure 30 below shows the resulting impact on total energy cost. 30 Total_Cost Total_Cost 50 40 30 20 10 0 0E0 1E6 2E6 3E6 4E6 5E6 6E6 7E6 8E6 Figure 30: Relative reduction in total HVAC system energy cost due to a widening of T-stat setpoints. After reviewing these results and taking conventional building operation practices into consideration, we have determined that the wider T-stat setpoint range is perhaps more appropriate for our shading device design simulation. 6. Outrigger Depth Increase By increasing the outrigger shading device depth, we expect that the solar radiation on the window will decrease. Figure 31 below shows the results for increasing the outrigger depth. 31 envelop_Load.window.WinExterior_Rad.Q_flow 700 envelop_Load.window.WinExterior_Rad.Q_flow 600 500 400 300 200 100 0 -100 0E0 1E6 2E6 3E6 4E6 5E6 6E6 7E6 8E6 Figure 31: Reduction in exterior window radiation due to increased outrigger depth. The results indicate that the solar radiation load on the window is indeed reduced by increasing the depth of the shading device. 7. Outrigger Depth Increase (additional) By increasing the outrigger depth further, we should expect to see a reduction in the cooling load for the HVAC system (see Figure 32 below). 32 hVAC.Cooling_Cost 45 hVAC.Cooling_Cost 40 35 30 25 20 15 10 5 0 -5 0E0 1E6 2E6 3E6 4E6 5E6 6E6 7E6 8E6 Figure 32: Comparison of cooling cost for increased outrigger depth. The results of this experiment confirm our expectation of reduced energy costs. For a full simulation year, we would expect the cooling cost reduction to be more pronounced during the summer months. 8. Increased Shading Device Height The shading device height above the top of the window is expected to have some impact (positive) on the solar radiation entering the room, but not the same degree of impact as the outrigger depth. This expectation is based on the fact that for a south-facing wall, the solar angle is greater than 45 degrees during the most intense periods of direct solar radiation. In this scenario, the device depth has more influence over the shaded area than does the device height (assuming a shaded area projected on a vertical surface). Figure 33 shows the results of our simulation experiment for increased device height. 33 envelop_Load.window.thermalConductor.Q_flow 200 envelop_Load.window.thermalConductor.Q_flow 0 -200 -400 -600 -800 0E0 1E6 2E6 3E6 4E6 5E6 6E6 7E6 8E6 Figure 33: Comparison of window solar radiation for an increase in outrigger height. The simulation results indicate that the solar radiation on the window is nearly identical for device heights of 0 and 0.5 m above the top of the window. These results indicate that an alternative shading device parameter may need to be included in our design study. 9. Increased Shading Device Height (additional) A further increase in the height of the shading device above the window was tested and the results are shown below in Figure 34. 34 hVAC.Cooling_Cost 45 hVAC.Cooling_Cost 40 35 30 25 20 15 10 5 0 -5 0E0 1E6 2E6 3E6 4E6 5E6 6E6 7E6 8E6 Figure 34: Comparison of cooling cost for increased outrigger depth (0 to 1m). The simulation results compare device heights of 0 and 1 m, which is nearly the practical maximum increase for typical building plenum heights. Since the results are nearly identical, we conclude that a different device design parameter will need to be included in our design study. Due to project time constraints, we have not been able to modify our shading device model component to include an alternate design parameter/variable. Our intent is to include window U-value and window SHGC as design variables for our study. These design parameters, in combination with shading device depth, are directly related to the thermal performance of the fenestration system. Thus, these parameters should be considered together by the decision maker (architect) when designing the fenestration system. Experimentation Summary The estimated heating, cooling, and total energy costs for each of the simulation scenarios are shown below in Figure 35. 35 9 8 7 Scenario 6 Total Cost 5 Heating Energy Cost 4 Cooling Energy Cost 3 2 1 0 0 10 20 30 Cost, $ 40 50 60 Figure 35: Comparison of heating, cooling, and total costs for the experimentation scenarios. Scenario “0” represents the original base design. The scenarios illustrate how the building design and operation context produces tradeoffs in heating and cooling energy cost. Furthermore, the total energy consumed or saved by changes in shading device design parameters (Scenarios 6 through 9) may be offset by changes in the design context (Scenarios 1 through 5). Even though our comparison scenarios are limited to the first few months of the year, we see that the depth of the outrigger has a considerable impact on cooling energy cost. For a longer simulation year that includes the cooling season, the impact is expected to be greater. Task 6: Lessons learned The experience of developing and testing a Dymola model of a shading device for a commercial office building presented many challenges. Heat transfer in building seems is conceptually simple, but in terms of the number and variety of interactions, it is considerably complex. The dymola software supported modular development of the building model components, but the connections between these components quickly became cumbersome. In one way, our model is perhaps overly complex, in that it includes heat transfer between opposite partitions and floor/ceiling sections (symmetrical 36 heat transfer to/from the space). After investing the time to develop the model in this manner, not enough time was available to compare the model results with and without partitions, floors, and ceilings. If we were to create another simulation model, we would have started with a simpler model of the room envelope. If we had more time we would like to include more sophistication and operational complexity in the HVAC system model, particularly for cooling operation. Our current model measures cooling cost during the winter months, whereas in reality, the cooling would be accomplished by “free cooling” in economizer mode (cold ventilation air would be used instead of vapor compression cooling). Despite the modular, object-oriented architecture of Dymola and the Modelica language, we were unfortunately not able to locate/utilize an existing building simulation library. Consequently, we had to build from scratch many model components that have likely been created by hundreds of other Modelica users. Even though we both have a solid background of building design and simulation principles, we also went through a hard time in training ourselves to think from the perspective of Modelica and simplified, lumpparamter modeling. Overall, we have improved our knowledge of the behavior of the system we modeled. By building a network of the main thermal components, we have improved our understanding of heat transfer in a complex system, and how those components interact with each other. Furthermore, by using Dymola, we have learned how the Modelica, object-oriented language actually works in terms of an operating modeling. 37 Appendix A: Envelope and Space Inputs Exterior Wall Area M^2 U W/m2-K G W/K Solar Absorptance 11.7 0.371 4.3407 0.7 Brick Density Thickness Volume Capacity C Total C/2 kg/m3 mm M^3 J/kgK J/K J/k Wood 1922.2 101.6 1.18872 836.8 1912052.506 1231150.741 Ho Hi Area*Ho Area*Hi 20 W/m2-K 5 W/m2-K MinwoolBatt Gypboard 592.7 25.4 0.29718 2510.4 442178.3063 9.61 90.2 1.05534 836.8 8486.6728 234 58.5 800.9 12.7 0.14859 836.8 99583.9957 Partition Wall Area U G M^2 W/m2-K W/K 29.4 1.639 48.1866 Hi Area*Hi Gypboard Density Thickness Volume Capacity C Total C/2 kg/m3 mm M^3 J/kgK J/K J/k 800.9 25 0.735 836.8 492591.9432 492591.9432 Gypboard 800.9 25 0.735 836.8 492591.9432 38 5 W/m2-K 147 Ceiling and Floor Area U G Plywood+Concrete Plasterboard M^2 42 42 Hi W/m2-K 2.128 20 W/K 89.376 840 Area*Hi Plywood Density Thickness Volume Capacity C Total C/2 Total C/2_Plasterboard kg/m3 mm M^3 J/kgK J/K J/k J/k 700 10 0.42 1420 417480 2728740 684902.4 5 W/m2-K 210 Concrete-Light Plasterboard 1200 2800 100 13 4.2 0.546 1000 896 5040000 1369804.8 Window: Dbl LoE (e3=.1) Clr 3mm/6mm Air Area U G SHGC M^2 W/m2-K W/K Density Thickness Volume Capacity C Total C/2 kg/m3 mm M^3 J/kgK J/K J/k 13.5 2.47 33.345 0.636 Ho Hi Area*Ho Area*Hi Generic Clear Pane Generic Low_E Clear Pane 2400 2400 3 3 0.0405 0.0405 840 840 81648 81648 81648 39 20 W/m2-K 5 W/m2-K 270 67.5 Occupancy Total 6 Person 110W/person 660 W ASHRAE Equipment Total 9 W/m^2 378 W 42 M^2 Lighting Total 12 W/m^2 504 W 42 M^2 Ventilation Min Total 8 L/s/Person 48 L/s 0.048 M^3/s Air Density (Kg/m^3) Cp (J/kg.k) Volume M^3 1.205 1005 58.13 W/k Infiltration Min Total 0.09 AC/h 17.01 m^3/h 0.004725 M^3/s Volume 5.72 w/k SetPoint Tem Cooling Heating 23-25 20-22 C C 189 m^3 Efficiency 2.8 0.85 40