A procedure for the performance rvaluation of a new commercial building: part I--calibrated as-built simulation.INTRODUCTION
Many building energy studies and ASHRAE research projects have been reporting on efforts to calibrate simulations to measured data from monthly utility data (Diamond and Hunn 1981; McLain et al., 1994), and to hourly measured data (Hsieh 1988; Kaplan et al. 1990, 1992; Bronson et al. 1992; Huang 1994; Haberl et al. 1995; Huang and Crawley 1996; Haberl and Bou-Saada 1998; Abushakra 2001; Reddy 2004; Ramirez 2006). In addition, in-situ measurements of HVAC&R equipment (Phelan et al. 1997a, 1997b; Haberl et al. 1997; Liu et al. 2002) have been developed to support the effectiveness of calibrated simulation.
Some of the earliest published calibration procedures were developed in the two office buildings reported by Hsieh (1988), including: calibration of tenant energy use, HVAC equipment operation schedules and thermostat setpoints, heating and cooling equipment performance, building shell heat loss coefficient, zone definitions in DOE-2, outside air intake, and weather data. Of these factors, the calibration technique for equipment performance was used in this study because the DOE-2 program only provides standard default performance values that may not be related to the high efficiency equipment installed in a building.
Kaplan et al. (1990) developed "day-type schedules" to incorporate monitored lighting and equipment data into the typical operating schedule in the DOE-2 model. Such day-type schedules showed that monitored data could be used to generate simulation inputs, as well as to verify simulation outputs for calibrating the simulation model. Abushakra et al. (2001) completed ASHRAE research project 1093-RP for developing procedures to derive diversity factors and typical load shapes of lighting and receptacle loads in office buildings. In their study, a percentile analysis was used to derive load shapes and diversity factors, which were used in the current study. Recently, as an integral part of the California Energy Commission Commercial End-Use Survey(CEC CEUS) effort, a Site Processor (Ramirez 2006) has been developed to interface survey data with the building simulation processor(eQuest) for the purpose of site simulation and calibration to actual energy consumption data. In the site processor, four day-types defined in the 16-day format were used to make adjustments to whole end uses, which are similar to the load and diversity factors used in the current study.
Haberl et al. (1995) evaluated the impact of using measured weather data that was repacked into Test Reference Year (TRY) format vs. TMY format in a DOE-2 simulation by comparing the results of simulated energy use. Huang and Crawley (1996) also compared the influence of the various weather data sets, including: TRY, TMY, TMY2, WYEC (Weather Year for Energy Calculations), and WYEC2, on simulated annual energy use and energy cost. Huang and Crawley (1996) recommended that TMY2 (Marion and Urban 1995) should be used in building energy simulations where solar radiation is critical to the results. The results from both of these studies have provided guidance to the current study.
An effective calibrated simulation often requires in-situ performance measurements of the mechanical equipment, especially for high efficiency equipment that is used for new high performance buildings. To assist in this effort, Phelan et al. (1997b) developed a set of in-situ testing methods for pumps, fans, and chillers under ASHRAE Research Project 827-RP to evaluate annual energy consumption and to account for part-load operations that are affected by overall system controls. They developed semi-empirical chiller models using a statistical regression analysis based on one year of hourly measured data, including: chiller power consumption, evaporator flow rates, and chilled water and condenser water supply and return temperature, which proved to be useful in the current study. Liu et al. (2002) developed a procedure to determine the in-situ performance of commonly used HVAC systems as part of ASHRAE Research Project 1092-RP. In their project, the research objectives were to develop a simplified model calibration procedure from short-term field measurement and validate the calibration procedure using a simulation program developed with the ASHRAE modified bin method, which was useful to this study.
Finally, several statistical methods have been developed to access the goodness-of-fit of a simulation model, including: percent difference, mean bias error (MBE), and the coefficient of variation of the root mean square error (CV(RMSE)) (Kreider and Haberl 1994), which were used in the current study. Graphical comparisons are also useful to effectively represent the difference between simulated and measured data. In relation, Wei et al. (1998) developed "calibration signatures" of different parameters on the heating and cooling energy consumption of typical air handling units (AHUs) for model calibration. In this study, the signature concept is enhanced with percentile expression (i.e., binned statistical plot) to help identify the deviation between measured and simulated results for the detailed whole-building simulation, which is expressed in the companion paper (Song and Haberl 2008).
In summary, many of the previous simulation and calibration methods were shown to be useful in improving the accuracy and reliability of new building simulations, including: HVAC&R equipment performance, day-type profiles, packed weather data with solar radiation, and statistical methods. In this study, selected methods from the previous studies were applied for developing the as-built simulation model of the case-study building. In addition, several new calibration factors were developed and applied to a calibrated simulation of the case-study building, including: supply and outside air (OA) flow rate adjustments, a building thermal mass adjustment, an exhaust air adjustment scheme, hot deck air temperature adjustments, and new procedures for packing a TRY weather file, which compared measured vs. synthesized direct normal solar radiation.
AS-BUILT SIMULATION MODEL
The case-study building is a six-story, 303,389 [ft.sup.2] (28,196 [m.sub.2]) office building for the state legislative support staff, which includes a large print shop and data processing center. The building was designed to be a sustainable project with numerous Energy Conservation Design Measures (ECDMs) (Eley and Tathagat 1998) designed to make the building more efficient than prevailing building code (i.e., ASHRAE Standard 90.1-1989). The building contains over 50% windows in the facade consisting of two types of low-e glazing. Deciduous live oak trees shade a significant portion of the south facade up to approximately the 3rd floor as shown in Figure 1.
[FIGURE 1 OMITTED]
The DOE-2.1e program (i.e., Version 119) was used to simulate the case-study building. Information from site visits, DOE-2 manuals, as-built drawings, and measured energy and operation data were used to create a DOE-2 input file for the calibrated as-built simulation. Figure 1 shows the south-west facade of the DOE-2 model using the DrawBDL program for viewing the complex input file (Huang 1994). The north facade of the case-study building faces approximately 14 degrees east of north, which exposes the west of north facade to direct sunlight in the late afternoon. The measured site weather data were packed into the Test Reference Year (TRY) weather file format along with measured solar radiation data and then incorporated into the DOE-2 simulation in this study. Monthly ground temperatures were automatically calculated using the method of Kusuda and Achenbach (1965) as part of the DOE-2 weather processor, based on the packed Austin TRY weather files. Space zoning was established using interior and perimeter zones, based on the as-built drawings. Table 1 shows the conditions for a typical space of the case-study building. Lighting and equipment load densities and schedules were determined based on the measured data using the ASHRAE 1093-RP toolkit (Abushakra 2001). Figures 2(a) and 2(b) represent the derived day-type profiles for weekday and weekend schedules in terms of whole-building lighting and receptacle loads. People schedules for the entire building were adjusted from the measured electric load profile of the typical 4th floor. The hourly values of the 50th percentile in the day-type plots were used in the DOE-2 schedules of the as-built simulation. No infiltration was assumed in the DOE-2 simulation because the HVAC is always on and the building is assumed to be pressurized. For typical wall and roof constructions, each material was selected from the DOE-2 material library corresponding to the actual materials in the as-built drawings. Window files for the two window types used low-e glazing generated by the Window 5.2 program and incorporated into the DOE-2 window library. For the underground wall and floor, U-effective was calculated to account for the ground heat transfer in the DOE-2 simulation, using the method by Winkelmann (1992). Table 2 shows the calculated U-effective in this study for the underground wall and floors of the case-study building.
Table 1. Space Conditions for the Typical Floor DOE-2 Keywords Space Condition (Office) Description TEMPERATURE 71[degrees]F Midpoint of heating and (21.7[degrees]C) cooling setpoint AREA/PERSON 275 [ft.sup.2] - (25.6 [m.sup.2]) PEOPLE-HG-SENS 230 Btu/hr - (67.4 W) PEOPLE-HG-LAT 190 Btu/hr - (55.7 W) PEOPLE-SCHEDULE Scheduled Based on Lighting Schedul (ASHRAE 1093-RP) LIGHTING-W/SQFT 1.27 W/[ft.sup.2] Measured Data (13.7 W/[m.sup.2]) LIGHTING-SCHEDULE Scheduled Measured Data (ASHRAE 1093-RP) EQUIPMENT-W/SQFT 0.74 W/[ft.sup.2] Measured Data (8 W/[m.sup.2]) EQUIP-SCHEDULE Scheduled Measured Data (ASHRAE 1093-RP) INF-METHOD Air-change - AIR-CHANGE/HR 0 HVAC is always on (No infiltration) Table 2. U-Effective for the Underground Wall and Floors Items Underground Construction Conduction Wall Height Factor (F2) Underground Wall 8ft (2.44 m) 8 ft (2.44m), R-10 0.78 (deep basement) interior, concrete Underground Floor - - - Items Effective R = A/ Effective U = Remarks (F2 * Pexp) 1 / Reff Underground Wall 20.94 (hr-[ft.sup.2]- 0.048 (Btu/hr- - F/Btu) 3.69 ([m.sup.2]- [ft.sup.2]-F) [.sup.0]C/W) 0.273 (W/ [m.sup.2]- [.sup.0]C) Underground Floor 1000 (hr-[ft.sup.2]- 0.001 (Btu/hr- Exposed F/Btu) 176.1 [m.sup.2]- [ft.sup.2]-F) parameter [.sup.0]C/W) 0.006 (W/ (Pexp)=0 ([m.sup.2]- [.sup.0]C/)
[FIGURE 2 OMITTED]
The majority of the conditioned area in the case-study building (i.e., 23,690 [m.sup.2]) is served by the Dual-Duct, Variable Air Volume (DDVAV) systems with preconditioned outside air flowing from a special dedicated outside air unit (DOAU) (not simulated). Two DOAUs on the roof of the case-study building provide the east and west Air Handling Units (AHUs) on each floor with pre-conditioned OA, which is controlled by [CO.sub.2] space sensors located in the respective zones. For the basement HVAC system (i.e., 6,175 [m.sup.2]), four types of systems were used according to each space need, including: a bypass multi-zone system, a single-duct variable air volume system (VAV) without heating coil, a single-duct constant air volume (CAV) system with a humidifier, a heat wheel heat-recovery unit (not simulated), and Computer Room Units (CRUs).
Energy monitoring and in-situ measurements were performed for measuring the whole-building energy use and HVAC&R equipment operation of the case-study building. Table 3 shows a description of the monitoring data points and channels, including: 1) The whole-building electricity (WBE) use, motor control center (MCC) electricity and other weather independent electric use (WBE-MCC). 2) The electricity use of the two chillers and the thermal energy use with chilled water flow, chilled water supply and return temperature, and condenser water supply and return temperature. and 3) Boiler energy use with hot water flow and supply and return temperature. Figure 3 shows the detailed thermal monitoring diagram of the central plant in the case-study building. Additional information about the whole-building energy monitoring can be found in Song & Haberl (2008) and Song (2006).
Table 3. Monitoring Channel Description Items Description Unit Channels Remarks WBE Whole kWh/h Building Building Electricity Electricity 1 Phase A (ch4497) Building WBE 1 (Phase A+B+C) Electricity 1 Phase B (ch4498) Building Electricity 1 Phase C (ch4499) Building Electricity 2 Phase A (ch4500) Building WBE 2 (Phase A+B+C) Electricity 2 Phase B (ch4501) Building Electricity 2 Phase C (ch4502) MCC Motor kWh/h MCC Electric Phase A + Phase C Control Phase A Center (ch4476) MCC Electric Phases C (ch4477) WBE-MCC kWh/h WBE-MCC Weather independent electric use (Lighting, receptacles & others) Chiller # 1 kWh/h Electricity Phase A + Phase C Phase A (ch4478) Electricity Phase C (ch4479) kBtu/h User (GBH * (supply- Defined return) temp)/2 Channel (ch4520) GPH Chilled - Water Flow (ch4484) [degrees]F Chilled - Water Supply Temp. (ch4485) [degrees]F Chilled - Water Supply Temp. (ch4485) Chilled - Water Return Temp. (ch4486) Chillers Chiller # 2 kWh/h Electricity Phase A + Phase C Phase A (ch4480) Electricity Phase C (ch4481) kBtu/h User Defined (GPH * (supply- Channel return) temp)/2 (ch4521) GPH Chilled - Water Flow (ch4489) [degrees]F Chilled - Water Supply Temp. (ch4490) [degrees]F Chilled - Water Supply Temp. (ch4491) Chiller # 3 - No Channels No sensors installed Boiler kBtu/h User Defined (GPH * (supply- Channel return) temp)/2 (ch4522) GPH Hot Water (GPH * (supply- Flow return) temp)/2 (ch4494) [degrees]F Hot Water - Flow (ch4494) [degrees]F Hot Water Return Temperature (ch4496)
The case-study building also contains high efficiency mechanical equipment, including: two low-NOx boilers, three high efficiency centrifugal chillers, two over-sized cooling towers, and other miscellaneous pumps. In the DOE-2 simulation, HVAC equipment efficiency was determined using the ratio of energy input to energy output at full load (normal capacity), including the Electric Input Ratio (EIR) for the electrical equipment (i.e., chillers, fans, and pumps) and Heat Input Ratio (HIR) for the equipment requiring fuel input (i.e., boiler) (LBNL 1993). Table 4 shows the plant model at normal operating conditions for the as-built simulation. According to the manufacturer's specifications, the two low-NOx boilers have a normal rated output capacity of 4.2 MMBtu (4.4 GJ) with a Heat-Input-Ratio (HIR) of 1.19. The two centrifugal chillers have a normal cooling capacity of 5.58 MMBtu (5.89 GJ) with an Electric-Input-Ratio (EIR) of 0.1547 (6.59 COP). The two cooling towers have an over-sized output capacity of 12 MMBtu (12.66 GJ) with an Electric-Input-Ratio (EIR) of 0.0046, which is determined by the fan power consumption of an open tower to 0.003 kW/(L/min) (0.0105 Btu/Btu) at the CTI (Cooling Technology Institute) rating conditions used in DOE-2 (LBNL 1993). In this study, the DOE-2 default efficiency was first adjusted with manufacturer's data and then measured data to account for the actual HVAC performance at normal operation conditions. For part-load conditions, the DOE-2 default chiller curves were used because the measured curves were found to be quite similar in shape to the DOE-2 default curves as shown in Figure 4(a). However, a method of switching chiller performance curves was also required to further account for the actual operation with either parallel or sequential operation at part-load conditions (not simulated). This can be seen as the two distinct clouds of points below a PLR of 0.60 in Figure 4(b).
[FIGURE 4 OMITTED]
Table 4. DOE-2 Plant Model of the Case-study Building Items DOE-2 Model Description BOILER HW Boiler PVI Industries b(125 WBE 250A-TP) SIZE 4.2. MMBtu (4.4 GJ) 4.4 GJ INSTALL NUMBER 2 - HW-BOILER-HIR 1.19 Input (5.25 GJ)/Output (4.42 GJ) CHILLER HERM-CENT-CHLR TRANE (CVHF-555) SIZE 5.58 MMBtu (5.89 GJ) 465 TON INSTALL NUMBER 2 - ELEC-INPUT-RATIO 0.1547 0.544 (kW/ton); 6.59 COP COOLING TOWER OPEN-TWR - SIZE 12 MMBtu (12.66 GJ) 1000 TON INSTALL NUMBER 2 - ELEC-INPUT-RATIO 0.0046 (14.9 kW/11,355 (L/min)* 3.5
AS-BUILT MODEL CALIBRATION
The as-built simulation was run again until the simulated data matched with measured data to a suitable level. Both graphical and statistical tools were used to perform the calibration. The as-built model was calibrated with measured data by changing the calibration factors cumulative to the base-model. Table 5 shows the primary input parameters used for each run, including: (1) Supply air and outside air (OA) flow rates, (2) Building thermal mass, (3) Undocumented exhaust air flow rates, (4) Max. supply air temperature, and (5) Measured weather file with synthesized direct normal solar radiation. In this process, the model was calibrated by comparing the hourly simulation results to the measured data, and then making adjustments. The simulation impact of each calibration factor is described in the following section.
Table 5. Primary Input Parameters Used in the Calibration Input Base Model Run 1 Run 2 parameters 1 Supply Air Assigned CFM Auto Auto Flow calculated calculated Outside Air Assigned CFM 0.1 0.1 Flow 2 Weighting Pre-calculated Pre-calculated Custom Factor 3 Undocumented 0 0 0 Exhaust Air Flow 4 Max Supply 105 [degrees]F 105 [degrees]F 105 [degrees]F Temp. (F) (41[.sup.0]C) (41[.sup.0]C) (41[.sup.0]C) 5 Direct Normal Measured Measured Measured Solar Radiation Input Run 3 Run 4 Run 5 parameters 1 Supply Air Auto Auto Auto Flow calculated calculated calculated Outside Air 0.1 0.1 0.1 Flow 2 Weighting Custom Custom Custom Factor 3 Undocumented 0.3 0.3 0.3 Exhaust Air Flow 4 Max Supply 105 [degrees]F 95 [degrees]F/ 95 [degrees]F/ Temp. (F) (41[.sup.0]C) 75 [degrees]F 75 [degrees]F (35[.sup.0]C/ (35[.sup.0]C/ 24 [.sup.0]C) 24 [.sup.0]C) 5 Direct Normal Measured Measured Calculated Solar Radiation
Once the as-built simulation was calibrated, the hourly simulated data were extracted from selected DOE-2 reports and then evaluated with statistical comparisons to measured data in order to determine how well the simulation model fits the data in the process of calibration (i.e., the lower the CV(RMSE), the better the calibration) (Haberl and Bou-Saada 1998). The coefficient of variation of the root mean square error (CV(RMSE)) is the root mean square error divided by the measured mean of all the data. The mean bias error (MBE) determines the non-dimensional bias between the simulated data and the measured data for each individual hour. Table 6 summarizes the statistical calibration results for each run. Figure 5a shows the overall daily CV(RMSE) and MBE (Kreider and Haberl 1994) analyzed in this study, in terms of heating, cooling, and electricity. Figures 5(b) through 5(c) show the daily CV(RMSE) and MBE for heating, cooling, and electricity (WBE) in each run, respectively.
[FIGURE 5 OMITTED]
Table 6. Summary of Statistical Results in Each Run Calibration Runs Daily MBE (%) Daily CV (RMSR) (%) 0 Cooling Heating WBE Cooling Heating WBE 0 -38.62 20.29 7.63 40.94 58.85 8.84 1 -49.49 -22.98 -3.84 53.04 82.35 6.01 2 -41.05 19.26 -2.32 44.61 55.49 5.07 3 1.75 42.08 11.93 9.49 58.18 12.52 4 -4.85 12.92 10.50 8.55 39.38 11.13 5 -7.01 -0.60 9.51 10.31 40.62 10.22 N-1 321 321 359 321 321 359 Calibration Runs Overall (%) MBE CV(RMSE) 0 -3.57 36.21 1 -25.44 47.13 2 -8.04 35.06 3 18.59 26.73 4 6.19 19.69 5 0.63 20.38 N-1 - -
1st Run: Supply Air and Outside Air Flow Rate
The as-built simulation was first modeled with assigned CFMs for the supply and outside air flow rates based on the design information from the as-built drawings. As a result, the simulated electricity use was much higher than the measured data (i.e., WBE-Chiller) as shown in Figure 6(a). In this calibration step, instead of the assigned CFM, the minimum supply air flow rate was set to 0.6 for the VAV systems, and the outside air flow rate was set at 10% of the total supply air flow for all the AHU systems. As a result of this adjustment, the system electricity use had an improved agreement with the measured data. Figure 6(b) shows the hourly time series plot of the system electricity use after this calibration with the adjusted supply and outside air flow rate. In the 1st calibration step, the CV(RMSE) and MBE for whole-building electricity (WBE) were improved as expected, but the CV(RMSE) and MBE for cooling and heating became worse as shown in Figure 5(b) and 5(c).
[FIGURE 6 OMITTED]
2nd Run: Building Thermal Mass
In the DOE-2 program there are two ways to incorporate the dynamic effects of thermal mass, including pre-calculated weighting factors and Custom Weighting Factors (CWFs). In the 2nd run, custom weighting factors were used instead of pre-calculated weighting factor, which was initially assumed to be 70 lb/[ft.sup.2] (342 kg/[m.sup.2]) for the floor weight. Unfortunately, making the change from pre-calculated weighting factors to CWFs requires substantial changes to the input file, including: layering of all walls, roofs, and floors. In the second run with custom weighting factors, the heating and cooling energy increased slightly. However, the cooling energy use was still much lower than the measured data as shown in Figure 7. The CV(RMSE) for heating and cooling improved in the 2nd run as expected, but not enough to reach a suitable range as shown in Figure 5(b).
[FIGURE 7 OMITTED]
3rd Run: Undocumented Exhaust Air Flow Rate
Undocumented exhaust air in the case-study building was addressed using the DUCT-AIR-LOSS command in DOE-2. In Figure 8, the base-case model had a much lower cooling load than the measured data. From the cooling load comparison between simulated and measured data, a 30% duct-air-loss was used for the 3rd calibration run, which included 10% exhaust air from the exhaust fans installed on the roof of the case-study building according to the fan schedules. The remaining of 20% exhaust air was assumed to be account for the unknown exhaust air, which could have occurred from the senate print shop, ductwork or other losses. The inclusion of the undocumented exhaust air had a significant improvement in the overall CV(RMSE) from 35.06 to 26.73, which included a decrease in the cooling CV(RMSE) from 44.61 to 9.49. MBE for cooling was also much improved. However, the MBE and CV(RMSE) for heating became worse as shown in Figures 5(b) and 5(c).
[FIGURE 8 OMITTED]
4th Run: Hot Deck Air Temperature
A careful inspection of the measured data revealed that the boiler hot water supply and return temperature suddenly dropped to 140 [degrees]F (60 [.sup.0]C) and then returned to the previous setpoint of 180 [degrees]F (82 [.sup.0]C) due to an operation change as shown in Figure 8. Unfortunately, there is no keyword to control the hot water supply temperature for a boiler in DOE-2 (i.e., version 119). Therefore, in this study, the boiler operation change was incorporated into the DOE-2 simulation based on the measured hot deck and cold deck air temperature for a typical AHU (DDVAV). Figure 9 shows the measured and simulated heating energy use before the hot deck air temperature adjustment. As a result, the simulated heating energy was separated to two groups that more closely matched the measured data, which is shown in the final calibration result (Figure 12(a)). This produced a significant decrease in the CV(RMSE) for heating from 58.18 to 39.38 as shown in Figure 5(b). CV(RMSE) for cooling and WBE also improved slightly. However, the MBE for cooling was slightly worse as shown in Figure 5(c). Overall CVRMSE and MBE were improved as shown in Figure 5(a).
[FIGURE 9 OMITTED]
5th Run: Calculated Direct Normal Solar Radiation
Measured direct normal solar radiation is often not available even though it is an important factor for preparing a weather file into the proper format for use by the DOE-2 program. In this study, measured beam and diffuse were available for the first measurement period but not in the second period. This caused problems with the use of the calibrated model in the second period which was based on calibrations from the first period that used a different source for the direct-normal solar radiation. Therefore, direct normal solar radiation was synthesized for both periods using the Erbs correlation (Duffie and Beckman 1991) and compared against the measured data for the first period. Figure 10 shows the measured and synthesized diffuse fraction against clearness index (Kt) after trimming the bad data. As expected, diffuse fraction of the hourly total radiation is strongly correlated with Kt, which is an indicator of the relative clearness of the atmosphere. Surprisingly, the use of synthetic data for both periods introduced significant changes to the direct normal component as shown in Figure 11. When simulations using a TRY weather file packed with measured direct normal solar radiation were compared with synthetic data, it was found that there was a 2% increase in cooling energy and a 15% increase in heating energy.
[FIGURE 10 OMITTED]
[FIGURE 11 OMITTED]
SUMMARY AND CONCLUSION
A detailed as-built simulation and calibration were performed for a case-study building that has energy conservation design measures (ECDMs), based on as-built drawings and measured data, including: weather data packed to TRY format, light & receptacle day-typing, Custom Weighting Factors (CWF) with U-effective calculation to account for thermal mass, low-e windows using the Window 5.2 library, and measured HVAC systems performance for chillers and AHUs. Calibrations were analyzed at each step using graphical and statistical comparisons to develop a calibrated simulation of the case study building, including: supply and outside air (OA) flow rates, building thermal mass, undocumented exhaust air, hot deck air temperature, and measured vs. synthesized direct normal solar radiation. As a results, the as-built calibrated model for a case-study building was improved from overall CV(RMSE) of 36.21 to 20.38% and MBE almost eliminated. Figures 12(a) through 12(c) show the base-case and final calibration vs. measured data. Consequently, the as-built simulation using the calibration factors analyzed in this study were considered sufficient to be used to perform further evaluations, including an evaluation of the energy performance of the energy-efficient design measures (ECDMs), which is reported in the companion paper (Song and Haberl 2008).
This study was partially funded by the State of Texas Senate Bill 5 program and the Texas State Energy Conservation Office. We would like to acknowledge the assistance of many people in the Energy Systems Laboratory at Texas A&M University. We also would like to acknowledge Dr. Dan Turner, Dr. Charles Culp, and Dr. Liliana Beltran for their assistance and comments. We are also thankful to Mr. Kelly Milligan and Mr. Jim Sweeney for their efforts to help gather data from the Robert E. Johnson (REJ) Building over a 4 year period. We are also thankful to Mr. Martin Wilford and Mr. Mel Bullock from the REJ building for their time and kind help.
Abushakra, B., A. Sreshthaputra, J. S. Haberl, and D. E. Claridge. 2001. Compilation of diversity factors and schedules for energy and cooling load calculations. Energy System Laboratory Report No. ESL-TR-01/04-01. College Station: Texas A&M University.
ASHRAE. 2002. ASHRAE Guideline 14: Measurement of energy and demand savings. Atlanta, GA: American Society of Heating, Refrigerating, and Air-Conditioning Engineers.
Ramirez, B. 2006. Site processor user guide. San Diego, Itron, Inc.
Bronson, D. J., S. B. Hinchey, J. S. Haberl, and D. L. O'Neal. 1992. A procedure for calibrating the DOE-2 simulation program to non-weather-dependent measured loads. ASHRAE Transactions 98(1): 636-652.
Duffie, J. A., and W. A. Beckman. 1991. Solar engineering of thermal processes. New York, John Wiley & Sons, Inc.
Haberl, J. S., D. J. Bronson, and D. L. O'Neal. 1995. Impact of using measured weather data vs. TMY weather data in a DOE-2 simulation. ASHRAE Transactions.101(2), 558-576.
Haberl, J. S., T. A. Reddy, I. Figueroa, and M. Medina. 1997. Overview of LoanSTAR chiller monitoring and analysis of in-situ chiller diagnostics using ASHRAE RP827 test method. Proceedings of the PG&E Cool Sense National integrated Chiller Retrofit Forum, pp. 1-19.
Haberl, J. S., and T. E. Bou-Saada. 1998. Procedures for calibrating hourly simulation models to measured building energy and environmental data. Journal of Solar Energy Engineering 120: 193-204.
Hsieh, E. S. 1988. Calibrated computer models of commercial buildings and their role in building design and operation. Master's Thesis. PU/CEES Report No. 230. Princeton University, Princeton, NJ.
Huang, Y. J. 1994. DrawBDL version 2.02. Joe Haung and Associates: Moraga, CA.
Huang, Y. J., and D. B. Crawley. 1996. Does it matter which weather data you use in energy simulation? Proceedings of ACEEE 1996 Summer Study on Energy Efficiency in Buildings, pp. 4.183- 4.192.
Kaplan, M. B., J. McFerran, J. Jansen, and R. Pratt. 1990. Reconciliation of a DOE2.1C model with monitored end-use data for a small office building. ASHRAE Transactions 96(1): 981-992.
Kaplan, M. B., P. Cancer, and G.W. Vincent. 1992. Guidelines for energy simulation of commercial buildings. Proceedings of ACEEE 1992 Summer Study on Energy Efficiency in Buildings, pp. 1.137- 1.147.
Kreider, J. and J. S. Haberl. 1994. Predicting hourly building energy usage: The great energy predictor shootout-overview and discussion of results. ASHRAE Transactions 100(2):1104-1118.
Kusuda, T., and P. R. Achenbach. 1965. Earth temperature and thermal diffusivity at selected stations in the united states. ASHRAE Transactions 71(1): 61-75.
LBNL. 1993. DOE-2.1E supplement manual. Berkeley, CA: Lawrence Berkeley National Laboratory.
Liu, M., D. E. Claridge, J. S. Haberl. 2002. Development of procedures to determine in-situ performance of commonly used HVAC systems (1092-TRP). Lincoln, NE: Energy Systems Laboratory.
Marion, W., and K. Urban. 1995. User's manual for TMY2s: Derived from the 1961-1990 national solar radiation data base. Golden, CO: National Renewable Energy Laboratory.
Phelan, J., M. J. Brandemuehl, and M. Krarti. 1997a. In-situ performance testing of chillers for energy analysis. ASHRAE Transactions 103(1): 290-302.
Phelan, J., M. J. Brandemuehl, and M. Krarti. 1997b. In-situ performance testing of fans and pumps for energy analysis. ASHRAE Transaction 103(1): 318-332.
Reddy, T. A. 2004. Procedures for reconciling computer-calculated results with measured energy data (1051-RP Rep2). Work-in-progress report, Drexel University, Philadelphia, PA.
Sylvester, K., S. Song, J. S. Haberl, and D. Turner. 2002a. Energy savings assessment of the Robert E. Johnson state office building. Proceedings of the Thirteenth Symposium on Improving Building Systems in Hot and Humid Climates, pp. 103-109.
Song, S. 2006. Development of new methodologies for evaluating the energy performance of new commercial buildings. Ph.D. dissertation, Texas A&M University, College Station, TX.
Song, S and Haberl, J.S. 2008. A procedure for the performance evaluation of a new commercial building: Part II-Overall methodology and comparisons of methods. ASHRAE Transactions (Accepted in Mar., 2008).
Wei, G., Liu, M., and D.E. Claridge. 1998. Signature of heating and cooling energy consumption for typical AHUs. Proceedings of the Thirteenth Symposium on Improving Building Systems in Hot and Humid Climates, pp. 387-402.
Winkelmann, F. C. 1998. Underground surfaces: How to get a better underground surface heat transfer calculation in DOE-2.1e. Building Energy Simulation User News, 19(1): 17-25.
Suwon Song, PhD
Jeff S. Haberl, PhD, PE
Suwon Song is a research professor at the Center for Sustainable Buildings, Department of Architecture, Yonsei University, Seoul, South Korea. Jeff S. Haberl is a professor at the Energy Systems Laboratory, Department of Architecture, Texas A&M University, College Station, TX.