Field testing of optimal controls of passive and active thermal storage.
Shifting building cooling loads using thermal energy storage (TES) systems provides several advantages including reduction of peak demands for the electrical utilities and reduction of operating costs for the building owners. Generally, two types of TES systems are typically utilized in buildings: passive and active.
Passive TES systems utilize precooling strategies of the building thermal mass during nighttime to shift and reduce peak cooling loads (Braun 2003). Simulation analyses of various precooling strategies have shown that energy cost savings of 10% to 50% and peak demand reductions of 10% to 35% are possible by utilizing a preconditioning control strategy (Braun 1990, Rabl and Norford 1991, Conniff 1991, Andreson and Brandemuehl 1992, Morris et al. 1994, Keeney and Braun 1996, Chen 2001, Braun et al. 2001, Chaturvedi and Braun 2002). Experimental studies have also shown comparable levels of cost savings and peak demand reduction (Braun et al. 2001, Keeney and Braun 1997, Morris et al. 1994). Control optimization geared toward specific outcomes can generally increase cost savings or peak demand reduction (Braun 2003).
Active TES systems refer to the use of chilled water or ice tanks on the plant chilled water loop as a heat storage medium. Active TES systems provide load shifting by allowing the chiller plant to be run during unoccupied periods, storing the heat absorption capacity, and discharging it during occupied and/or peak periods to reduce the need for mechanical cooling of the chilled water loop. Chilled water tanks and ice storage tanks are the most common active TES equipment. The dispatchable load shifting capacity with active TES systems allows for a reduction in chiller size due to a reliable reduction in peak loads, and the lower chilled water supply temperature allows for unique airside HVAC designs (Henze and Krarti 2002). Several control strategies have been proposed for active TES systems including chiller-priority, storage-priority, constant-proportion, and optimal controls (Henze 2003). While some active TES systems in the field have been found to be underperforming (Guven and Flynn 1992, Tran et al. 1989), these systems have demonstrated overall cost savings and increased energy consumption compared to systems without active TES (Sohn 1991, Henze and Krarti 1998, Ihm et al. 2004). Simulation work on optimal control of active TES has shown that it is possible to reduce costs by as much as 20% without increasing overall energy consumption (Henze and Krarti 1998).
The combined utilization of passive and active TES systems has been investigated and found to be capable of reducing costs by up to 45% when optimal controls are considered (Henze and Krarti 2002, Henze et al. 2004, Zhou et al. 2005, Krarti et al. 2007).
In this paper, the performance of combined passive and active TES systems is investigated through field testing of various control strategies. The field testing is carried in an elementary school in Colorado equipped with an ice storage system. A simulation environment based Energyplus (Crawley, 2000), a detailed whole-building simulation program is used to determine the optimal control strategies (Zhou et al. 2005, Krarti et al. 2007). First, the building and its cooling system is presented. Then, the testing procedures as well as the simulation environment are briefly outlined. Finally, the testing results are summarized and discussed.
The field testing for the TES control strategies has been carryout in an elementary school located in Fort Collins, CO. This building, shown in Figures 1 and 2, part of the Poudre School District (PDS), was built in 2002 as a model of a high-performance school building. The school has a total floor area of 65,000 [ft.sup.2] including roughly 48,000 [ft.sup.2] is on the ground level, with the remaining 17,500 [ft.sup.2] on the second floor. There are 25 classroom spaces, a gymnasium, cafeteria, library, computer lab, and staff offices (see Figures 1 and 2).
[FIGURE 1 OMITTED]
[FIGURE 2 OMITTED]
The school is occupied Monday through Friday from 8:30 am to 3:30 pm by 630 students and staff from August 24th to May 31st. The building is unoccupied from June 1st to August 15th when staff returns to prepare for the academic year. The building is also unoccupied for two weeks over the holidays of Christmas and New Years. Table 1 summarizes the basic features of the elementary school.
Table 1. Elementary School Building Characteristics Category Description Occupancy Schedule 630 students and staff 8:30 am to 3:30 pm M-F Aug 24 to June 1 Unoccupied during holidays Exterior Wall 10 cm (4-in) brick facade, 2.5 cm (1-in) air Construction gap, vapor barrier, 15 cm (6-in) metal stud with R-19 batt insulation, brick or Roof Construction 1 cm(0.5-in) gravel, metal sheeting, 10 cm (4-in) rigid insulation, vapor barrier Floor Construction 10 cm (4-in) concrete slab Glazing Type Eye-level: 24 mm (0.95-in) Double Pane Glass - Conductivity = 0.9 W/m*K Shading Coefficient = 0.66 Clerestory: 24 mm (0.95-in) double pane glass Equipment Power 5 W/[m.sup.2] (0.47 W/[ft.sup.2]); Schedule Density 100% 9am-12pm, 1pm-3pm, 50% 8am-9am, 12pm-1pm, 3pm-4pm, 0% all other times Lighting Power 10 W/[m.sup.2] (0.94 W/[ft.sup.2]); Schedule Density 100% 8am-6pm, 10% all other times Conventional Cooling 24[degrees]C (75[degrees]F) from 8am-5pm, Setpoints 30[degrees]C (85[degrees]F) all other times HVAC System 7 AHUs combined 70 hp (52 kW) and 51,700 cfm with 22[degrees]C (72[degrees]F) economizers Chiller Trane CGAFC 50 ton capacity Characteristics Ice Storage System Calmac 1500C 570 ton-hr unit 15% latent
To cool the school, a chiller and a thermal energy storage (TES) system are utilized. The cooling setpoint is 24[degrees]C (75[degrees]F) during occupied periods from 8:30 am to 5 pm, Monday - Friday. During unoccupied periods, the temperature is set to 32[degrees]C (90[degrees]F), effectively allowing the temperature to float. When the cooling system is not engaged (i.e. during offseason), the setpoint is left to float as well.
The air-cooled chiller with a 50-ton scroll compressor is run only at night, charging the TES system when the building is unoccupied and the building electrical demand is at a minimum. The chiller is kept in assist mode during the day in order to meet unexpected cooling loads. The TES system handles all cooling loads during the occupied period. Typical operation is from 2 or 3 am until 7 or 8 am. Sometimes, the chiller is operated earlier when additional charging of the storage tank is needed. The chiller is never operated after 8 am when the interior lighting and AHUs come on as the building starts to be heavily occupied by staff and students.
The TES system, consisting of 3 ice-tanks with a total capacity of 570 ton-hr, is an internal melt ice-on-coil system. Its refrigerant is a water/glycol (brine) solution operated at - 2.8[degrees]C (27[degrees]F). The TES system was sized to meet 100% of cooling load for design-day conditions. Thus, it supplies chilled water to the cooling coils during occupied hours (on-peak demand-setting period). It is generally discharged during occupied period as needed to meet the building cooling load. The ice storage tanks are fully charged every night except on weekends. The tanks are located outside, buried underground.
The TES system is charged fully every weekend, sometimes operating from Saturday evening until Monday morning, immediately prior to occupancy. This is possible because the building is completely unoccupied on the weekends, so the use of the chiller during the day does not increase on-peak electrical demand. During periods of high cooling load, the TES is typically depleted just before the end of the occupied period. This is in part due to the small chiller size and in part to the restrictions on demand due to outdoor lighting and other small evening loads. Figure 3 illustrates a typical weekly charge cycle for the TES during a period of peak cooling demand.
[FIGURE 3 OMITTED]
Two types of tests were carried out at the elementary school in addition to the tests that evaluate the performance of the existing control strategies. Due to demand constraints, the chiller was almost never used during occupied periods, and since the ice storage system was so large the cooling energy consumption was measured by monitoring the discharge of the ice tanks. All system information was obtained from the PSD online building automation system. Weather data was obtained from a nearby weather station operated by the Northern Colorado Water Conservation District (NCWCD). The weather data included hourly averaged dry bulb temperature, wind speed, and solar radiation (Krarti et al., 2007).
The first type of test was a simple precooling test to determine the dynamic behavior of the building when its passive thermal mass was utilized. Three of these tests were performed in May just before classes ended for summer recess. Two additional precooling tests were completed at the end of August. In these tests, indoor temperature setpoints were determined prior to the tests through a series of simulation analyses and discussions with the PSD energy manager. After determining the appropriate precooling temperatures and periods, zone temperature setpoints were changed remotely for the specified periods.
During the second half of August when the building is fully occupied, optimized and predictive tests were carried out. A simulation environment, briefly presented in the following section, is used to determine the optimal settings (Zhou et al. 2005). These tests were performed for predictive update windows of 4, 6, and 24 hours. The simulation analysis was executed based on the latest available weather data, and the temperature setpoints were changed to reflect the optimized control determined by the simulation environment. Due to control restrictions for the cooling system, the ice storage was discharged to meet the cooling load for all occupied periods.
To determine and implement various control strategies including predictive optimal controls, a simulation environment has been developed. The basic structure of the simulation environment is shown in Figure 4. A weather predictor is used to generate future weather information that can be used by the simulation environment (consisting of a whole building simulation program with internal optimization algorithms) to determine the optimal temperature setpoints and/or charging/discharging rates to minimize the building utility costs while maintaining comfort. A detailed description of the simulation environment as well as the weather predictor is presented by Zhou et al. (2005) and Krarti et al. (2007).
[FIGURE 4 OMITTED]
In this study, the school was modeled using EnergyPlus program (Crawley et al 2000). The elementary school was modeled using the basic characteristics provided in Table 1. Both monthly and hourly calibrations of the simulation model were carried out.
Figure 5 shows the results for the monthly calibration. The simulation model predictions and the utility data were within 5% for all the months.
[FIGURE 5 OMITTED]
Hourly calibration of the simulation model predictions were performed using hourly monitored data obtained during regular building operation during a hot day. For this day, the TES, fully charged in the morning, was depleted throughout the day to meet the building cooling load. Zone setpoints were set to float during the unoccupied periods and were set to 24[degrees]C (75[degrees]F) from 8 am to 5 pm. The building simulation model predicted reasonably well the actual building performance as presented in Figure 6 for the indoor temperature variation and Figure 7 for the ice level in the TES tank. The discrepancies between hourly model predictions and measured data are within 10%. The overall energy consumption predicted by the simulation model was within 5% of measured values on an hourly and daily basis.
[FIGURE 6 OMITTED]
[FIGURE 7 OMITTED]
FIELD TESTING RESULTS
Once the simulation model was calibrated, several tests were performed at the elementary school to evaluate various TES control strategies. Conventional controls and basic preconditioning tests were carried out first. Tests were completed for 4 hour and 8 hour precooling to 21[degrees]C (70[degrees]F) for the periods immediately prior to occupancy. After these test results were analyze, several predictive optimal control tests were performed for various update-window periods. Due to the time delay for uploading the weather data as well as the time required for simulation runs, the shortest update-window used was 4 hours. The results from all tests are shown, discussed, and compared in this section.
Existing controls, referred to as conventional controls, of the cooling system, consist of night charging of the TES, a brief preconditioning period at 24[degrees]C (75[degrees]F) from 8 am to 8:30 am to ensure occupant comfort upon arrival, occupied space cooling to 24[degrees]C (75[degrees]F) with the chilled water load met by discharging the TES, and floating zone setpoints during unoccupied periods usually beginning at 5 pm. The chiller is not operated during the occupied period due to the size of the TES and peak demand constraints.
The thermal performance of the actual building and the results obtained from the simulation model operated using the conventional controls are shown in Table 2 and Figure 8. For all the variables, the model predictions are within 5% of the measured test results. The total TES cooling energy consumption obtained from measured data and from the simulation model predictions were estimated to be 930 kWh and 959 kWh, respectively (i.e., a relative difference of 3.1%). The total building electricity use is shown in Figure 8. Total energy consumption and electrical demand matched very closely between measured data and model predictions. The energy use profile by the HVAC system, which includes fans, pumps, and the chiller predicted by the simulation model, agrees well with the actual data. The spike in HVAC energy use noted between 5 and 6 pm reflects an unusual period of evening occupancy that was due to parent-teacher meetings for this day, when the typical late afternoon electrical load is 20 kW lower. The spike in actual building HVAC consumption between 10 pm and midnight is due to the chiller operation beginning the next day's charging cycle.
[FIGURE 8 OMITTED]
Table 2. Building and Model Energy and Operation Cost Comparison for the Conventional Control Energy Peak Energy Cost Demand Cost Total Cost Consumption Demand (kWh) (kW) Actual 1203 94 $19.73 $35.22 $54.95 Model 1178 94 $19.32 $35.22 % Difference 2.1% 0.0% 2.1% 0.0% 0.7%
Daily total energy costs for all the tests were calculated using the actual utility rate with two assumptions. Peak demand is calculated on a monthly basis for the actual billing, but for daily tests it is necessary to assume that the day of the test contains the monthly peak demand (or that all the days in the month have the same energy use profile). To fairly compare the demand cost with the energy cost, the demand charge was divided by 30 before it is added to the daily energy charges.
The model predictions agreed generally well with the measured data obtained for this test. All total energy consumption and cost data were within 6%. The demand charges made up the bulk of the total energy cost. This is reasonable because the utility rate is structured with a flat consumption rate of $0.0164/kWh and a demand charge of $11.24/kW. This rate provides a strong incentive for load shifting or leveling, and it is the main reason for the unusual type of cooling system control that is considered "conventional" for Zach Elementary building.
STANDARD PRECOOLING CONTROLS
Tests of standard precooling periods of 4 and 8 hours were performed to determine the performance of the building thermal mass used as a storage medium to complement the ice TES system. For these tests the precooling setpoint was 21[degrees]C (70[degrees]F), and the precooling period was set to end immediately prior to the occupancy period to minimize heat gain to the mass between the end of the precooling period and the beginning of the occupied period.
A precooling test was carried out using 8 hours of cooling all spaces to 21[degrees]C (70[degrees]F) from 12 am to 8 am, which corresponded to the eight hours immediately prior to occupancy. Setpoints were changed to 24[degrees]C (75[degrees]F) during the occupied period of 8 am to 5 pm and left to float thereafter. Figure 9 and Table 3 summarize the performance of the simulated and actual cooling systems during the 8-hr precooling test. Note the effect of the precooling on measured zone temperatures. Measured temps are below setpoints for most of the day, reflecting the discharge of the preconditioned building mass.
[FIGURE 9 OMITTED]
Table 3. Building and Model Energy and Operation Cost Comparison for 8-Hour Precooling Test Energy Peak Energy Cost Demand Cost Total Cost Consumption Demand (kWh) (kW) Actual 1395 100 $22.88 $37.47 $60.35 Model 1340 96 $21.98 $35.97 $57.95 % Difference 3.9% 4.0% 3.9% 4.0% 4.0%
As shown in Table 3, the model predictions of energy consumption, peak demand, and cost are within 5% of the measured data. This test demonstrates that the simulation model is well suited for predicting precooling effects.
A second test was performed using 4 hour precooling. The setpoints were the same as those set for the 8-hr precooling test, except that the zone temperatures were left to float for the first 4 hours of the day. The effect of the load shifting on measured zone temperatures is noticeable here as illustrated in Figure 10 and Table 4.
[FIGURE 10 OMITTED]
Table 4. Building and Model Energy and Operation Cost Comparison for 4-Hour Precooling Test Energy Peak Energy Cost Demand Cost Total Cost Consumption Demand (kWh) (kW) Actual 1332 95 $21.85 $35.54 $57.40 Model 1271 96 $20.84 $36.12 $56.95 % Difference 4.6% 1.6% 4.6% 1.6% 0.8%
Again the values for energy consumption, peak demand, and cost are within 5% when comparing the model predictions to the actual values. Note that this test was run on a cooler day than the 8-hour precooling test. Total energy use and demand are both slightly lower for this test, which is expected.
Predictive Optimal Controls
Once the test procedure had been used for conventional and standard precooling control strategies, predictive optimization control tests were carried out using the simulation environment (refer to Figure 3) for different updating periods. Preliminary simulation runs revealed that using the actual utility rate would not provide enough incentive for precooling based on the results from optimization control module. This result is most likely due to the fact that the energy costs were constant regardless of time of day, and the demand charge was not significant enough for load shifting. Moreover, the TES capacity was sufficient to meet 100% of all cooling loads for cooling design-day conditions, preconditioning the space provided little additional load-shifting potential. Because of these factors a different utility rate was used to increase the incentive to shift loads to off-peak hours, thereby calling for some preconditioning of the building when optimized control was used. The new utility rate and the original utility rate are shown in Table 5.
Table 5. Original and High-Incentive Utility Rates Actual Peak Off-Peak High-Incentive Peak Off-Peak Energy ($/kWh) 0.0164 0.0164 Energy ($/kWh) 0.164 0.0164 Demand ($/kW) 0 11.24 Demand ($/kW) 0 11.24
The variables selected for optimization were the zone temperature setpoints and the TES charging/discharging rate. The ranges used for the zone temperatures were from 21[degrees]C (70[degrees]F) to 29.5[degrees]C (85[degrees]F) [i.e., floating] for the period from 12 am to 8 am, from 21[degrees]C (70[degrees]F) to 24[degrees]C (75[degrees]F) during the occupied period of 8 am to 5 pm (to maintain thermal comfort), and again from 21[degrees]C (70[degrees]F) to 29.5[degrees]C (85[degrees]F) from 5 pm to midnight. This range allowed for up to 15 hours of precooling if needed. The range for the TES charging/discharging rate was from 100% charging to 100% discharging for all hours of the day.
24-Hour Update Window
First, an optimized control test was conducted with a 24hour update-window. Thus, during the day (24-hour period), only one optimization simulation was carried out. The setpoints output from that simulation run were used for the entire 24 hour period during the test. The results from this test called for a full 8 hours of precooling at 21 [degrees]C (70[degrees]F), followed by setpoints of 24[degrees]C (75[degrees]F) for the entire occupied period. The zone temperatures were left to float after 5 pm. Therefore, this test was performed with the identical setpoints to the 8 hour precooling test discussed above. Figure 11 and Table 6 show the results.
[FIGURE 11 OMITTED]
Table 6. Building and Model Energy and Operation Cost Comparison for 24-hour Predictive Optimal Control Test Energy Peak Energy Cost Demand Cost Total Cost Consumption Demand (kWh) (kW) Actual 1379 93 $22.61 $34.99 $57.60 Model 1334 92 $21.88 $34.53 $56.41 % Difference 3.2% 1.3% 3.2% 1.3% 2.1%
The simulation model results are within 5% of the actual data, indicating that the model was well-calibrated for this control strategy.
6-Hour Update Window
In this test, the temperature settings and the discharging/ charging rates are updated from the simulation runs every 6 hours. This would allow for some adjustment during the last 2 hours of precooling, depending on the changes in weather from the previous update. For this test, the precooling temperature setpoint changed by 0.1[degrees]C (0.2[degrees]F) between the first and second updates. For the other updates, the optimal setpoints remained the same, 24[degrees]C (75[degrees]F) for occupied periods and floating for unoccupied periods. The results are shown in Figure 12 and Table 7.
[FIGURE 11 OMITTED]
Table 7. Building and Model Energy and Operation Cost Comparison for 6-Hour Predictive Optimal Control Test Energy Peak Energy Cost Demand Cost Total Cost Consumption Demand (kWh) (kW) Actual 1397 95 $22.90 $35.51 $58.42 Model 1342 93 $22.01 $34.78 $56.79 % Difference 3.9% 2.1% 3.9% 2.1% 2.8%
4-Hour Update Window
This test was a predictive optimal control test with a 4 hour update window. This would again allow for altered precooling setpoints depending on the weather conditions. Also, due to a one to two hour delay for the updated weather information, this test was the shortest moving-window predictive control test. Table 8 summarizes the results of this test.
Table 8. Building and Model Energy and Operation Cost Comparison for 4-Hour Predictive Optimal Control Test Energy Peak Energy Cost Demand Cost Total Cost Consumption Demand (kWh) (kW) Actual 1413 96 $23.18 $35.92 $59.09 Model 1366 95 $22.41 $35.71 $58.12 % Difference 3.3% 0.6% 3.3% 0.6% 1.6%
Several conclusions can be drawn from the results of the field tests presented in this paper. The simulation environment for modeling the building thermal behavior using Energyplus worked well for all the tests. Daily simulation results for energy consumption, peak demand, and utility cost were within 5% of the actual building data. This indicates that the modeling procedure and the simulation environment can be applied successfully to control the cooling system for the elementary school considered in this study for different ranges of weather and utility rate conditions.
The results have indicated that the utility rate had a significant impact on the optimal control strategy. This was evidenced when the energy rate schedule was changed (with a higher incentive for load shifting), resulting in the need for precooling of the building instead of conventional control. For the different optimal control tests with different weather conditions the optimal control for elementary school called for precooling to 70[degrees]F for 8 hours prior to occupancy. Due to thermal comfort concerns, the setpoints for precooling were limited to 21[degrees]C (70 [degrees]F).
Since none of the tests were performed on identical days, it is not possible to compare the actual utility cost savings due to precooling controls and/or optimal controls. For the actual utility rate, however, optimized control simulations called for no precooling. Since the cooling load is fully met by the TES there is no need to shift load at all unless peak demand can be reduced. However, when the high-incentive rate is used, a shift of cooling load using precooling was recommended by the optimal control.
It was, however, possible to compare the impact of different control strategies for this building on identical days using the calibrated model. The results for all tests demonstrated that the simulation model accurately represents the actual performance of the elementary school. Therefore, the simulation environment should give a reasonable estimate of the impact on costs for the actual building for varying control strategies under the same weather and occupancy conditions. Table 9 shows the energy and cost data for the simulated building using several control strategies.
Table 9. Standard Rate Comparison of Simulated Control Strategies for One Typical Day Control Strategy Conventional 4-Hour Precool 8-Hour Precool Energy $25.71 $26.37 $26.39 Standard Demand $40.01 $39.85 $39.80 Rate Total $65.72 $66.21 $66.19 High Energy $162.77 $161.70 $161.38 Incentive Demand $40.01 $39.85 $39.80 Rate Total $202.78 $201.55 $201.18 Control Strategy NO TES Conventional Optimal Control High Incentive Energy $36.65 $25.75 Standard Demand $86.93 $40.02 Rate Total $123.57 $65.77 High Energy $299.78 $162.10 Incentive Demand $86.93 $40.02 Rate Total $386.71 $202.12
For the actual utility rate, conventional (i.e., current) control strategy is the most cost effective. This control saves 47% of utility charges over a standard building with no ice storage system using conventional control. However, when the high incentive rate structure was used, precooling became slightly advantageous yielding only a 0.5% savings. This indicates that precooling is not an effective control strategy for this building.
In summary, the simulation environment was found to be effective to implement real-time predictive optimal controls for both passive and active TES systems for the building considered in this study. Additional field studies are needed to assess the performance of the simulation environment for different climates and for other building types.
Andreson, I., and Brandemuehl, M.J., 1992, "Heat Storage in Building Thermal Mass: A Parametric Study," ASHRAE Transactions, 98.
ASHRAE. 2007, Applications Handbook, "Chapter 34: Thermal Storage", American Society of Heating, Refrigeration, and Air-Conditioning Engineers, Atlanta, GA.
Braun, J.E. (1990), "Reducing Energy Costs and Peak Electrical Demand Through Optimal Control of Building Thermal Storage," ASHRAE Transactions, 96(2), pp. 876-888.
Braun, J.E., Montgomery, K.W., and Chaturvedi, N., 2001, "Evaluating the Performance of Building Thermal Mass Control Strategies," International Journal of HVAC&R Research, 7, pp. 403-428.
Braun, J.E., 2003, "Load Control Using Building Thermal Mass," Journal of Solar Energy Engineering, Vol. 125, pp. 292-301.
Chen, T.Y., 2001, "Real-Time Predictive Supervisory Operation of Building Thermal Systems with Thermal Mass," Energy Build., 33, pp. 141-150.
Conniff, J.R, 1991, "Strategies for Reducing Peak Air-Conditioning Loads by Using Heat Storage in the Building Structure," ASHRAE Transactions, 97, pp. 704-709.
Crawley, D.B., Lawrie, L.K., Pedersen, CO., Winkelmann, EC, 2000, "Energy Plus: Energy Simulation Program," ASHRAE Journal, Vol. 42, No. 4 (April), pp. 49-56.
Guven, H., and Flynn, J., 1992, "Commissioning TES Systems," Heating/Piping/Air Conditioning, Vol. 64, pt. 1.
Henze, G.P., and Krarti, M., 1998, "Ice Storage System Controls for the Reduction of Operating Cost and Energy Use," Journal of Solar Energy Engineering, Nov. 1998.
Henze, G.P., and Krarti, M., 2002, "Predictive Optimal Control of Active and Passive Building Thermal Storage Inventory--Final Report for Phase I."
Henze, G.P., 2003, "An Overview of Optimal Control for Central Cooling Plants with Thermal Energy Storage," Journal of Solar Energy Engineering, Vol. 125, pp. 302-309.
Henze, G. P., Felsmann, C, Knabe, G., 2004, "Evaluation of Optimal Control for Active and Passive Building Thermal Storage," International Journal of Thermal Sciences, 43, pp. 173-181.
Ihm, P., Krarti, M., and Henze, G.P, 2004, "Development of a Thermal Energy Storage Model for EnergyPlus," Energy and Buildings, 36, pp. 807-814.
Keeney, K.R., and Braun, J.E., 1997, "Application of Building Precooling to Reduce Peak Cooling Requirements," ASHRAE Transactions, 103, pp. 463-469.
Krarti, M., Henze G. Zhou G., Ihm P, Liu S., and Morgan S., 2007, "Real-Time Predictive Optimal Control of Active and Passive Thermal Energy Storage systems. Final Report", report to Department of Energy, Contract DE-FC-36-03G) 13026.
Morris, F.B., Braun, J.E., and Treado. S.J., 1994, "Experimental and Simulated Performance of Optimal Control of Building Thermal Storage," ASHRAE Transactions, 100, pp. 402-414.
Rabl, A., and Norford, L.K., 1991, "Peak Load Reduction by Preconditioning Buildings at Night," International Journal of Energy Research, 15, pp. 781-798.
Sohn, C.W., 1991, "Field Measurement of an Ice Harvester Storage Cooling System," ASHRAE Transactions, Vol. 97, pt.2, pp. 1187-1193.
Tran, N., Kreider, J.R, and Brothers, P., 1989, "Field Measurement of Chilled Water Storage Thermal Performance," ASHRAE Transactions, Vol. 95, pt. 1.
Zhou, G., Ihm, P., Krarti, M., Liu, S., and G.P. Henze, 2005, "Integration of an Internal Optimization Module within EnergyPlus" IBPSA Proceedings.
Stephen Morgan is a graduate student and Moncef Krarti is a professor and associate chair in the Civil, Environmental, and Architectural Engineering Department at the University of Colorado, Boulder, CO.
|Printer friendly Cite/link Email Feedback|
|Author:||Morgan, Stephen; Krarti, Moncef|
|Date:||Jan 1, 2010|
|Previous Article:||Consumption analysis of telco and data center cooling and humidification options.|
|Next Article:||Energy efficient TES designs for commercial DX systems.|