ESL-HH-08-12-42

STUDY OF THE OUTSIDE AIR EFFECTS IN THE SCREENING OF METERED BUILDING ENERGY DATA

Jing Ji, Ph.D. Juan Carlos Baltazar, Ph.D. David E. Claridge, Ph.D., P.E. Research Associate Associate Research Engineer Professor / Director

Energy Systems Laboratory, Texas Engineering Experiment Station Texas A&M University, College Station, TX

ABSTRCT INTRODUCTION The energy consumption in a building is affected The methodology of energy balance load has by many parameters including the occupancy, been proposed as a tool for screening building equipment, schedule time, HVAC systems and energy data which are separately recorded outside environment conditions. Currently, the according to individual heating, cooling and outside air dry-bulb temperature (TOA) is the electricity consumptions (Shao and Claridge, primary variable used in the data driven analysis 2006). Although different air handle units for building energy use, including development (AHU) have different energy use patterns on of energy consumption models and measurement individual heating and cooling, the air side of savings. The measured building energy data simplified simulation shows that the defined EBL analysis based on the variable of TOA has the has the similar pattern for four basic AHUs. drawback of overlooking several effects such as (CVRH: single constant air volume with , wind and solar. terminal reheat; DDCV: dual duct constant air volume; DDVAV: dual duct ; The application of outside air enthalpy (hOA) on a and SVAV: single duct variable air volume) methodology name as “Energy Balance” for (Shao, 2005). The methodology of energy screening building energy data to study the balance load provides the prospect application in influence of outside air humidity on building the analysis of energy use for buildings having energy analysis has been presented in this paper. various AHUs. This methodology has been The variable of hOA has been implemented in a implemented into energy use analysis for quality developed screening tool, which is the assurance of the energy use data for buildings on application of first law energy balance, to the Texas A&M University campus (Baltazar et analyze the building energy use. The energy al., 2007). balance load (EBL) for a whole building, which is the difference between the heating requirements It was found the slope and intercept of the plus the electric gain and the cooling load, has defined EBL is changed with buildings. It is been presented by the two variables of hOA and necessary to further study the effects of TOA, respectively. A design of experiment parameters and building functions on the EBL process is conducted to study the linear relations behaviors. Generally, the outside air temperature for the EBL as the function of hOA and as the has been used to study the EBL behaviors. function of TOA. The comparison results lead the However, the pattern for EBL in the high conclusion that the EBL in the high temperature temperature range has less linearity with the TOA. range could be better presented with the application of hOA instead of TOA. The energy This paper presents the analytical study of the analysis for buildings located in hot and humid behaviors for EBL as function of TOA and hOA. The climate would be better performed by using hOA. slope and cross point of the EBL have been Study cases are also presented to illustrate the investigated through simplified air side model difference between application of hOA and TOA in simulation, the analytical model, and validation energy use data analysis for buildings with by actual energy use data. The parameters effects different functions. The statistics study shows have also been demonstrated using the method of that the energy use analysis for buildings design of experiments (DOE) to characterize the classified as laboratory would be improved in the various effects of parameters. In order to

application of hOA as the variable instead of TOA. improve the analysis of building consumption in high temperature, the enthalpy of outside air is used in study the behaviors of EBL. Additionally,

1

Proceedings of the Sixteenth Symposium on Improving Building Systems in Hot and Humid Climates, Plano, TX, December 15-17, 2008 ESL-HH-08-12-42

the paper illustrates the application of hOA to windows and air handling units, QCon is the heat analyze the energy data in summer. transmission through the building structure, and QOcc is the heating gain from occupants. ANALYTICAL STUDY OF EBL Slope of the EBL as the function of TOA The energy balance load (EBL) is defined from the whole building thermodynamic model based Eqn. 2 indicates that the slope of EBL as a on the analytical redundancy (Shao and Claridge, function of TOA depends on the amount of intake 2006). The equation to represent the EBL is outside air and the UA values. expressed as: WbeleE Wbheat−+= Wbcool BL E Wbele Wbheat −+= Wbcool ( +++−= QQQQ ) B L ConAirSol Occ V ( Sol Air Con +++−= QQQQ Occ ) (1) ( +−= ρ +− ρhTTCVQ OA −ww )(([)( fgzOAPOAsol X CLR OA + +−+−+ QTTUAwwV ))())]( Where Wbele is the whole building electricity ROAOA OCCzOA ( ρ ())( +−−+−= QQTTUACV use for lighting and equipment in the building, POA occsolzOA −ww )( Wbheat is the input heating to maintain the + ρ Vh ([ CLR −+ ww +)))]( OAfg X ROA conditions in a building, and Wbcool is the input OA (2) cooling to maintain the conditions in a building. QSol is the solar heating gain, QAir is the ventilation and air via doors, A CVRH OA ratio effects B CVRH UA effects 2.5 CVRH OA 5% 2.5 2.0 CVRH UA 5000 Btu/(°F·hr) CVRH OA 10% 2.0 CVRH UA10000 Btu/(°F·hr) 1. 5 CVRH OA 15% 1. 5 1. 0 CVRH UA 15000 Btu/(°F·hr) 1. 0 0.5 0.5 0.0 0.0 -0.5 -0.5 -1.0

(MMBtu/Hour) -1.0 (MMBtu/Hour) -1.5 -1.5 Energy Balance Load -2.0 Load Balance Energy -2.0 -2.5 -2.5 0 102030405060708090100110 0 102030405060708090100110 o Outside Air Temperature ( F) Outside Air Temperature(oF)

(a) (b) Figure 1 Plots of energy balance load vs. outside air temperature with the variation of Intake outside air value (a), and the UA value (b)

Energy Balance Load vs. OA Temperature Bib Data : Energy Balance Load vs. OA Temperature 1,200 1,200 EBL-Building 1 Lab EBL- Building 1 Lab 900 EBL-Building 2 Library 900 EBL- Building 2 Library EBL-Building 3 Residence Hall EBL-Building 3 Residence Hall 600 600 ) ) 2 2 300 300

- -

(300) (300) (Btu/day/ft (Btu/day/ft (600) (600) Energy Balance Load Energy Balance Load Energy Balance (900) (900)

(1,200) (1,200) 0 102030405060708090100110120 0 102030405060708090100110120 Outside Air Dry-bulb Temperature (°F) Outside Air Dry-bulb Temperature (°F)

(a) (b) Figure 2 Plots of energy balance load vs. outside air temperature for three different buildings using the daily data(A) and bin data (B)

Figure 1 shows simulation results of the slope results have been achieved in the other three change with the change of the value of intake systems (DDCV, DDVAV, and SDVAV). outside air and the UA value in the CVRH Figure 2 shows the plot of energy balance load system ((a). change of the value of intake outside vs. outside air temperature (EBL vs. TOA) using air; (b). change of the UA value). The same the actual data for three buildings with difference

2

Proceedings of the Sixteenth Symposium on Improving Building Systems in Hot and Humid Climates, Plano, TX, December 15-17, 2008 ESL-HH-08-12-42

functions ((a) using the daily data; (b) using the have the clear linearly relationship with TOA bin data). The value of EBL in Figure 2 is when the latent load existed in high temperature, normalized with the area of a building with the as shown in Figure 4 (a), but Figure 4 (b) shows 2 unit of Btu/ft /day. The slope of EBL in building a better linear behaviors when the enthalpy of with the function of laboratory is steeper than outside air is applied to present the pattern of other two buildings which are library and EBL. residence hall, respectively. Following derived equations confirmed that the EBL has more linear behaviors as the function of Cross point temperature in the EBL enthalpy.

The cross point temperature is defined as the = dpa + ( , refg + TchwTch dpw ) (7) temperature at the EBL equal to zero. Eqn. 3 Cpa: Specific heat of dry air shows the explanation of the items of QAir, and Td: Dry bulb temperature QCon. Cpw: Specific heat of superheated water vapor hg,ref: Enthalpy of water vapor at appropriate

, += QQQ ,LatAirsenAirAir reference temperature

= ρ TTCV ZOAPOA )( +− w: Humidity ratio

VOA VOA + ρhfg 1 )(([ wVXw OAOAOAR −+− wCL)] BL = −(E ρ POA + +() ρ POAOA + () QQTUACVTUACV occsolZ ++− XOA XOA XR + [ρ OAOAfg + ρ VhwVh OAfg −ρ wVhw CLTotfgR ))] XOA

Con −= TTUAQ ZOA )( −= [( OAPOA + ρρ OAOAfg UATwVhTCV OA]) ++ X (3) [( ρ ) −+ ρ VhTUACV R w ] POA OAfgZ X R Substitute Eqn. 3 into Eqn. 1, Eqn. 4 has been OA ( −+− ρ wVhQQ ) achieved: occsol CLTotfg (8) BL WbeleE Wbheat−+= Wbcool ( +++−= QQQQ ) Compare the left of Eqn. 7 with the item of AirSol Con Occ (V ρC T +ρh V w ), the Eqn. 8 can be ( +−= ρ −TTCVQ )( OA p OA fg OA OA zOAPOAsol expressed with similar Eqn. 9. ,LatAir +−++ QTTUAQ OcczOA ))( (4) − hwh ,refgOAOA BL [( ρ OAOA +−≈ UAhVE ] The boundary condition of EBL=0 occurred at the + cwc pwOApa

cross point temperature, [( ρ ) +++ −ρρ wVhwVhTUACV RTotfgROAfgZPOA ]

( −+− ρ wVhQQ CLTotfgoccsol ) BL (QE +−= V ρC T T )( +− Q ,LatAirzOAPOAsol UA UA (V ρ+−≈ () Vh ρ++ )h QTTUA ))( =+−+ 0 OA + cwc OAOA + cwc Z OcczOA (5) pwOApa pwRpa hw hw Rearrange the Eqn.5: +UA( ,refgOA + ,refgR ) + cwc + cwc pwOApa pwRpa ρ ( −+−− ρ wVhQQwVh CLTotfgoccsolRTotfg ) Sal ,LatAir ++ QQQ Occ TT −= (9) ZOA ρ + UACV POA The Eqn. 9 indicates that EBL as a function of (6) hOA with the slope of UA From the Eqn. 6, it could be drawn the (Vk ρ +−= ) OA + cwc conclusion that the cross point temperature is pa pwOA always lower than the zone temperature. and the interception of UA (Vb OAρ += )hZ Application of enthalpy (hOA) in the EBL + cwc pwRpa methodology hw ,refgOA hw ,refgR Figure 3 shows the energy balance sensible load + UA( + ) pa + cwc pwOA + cwc pwRpa and latent load as the function of T using the OA − ρ ( −+− ρ wVhQQwVh ) simplified air side model simulation. The figure solRTotfg occ CLTotfg shows that the energy balance sensible load has the linear relationship with TOA, but the energy balance latent load doesn’t have the linear relationship with TOA. Because of this phenomena, the energy balance pattern didn’t

3

Proceedings of the Sixteenth Symposium on Improving Building Systems in Hot and Humid Climates, Plano, TX, December 15-17, 2008 ESL-HH-08-12-42

A Energy Balance Sensible Load vs. OA Temperature B Energy Balance Latent Load vs. OA temperature 1.5 1.5 CV RH CV RH DDCV 1.0 DDCV DDVAV 1.0 SDVAV DDV A V 0.5 0.5 SDVAV

0.0 0.0

-0.5 -0.5

-1.0 -1.0 Energy Balance Load Load Balance Energy Energy Balance Load Load Balance Energy (Sensible) (MMBtu/Hour) -1.5 (Latent) (MMBtu/Hour) -1.5

-2.0 -2.0 0 102030405060708090100110 0 102030405060708090100110 o o OA Temperature( F) OA Temperature( F) (a) (b) Figure 3 The simplified air side model simulation results of (a) Energy balance sensible load vs. Outside air temperature and (b) Energy balance latent load vs. Outside air temperature for the four basic AHUs

Energy Balance Load (EBL) vs. OA Enthalpy A Energy Balance Load (EBL) vs. OA Temperature B 1.5 1.5 CV RH CV RH 1.0 DDCV 1.0 DDCV DDV A V DDV A V 0.5 0.5 SDVAV SDVAV 0.0 0.0

-0.5 -0.5 (MMBtu/Hour) (MMBtu/Hour) -1.0 -1.0 Energy Balance Load Energy Balance Load Load Balance Energy -1.5 -1.5

-2.0 -2.0 0 102030405060708090100110 0 5 10 15 20 25 30 35 40 45 o OA Temperature( F) OA Enthalpy (Btu/lb) (a) (b) Figure 4 The simplified air side model simulation results of (a) Energy balance load vs. Outside air temperature and (b) Energy balance load vs. Outside air enthalpy for the four basic AHUs

Figure 5 (a) shows the plot EBL and energy use of hOA using the daily consumption data. Figure 6 electricity (ELE), (CHW) and show the similar plots to Figure 5 with the bin heating hot water (HHW) as the function of TOA data. and Figure 5 (b) shows the plot EBL and energy use of electricity (ELE), chilled water (CHW) and heating hot water (HHW) as the function of

1,600 1,600 Wbebl Wbele Wbco o l Wbheat Wbebl Wbele Wbco o l Wbheat ) ) 2

1,200 2 1,200

800 800

400 400

- -

(400) (400)

(800) (800)

(1,200) Energy Consumption (Btu/day/ft

Energy C onsum ption(1,200) (B tu/day/ft

(1,600) (1,600) 0 102030405060708090100110120 0 5 10 15 20 25 30 35 40 45 50 Outside Air Dry-bulb Temperature (°F) Enthalpy (Btu/lb) (a) (b) Figure 5 (a) Plot of the EBL, energy use of ELE, CHW and HHW vs. OA temperature (b) Plot of the EBL, energy use of ELE, CHW and HHW vs. OA Enthalpy in a building using daily data

4

Proceedings of the Sixteenth Symposium on Improving Building Systems in Hot and Humid Climates, Plano, TX, December 15-17, 2008 ESL-HH-08-12-42

1,600 1,600

1,200 1,200

800 800 ) ) 2 2 400 400

- -

(400) (400) (Btu/day/ft (800) (Btu/day/ft Energy Balance Balance Energy

Energy Balance Energy Balance (800)

(1,200) (1,200)

(1,600) (1,600) 0 102030405060708090100110120 0 5 10 15 20 25 30 35 40 45 50 Outside Air Enthalpy (Btu/lb) Outside Air Dry-bulb Temperature (°F) (a) (b) Figure 6 (a) Plot of the EBL, energy use of ELE, CHW and HHW vs. OA temperature (b) Plot of the EBL, energy use of ELE, CHW and HHW vs. OA Enthalpy in a building using daily bin data

with zero mean, and the plotted effects will lie FACTORS EFFECTS INVESTIGATION approximately along a straight line. Those effects FOR THE EBL that do not plot on the line are significant factors. (Montgomery and Runger, 2004) In the normal The method of full factorial design was used to probability plot of the effect, the x-axis study the effects of input parameters on E BL represents the effects, and the y-axis represents behaviors. A factorial design is widely used the cumulative probabilities. The scale of y-axis when there are several factors of interest in the is constructed in such a way that if the data experiments. In such designs factors are varied points follow a normal distribution, the together. Specifically, with a factorial cumulative probabilities will plot as a straight experiment, all possible combinations of the line. For effects that are from a normal levels of the factors are investigated in each distribution with mean zero, the plot of the complete trial or replicate of the experiment. The effects should approximate a straight line with effect of a factor is defined as the change in the line passing through the point (x=0,y=0.5). response produced by a change in the level of the Significant effects much different from 0 will fall factor. Five parameters are intake OA ratio, cold away from this line. Effects that are unusually deck temperature, zone temperature, UA value, small or large and fail to follow the straight line and occupant density. In this factorial design, all pattern are judged to be significant. (Ledolter five factors have two levels, denoted by “-1” and and Swersey, 2007) “+1”. Table 1 shows the set point for the

parameters For the 25 design in the operation, the number of Table 1 Parameter level for factor design estimated effects is 31× (25-1). The procedure for -1 +1 plotting normal probability of the effect is the OA Ratio 5% 15% following. First, the 31 effects have been ordered o o TCL 45 F 65 F from small to large. The smallest among the 31 Tz 68oF 82oF effects represents a cumulative probability UA 5000 15000 between 0 and 1/31 and is assigned a cumulative Btu/(hr*F) Btu/(hr*F) probability (y-value) at the midpoint of that Occupant 300 ft2/person 500 ft2/person interval. The second smallest among the 31 density effects represents a cumulative probability between 1/31 between 2/31 and is assigned a y- The corresponding effects of factors are analyzed value at the midpoint of that interval. The third by the normal probability plot of the effect and smallest effect is assigned a cumulative the Pareto charts. The normal probability plot of probability at the midpoint of the interval 2/31 to the effect estimates is a very helpful method in k 3/31, and so forth. In general, with m effects, the judging the significance of factors in a 2 ith smallest effect is plotted at a cumulative experiment, especially when many effects are to probability of ( i-0.5)/m. (Ledolter and Swersey, be estimated. If none of the effects are 2007) significant, then the estimates will behave like a random sample drawn from a normal distribution

5

Proceedings of the Sixteenth Symposium on Improving Building Systems in Hot and Humid Climates, Plano, TX, December 15-17, 2008 ESL-HH-08-12-42

The software Minitab was used to analyze the relationship either TOA or hOA according to the parameter effects on the slope and intercept of simplified air side model simulation results. the simulation results. The scale on the Generally, the relationship can be expressed as: cumulative probability (y-axis) in the normal y=k1x + k2 probability of the effect, such as Figure 7, Figure 8, Figure 9 and Figure 10, is not linear. This is Figure 7and Figure 8 show the parameter effects where the normal distribution comes into play. on the EBL when the TOA is used in presenting the In the normal probability plot of the effect, the function. Figure 7 show the plot of the parameter insignificant effects and significant effects have effects on the slope (variable: k1, (a): the normal been distinguished by fitting a straight line to the probability plot; (b): the Pareto plot.) Those middle portion of the graph. The value of plots show that the outside air ratio and the UA Lenth’s pseudo standard error (PSE) (Lenth, values have the significant negative effect on the 1989) is also added in normal probability plot in slope while the cold deck temperature has the Minitab. The pseudo standard error (PSE) is significant positive effect on the slope. Other defined as: factors (zone temperature and occupant density)

PSE 5.1 ×= mmedian i have fewer effects on the slope of the energy The Pareto chart is a special type of bar chart balance load. Figure 8 shows the plot of the where the values being plotted are arranged in parameter effects on the intercept (variable: k2, descending order. In Minitab, the line is drawn at (a): the normal probability plot; (b): the Pareto the margin of error (ME) on the Pareto chart. The plot). The effects of factor A (OA ratio) and D margin of error is defined as: (UA value) are far away from the normalized line. The plots indicate the OA ratio and UA

tME α n),2/( ×= PSE value have the highest influences on the intercept To analyze the parameter’s effect on energy of the assumed regression line. balance load, it is assumed the energy balance load variable has an approximately linear

Normal Plot of the Effects Pareto Chart of the Effects (response is slope, x is OA temperature, Alpha = 0.05) (response is slope, x is OA temperature, Alpha = 0.05, only 30 largest effects shown) 99 23 Effect Type B A Factor Name Not Significant B A OA Ratio(XOA) 95 D AB Significant AB BCold Deck Temp(Tcl) BD C 90 Factor Name BC CZone Temp(Tz) E BCD BE A OA Ratio(XOA) DUA Value AE CD 80 BCold Deck Temp(Tcl) BD EOcc.Density E 70 CZone Temp(Tz) ABC AC 60 DUA Value BE EOcc.Density ABE 50 BDE 40 DE

Term AE Percent DE CDE 30 ABEBDE AC BCDE ABC CE 20 CD BCE BCD AD BC ACE 10 C ABCE ABCDE 5 D ABD ACDE ABCD A ADE ACD 1 -25000 -20000 -15000 -10000 -5000 0 5000 10000 0 5000 10000 15000 20000 25000 Effect Effect Lenth's PSE = 10.5003 Lenth's PSE = 10.5003 (a) (b) Figure 7 (a) Parameters effects on the slope (x-TOA, Y-EBL) and (b) Pareto chart of parameters effects on the slope (x-TOA, y-EBL)

Normal Plot of the Effects Pareto Chart of the Effects (response is intercept, x is OA Temperature, Alpha = 0.05) (response is intercept, x is OA temperature, Alpha = 0.05, only 30 largest effects shown) 99 4884 Effect Type A A Factor Name Not Significant D BC A OA Ratio(XOA) 95 Significant AB D AC BCold Deck Temp(Tcl) AC BD 90 Factor Name BCD CZone Temp(Tz) BCD CD CD A OA Ratio(XOA) B DUA Value B ABC 80 ABC BCold Deck Temp(Tcl) BE EOcc.Density E E 70 ABE CZone Temp(Tz) ABE AE 60 DUA Value CE EOcc.Density BCE

50 Term BCDE CDE 40 ACD ABDE Percent ADE 30 ABCD ACDE 20 ABD ABCDE AD BE ACE 10 BD ABCE C 5 AB BDE

BC 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 400000 800000 1200000 1600000 2 4 6 8 10 12 14 16 Effect Effect Lenth's PSE = 2201.50 Lenth's PSE = 2201.50 (a) (b) Figure 8(a) Parameters effects on the intercept (x-TOA, y-EBL) (b) Pareto chart of parameters effects on the intercept (x-TOA, y-EBL)

6

Proceedings of the Sixteenth Symposium on Improving Building Systems in Hot and Humid Climates, Plano, TX, December 15-17, 2008 ESL-HH-08-12-42

Alternative analysis is using the hOA instead of implies that the effects of the correlative items TOA for analysis of parameters effects on energy could be neglected in the engineering balance load. Figure 9 shows the plot of the application. Figure 10 shows the plot of the parameter effects on the slope (variable: k1, (a): parameter effects on the intercept (variable: k2, the normal probability plot; (b): the Pareto plot). (a): the normal probability plot; (b): the Pareto Similar to the plots of parameter effects using plot). It is obvious that the outside air ratio and TOA, it is shows that the outside air ratio and UA the UA value have the significant positive effect value have the significant negative effect on the on the intercept. It presents that more conclusive slope while the cold deck temperature has the results of parameter effects would be drawn by significant positive effect on the slope. Other using the enthalpy in analysis instead of TOA factors (zone temperature and occupant density) based on the Figure 10, which shows that fewer have fewer effects on the slope of the energy items are marked having significant effects on balance load. Additionally, the correlative items the intercept of the energy balance load. are distributed along the normalized line. It

Normal Plot of the Effects Pareto Chart of the Effects (response is slope, x is enthalpy, Alpha = 0.05) (response is slope, x is enthalpy, Alpha = 0.05, only 30 largest effects shown) 99 44 Effect Type B A Factor Name Not Significant B A OA Ratio(XOA) 95 AB D Significant AB B C old Deck Temp(Tcl) BD C 90 E Factor Name BC CZone Temp(Tz) BE BCD DUA Value 80 AE A OA Ratio(XOA) CD BCDE BD E O cc.Density CDE B C old Deck Temp(Tcl) E 70 CZone Temp(Tz) ABC 60 AC DUA Value BE 50 E O cc.Density ABE 40 AE DE BDE Percent

30 BDE Term DE ACABE BCDE ABC CDE 20 CD BCD BCE BC CE 10 ABCE C ACE D AD 5 ABCDE ABD A ACDE ADE 1 ABCD ABDE 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 10000 20000 30000 40000 50000 6 5 4 3 2 1 1 2 3 ------Effect Effect Lenth's PSE = 19.9615 Lenth's PSE = 19.9615 (a) (b) Figure 9 (a) Parameters effects on the slope (x-hOA, y-EBL); (b) Pareto chart of parameters effects on the slope (x-hOA, y-EBL)

Normal Plot of the Effects Pareto Chart of the Effects (response is intercept, x is enthalpy, Alpha = 0.05) (response is intercept, x is enthalpy, Alpha = 0.05, only 30 largest effects shown) 99 86479 Effect Type A A Factor Name Not Significant D A OA Ratio(XOA) 95 BC D Significant B BCold Deck Temp(Tcl) B AC 90 Factor Name AB CZone Temp(Tz) AC BE A OA Ratio(XOA) ABE DUA Value 80 BCold Deck Temp(Tcl) CE EOcc.Density BCE 70 CZone Temp(Tz) BCDE CDE 60 DUA Value ABDE EOcc.Density ADE 50 ABD 40 AD

Term BDE Percent DE 30 ABCE ABCDE 20 ACE ACDE ABCD 10 ACD BD 5 AE BCD CD BC E 1 C 0 500000 1000000 1500000 0 200000 400000 600000 800000 1000000 1200000 1400000 Effect Effect Lenth's PSE = 38982.0 Lenth's PSE = 38982.0 (a) (b) Figure 10 (a) Parameters effects on the intercept (x-hOA, y-EBL); (b) Pareto chart of parameters effects on the intercept (x-hOA, y-EBL)

metering data for electricity, heating and cooling APPLICATION OF EBL IN THE consumption. The methodology of the EBL as a ENERGY DATA QUALITY function of TOA has the limitation in analysis of ASSURANCE the summer data because of the high temperature and limited data range. Since the latent load in The methodology of the EBL have been applied to summer is larger than that in other seasons, and analyze the energy data to fulfill the quality the latent load also varied with the buildings, it is assurance of utility bills (Baltazar et al , 2007). It difficult to analyze the energy data in one month is has been approved to be an effective data in summer. Application of h in presenting E quality screening method for buildings having OA BL 7

Proceedings of the Sixteenth Symposium on Improving Building Systems in Hot and Humid Climates, Plano, TX, December 15-17, 2008 ESL-HH-08-12-42

for summer data has the advantage in better building during summer in College Station, TX representative patterns of building energy use ((a): the EBL and energy consumptions as a than using the TOA. Figure 11 shows the plot of function of the TOA; (b): the plot of EBL, energy EBL for three buildings on campus during August consumptions as a function of the hOA). From the 2007 ( a: EBL as a function of TOA ; b EBL as a plot of Figure 12 (a), it is very hard to detect the function of hOA ). The cross point temperature in failure in the consumption for the building since EBL as function of TOA has the large range from the data is drawn on a scale that may make the negative to positive value. It is very hard to use slope difficult to notice. Energy consumption the slope and intercept to apply the methodology trends are also correct. The bin data then was of the energy balance in the data quality grouped to analyze the energy use for the assurance. With the application the hOA in building as shown in Figure 13. The linear model analysis of EBL, the ranges of slope and intercept equation for the EBL shown in Figure 13 (b) has have been narrowed down. The equations to the negative x-intercept. That means that the present the EBL pattern have been generated in building balance condition is at very low Figure 11 (a) and (b). It shows that the linearity enthalpy. It indicates that there are some errors in of EBL could be expressed better through hOA in the energy consumption. After further check, it stead of TOA in summer. was found that the CHW flow meter has wrong readings. Figure 12 shows the energy use (ELE, CHW and HHW) and energy balance load for a office

Energy balance Load vs. OA temperature Energy balance Load vs. OA enthalpy

- - ) ) Building 1 Building 1 2 2 (50) Building 2 (50) Buidling 2 Building 3 Building 3 (100) (100) y = -5.2962x + 72.845 y = -1.7097x + 6.2446 (150) (150) y = -6.8326x + 106.31 y = -1.1462x - 68.679 (200) (200)

y = -7.2081x + 36.778 (250) y = -0.3758x - 219.61 (250) Energy Balance (Btu/day/ft Energy Balance (Btu/day/ft Balance Energy (300) (300) 0 102030405060708090100110120 0 5 10 15 20 25 30 35 40 45 50 Outside Air Enthalpy(Btu/lb) Outside Air Dry-bulb Temperature (°F) (a) (b) Figure 11 (a) Plot of EBL vs. TOA and (b) Plot of EBL vs. hOA for three buildings during August 2007.

1,000 1,000 ) )

EBL 2 EBL 2 800 Wbele 800 Wbele 600 Wbcool 600 Wbcool Wbheat Wbheat 400 400 Linear (EBL) Linear (EBL) 200 200 - - (200) (200) (400) (400) (600) (600) y = -10.874x - 136.23 y = -9.0928x + 190.68 (800) (800) Energy Consumption (Btu/day/ft Consumption Energy Energy Consumption (Btu/day/ft Consumption Energy (1,000) (1,000) 0 5 10 15 20 25 30 35 40 45 50 0 102030405060708090100110120 Outside Air Enthalpy (Btu/lb) Outside Air Dry-bulb Temperature (°F) (a) (b) Figure 12 (a) Plots of EBL, energy use of ELE, CHW and HHW vs. OA temperature (b) Plots of EBL, energy use of ELE, CHW and HHW vs. OA enthalpy for a building during summer using daily data

8

Proceedings of the Sixteenth Symposium on Improving Building Systems in Hot and Humid Climates, Plano, TX, December 15-17, 2008 ESL-HH-08-12-42

1,000 1,000 ) ) 2 2 EBL EBL 800 800 Wbele Wbele 600 Wbcool 600 Wbcool Wbheat Wbheat 400 Linear (EBL) 400 Linear (EBL) 200 200 - - (200) (200) (400) (400) (600) (600) y = -9.7061x + 241.6 y = -9.7357x - 184.97 (800) (800) Energy Consumption (Btu/day/ft Energy Consumption (Btu/day/ft Consumption Energy (1,000) (1,000) 0 102030405060708090100110120 0 5 10 15 20 25 30 35 40 45 50 Outside Air Enthalpy (Btu/lb) Outside Air Dry-bulb Temperature (°F) (a) (b) Figure 13 (a) Plots of EBL, energy use of ELE, CHW and HHW vs. OA temperature (b) Plots of EBL, energy use of ELE, CHW and HHW vs. OA enthalpy for a building during summer using bin data

SUMMARY AND CONCLUSIONS REFERECE The further study for the methodology of the Baltazar, J. C., Sakurai,Y., Masuda, H., Energy balance load (EBL) has been presented. Feinauer, D., Liu, J., Ji, J., Claridge, D.E., Deng, The behaviors of the EBL have been investigated S., & Bruner, H. (2007). Experiences on the by the analysis of the defined mathematical Implementation of the “Energy Balance” model of whole building energy balance. The Methodology as a Data Quality Control Tool: simplified air side simulation has been conducted Application to the Building Energy Consumption th to study the patterns of EBL. The actual measured of Large University Campus, 7 International energy use data has been applied to prove the Conference for Enhanced Building Operations, conclusion of analytical study and simulation Nov. 1-2, 2007 San Francisco, California results. The factor effects analysis through the Ledolter, J., & Swersey, A.J. (2007). Test 1- experimental design lead the conclusion of that 2-3: Experimental Design with Applications in the outside intake air volume and the UA value Marketing and Service Operations, Stanford have the most negative effects on the slope of the Business Books, Stanford, California EBL pattern; while the cold deck temperature has Lenth, R.V., (1989), Quick and Easy the most significant positive effect on the slope Analysis of Unreplicated Factorials, of the EBL pattern. Technometrics, Vol. 31 No.4, 469-473 Minitab Inc. (2007), Minitab 15 Manual. The behaviors EBL have been studied using the Retrieved November 28, 2007, from variable of beside the TOA. Analysis of EBL as a http://www.minitab.com/products/minitab/docu function of hOA have the advantage to present the mentation.aspx energy data in high temperature since the latent Montgomery, D.C. and George C.R (2004). load effect can be represented though the Applied Statistics and probability for Engineers, variable of enthalpy instead of the temperature. 4th Edition, John Wiley & Sons, Inc. The EBL has a better liner relationship to the hOA Shao, X., (2005). First Law Energy Balance than the TOA. Using enthalpy in EBL analysis for as a Data Screening Tool, M.S. Thesis, energy use data provided a beneficial approach Department, Texas to analyze the energy data in summer. The A&M University, College Station, TX, May. application of the EBL as the function of hOA to Shao, X. and Claridge, D.E., (2006). Use of analyze the summer data for actual buildings has First Law Energy Balance as a Screening Tool been presented. An example of the methodology for Building Energy Data, Part I – Methodology, of enthalpy for EBL analysis applied to detect the ASHRAE Transactions - Research, Vol. 112, failure of metering data of energy consumption Part 2, QC-06-068. has been also illustrated.

9

Proceedings of the Sixteenth Symposium on Improving Building Systems in Hot and Humid Climates, Plano, TX, December 15-17, 2008