THE PERFORMANE OF ONVENTIONAL AND HUMID-LIMATE VAPOR-OMPRESSION SUPERMARKET AIR-ONDiTIONING SYSTEMS by RUTH ELLEN URBAN A thesis submitted in partial fulfillment of the requirements of the degree of MASTER OF SIENE (Mehanial Engineering) at the UNIVERSITY OF WISONSIN--MADISON 1988
AKNOWLEDGEMENTS Funding for this projet was provided by the Eletri Power Researh Institute. Bob Parkes of Parkes Assoiates, Ken Bowlen of argoaire, arl Laverrenz of Tyler, Keith Oliver and Bob Stevens of Kroger, David Knebel of Turbo, Peter Jenkins of Trane and Massoud Neshan of Hill supplied numerous diffiult-to-find piees of information. Several also reviewed model desriptions. I would like to thank my advisors John Mithell and Bill Bekman for their enthusiasm and areful guidane, even while on the other side of the Atlanti or fighting illness. huk Dorgan onsistently knew the right value for the parameter I needed, Jak Duffie smoothed away many obstales, and Sandy Klein taught me that "thermo" ould indeed be fun. My fellow students and the Solar Lab. staff reated a omfortable, helpful environment and provided many muh-needed diversions. The Jims deserve speial thanks: Jim Braun for helping me out of numerous dead-ends, and Jim Kummer for making TRNSYS and the omputers behave. A speial thank you also to Byron Prithard for writing the ie-storage omponent to be used in part two of this projet. Friends near and far shared with me assette tapes, Thursday morning breakfasts, letters and phone alls from distant plaes, silene, sunny vaations, ontra danes, meals and ommunity; making my world wide and satisfying this past year and a half. And finally, my love goes to my family. Regardless of the venture, they have always baked me.
ABSTRAT Reduing supermarket energy osts is important both for the store owner and the general publi. This study evaluates the energy onsumption patterns of two types of vapor-ompression air-onditioning systems (a onventional system and a humidlimate system) in a 30,000 ft 2 Miami, FL store. Supermarket, refrigerated ase, air-onditioning units, reheat and fan models were developed and implemented as TRNSYS omputer simulation omponents. These omponents were used to build onventional and humid-limate air-onditioning system models, whih were then used to perform annual simulations. The effets of varying limate, humidity setpoint, ventilation and irulation flow rates, refrigerated ase apaity and internal sensible load were studied. An annual air-onditioning energy onsumption savings of 63% (37,100 kwh) was found with the humid-limate system in Miami. However, as a result of the smaller demand by the humid-limate system for reheat energy relaimed from the refrigerated ase ondenser(s), the refrigerated ase OP dereased. The inreased refrigerated ase energy onsumption dereased the total annual energy onsumption savings to 1% (9,700 kwh). This represents a savings on the order of $1,000. Given a paybak period of 2 to 3 years, the ost of a humid-limate system annot exeed the ost of a onventional system by more than $600 to $1,800. The energy requirements of both systems are sensitive to limate, humidity setpoint, irulation flow rate and refrigerated ase apaity; but are not highly sensitive to ventilation flow rate or internal sensible load.
TABLE OF ONTENTS Aknowledgments... Abstrat...ii Table of ontents...m List of Tables...v List of Figures... vi Nomenlature...viii hapter 1: Introdution...1 1.1: SopeofStudy...1 1.2: Literature Review... 2 hapter 2: Supermarket Air-onditioning Systems...5 2.1: onventional Vapor-ompression System... 5 2.2: Humid-limate Vapor-ompression System... 7 2.3: General Vapor-ompression System Model...10 hapter 3: System omponents... 14 3.1: Supermarket Building... 14 3.2: Refrigerated ases...26 3.3: Vapor-ompression Air-onditioning Units... 31 3.3.1: Sensible and Latent ooling Unit...38 3.3.2: Sensible ooling-only Unit... 38 3.4: Reheat Heat Exhanger... 40 3.5: Fan Power...40 3.6: Misellaneous omponents... 41 iii
hapter 4: System Simulations... 43 4.1: Base ases... 43 4.1.1: Design Loads and System Sizing... 43 4.1.2: Energy onsumption Breakdown...45 4.1.3: onventional System vs. Humid-limate System...48 4.2: Sensitivity Analyses... 48 4.2.1: limate...48 4,2.2: Humidity Setpoint...50 4.2.3: Ventilation Flow Rate... 56 4.2.4: irulation Flow Rate...59 4.2.5: Refrigerated ase apaity...59 4.2.6: Internal Sensible Load...62 hapter 5: Eonomi onsiderations... 64 5.1: Annual Operating ost Savings... 64 5.2: Allowable First osts Inrement...67 hapter 6: onlusions and Reommendations for Future Researh...71 6.1: onlusions... 71 6.2: Reommendations for Further Researh... 73 Appendix A: TRNSYS Deks... 75 Appendix B: FORTRAN ode...90 Bibliography...121 iv
LIST OF TABLES Table Page 3.1.1 Store Model Desription...17 3.1.2 Additional Store Model Parameters Required by TRNSYS Type 19... 23 3.2.1 Design Values Assumed for Refrigerated ase OP Determination... 30 3.5.1 Pressure Drops Aross Air-onditioning System omponents...42 4.2.2.1 Air onditioner apaities for Humidity Setpoint Tests... 57 5.1.1 Flat Rate Eletriity Rate Shedules...64 5.1.2 Time-of-Use Eletriity Rate Shedules... 65 5.2.1 Eonomi Analysis Parameter Values... 70
LIST OF FIGURES Figure 2.1.1 onventional Air-onditioning System...6 2.1.2 Psyhrometri Diagram for onventional System...8 2.2.1 Humid-limate Air-onditioning System...9 2.2.2 Psyhrometri Diagram for Humid-limate System...11 2.3.1 General Vapor-ompression Air-onditioning Model...12 3.1.1 Store Model Energy and Mass Flows... 15 3.1.2 Store Model Ventilation Flow Rate Shedule...19 3.1.3 Store Model Infiltration Flow Rate Shedule...20 3.1.4 Store Model Number-of-People-in-Store Shedule...21 3.1.5 Store Model Lighting and Equipment Gain Shedule...22 3.2.1 Store Model Refrigerated ase On-Time... 27 3.2.2 Energy Balane on ase-store Air Interfae... 28 3.3.1 Atual and alulated 18-Ton arrier Performanes: Total apaity vs. Entering Wet-Bulb Temperature 1/OP vs. Entering Wet-Bulb Temperature... 34 3.3.2 Atual and alulated 18-Ton arrier Performanes: Total apaity vs. ondenser Dry-Bulb Temperature 1/OP vs. ondenser Dry-Bulb Temperature... 35 3.3.3 Atual and alulated 18-Ton arrier Performanes: Total apaity vs. Flow Rate 1/OP vs. Flow Rate vi
Bypass Fator vs. Flow Rate...36 3.3.1.1 Regression Fit for Saturation Temperature as a Funtion of Saturation Humidity...39 4.1.1.1 Supermarket Loads/redits at Design onditions...44 4.1.2.1 Annual Energy onsumption Breakdown (onventional System Base ase)...46 4.1.2.2 Annual Energy onsumption Breakdown (Humid-limate System Base ase)...47 4.1.3.1 Energy onsumption for onventional and Humid-limate System Base ases...49 4.2.1.1 limate Dependene...51 4.2.2.1 Store Setpoint Dependene (onventional System).53 4.2.2.2 Store Setpoint Dependene (Resized onventional System)...54 4.2.2.3 Store Setpoint Dependene (Resized Humid limate System)... 55 4.2.3.1 Ventilation Flow Rate Dependene... 58 4.2.4.1 irulation Flow Rate Dependene... 60 4.2.5.1 Refrigerated ase apaity Dependene...61 4.2.6.1 Internal Sensible Load Dependene... 63 5.1.1 Hour-of-Day Energy onsumption Breakdown...66 vii
NOMENLATURE BF equip db fm OP d ewb fbffm ft fmed flow fl/op fm fl/op plr fl/op_t Fpw h i if L mi air onditioner oil bypass fator apaity, either refrigeration or air onditioner initial ost of equipment dry-bulb temperature of air entering air onditioner ondenser ubi feet per minute oeffiient of performane disount rate wet-bulb temperature of air entering air onditioner evaporator air onditioner bypass modifying funtion for flow rate air onditioner apaity modifying funtion for flow rate air onditioner apaity modifying funtion for dry-bulb temperature medium temperature refrigerated ase redit modifying funtion low temperature refrigerated ase redit modifying funtion air onditioner 1/OP modifying funtion for flow rate air onditioner 1/OP modifying funtion for partial-load-ratio air onditioner l/op modifying funtion for dry-bulb temperature present worth fator refrigerated ase height general inflation rate fuel inflation rate refrigerated ase linear length mass flow rate 0o0 vii
Ne plr Pr Pf P 1 number of years in eonomi analysis air onditioner partial load ratio total power onsumption of refrigerated ases fan power ratio of life yle savings due to operating ost savings over the first year operating ost savings P 2 ratio of life yle ost due to equipment ost over the equipment ost QI, qs Qsj R T ti tp Sfirst year SL latent refrigerated ase redit anti-sweat heater power sensible refrigerated ase redit thermal resistane dry-bulb temperature effetive inome tax rate property tax rate first year operating ost savings life yle savings latent-to-total refrigerated ase redit ratio AP 7 p 01) stati pressure drop refrigerated ase on-time fration density absolute humidity ix
Subsripts: atual air ase design dp e exit i in low med rated ref surf store indiates atual value pertaining to store air pertaining to refrigerated ase ondenser(s) pertaining to refrigerated ase(s) indiates design value dew point pertaining to refrigerated ase evaporator(s) pertaining to exiting air pertaining to i th omponent pertaining to inoming air pertaining to low temperature refrigerated ases pertaining to medium temperature refrigerated ases indiates value at ARI rating onditions referene value pertaining to air state at surfae of air onditioner evaporator oil pertaining to store air state Supersripts: indiates per area value + indiates value used only if positive, otherwise replaed with 0 x
HAPTER 1: INTRODUTION Approximately 4% of the annual U.S. eletriity onsumption ours in supermarkets (FMI, 1979). The ost of energy in a supermarket an be equal to the store profit (argoaire, 1985, np) and, until reently, the ost of energy was one of the most pressing problems of supermarket owners and managers (Proressive Groer, 1985, 38). For these reasons, this study and others have attempted to develop methods of reduing supermarket energy osts. The emphasis in this study was on reduing supermarket air-onditioning osts. 1.1: Sope of Study Due to the presene of the refrigerator and freezer display ases, supermarket air-onditioning systems present design problems not found with other ommerial building air-onditioning systems. The display ases perform part of the neessary aironditioning, making the remaining ooling load smaller than that of a omparablysized non-supermarket building. The ases remove proportionately more of the sensible load than the latent load, resulting in an air onditioner load with a higher than usual latent-to-total load ratio. Additionally, the display ases require the areful ontrol of store humidity levels. Exess moisture auses inreased refrigeration energy onsumption and the deterioration of the appearane of frozen food pakages (argoaire, 1985, np). Supermarket air onditioning is most ommonly done with onventional vaporompression systems whih proess a mixture of fresh and return air. Vaporompression air-onditioning systems whih dehumidify only the in-oming outside air have been designed for humid-limates ("humid-limate systems"). Hybrid desiant
2 air-onditioning systems are being used in a few stores. These systems use a solid desiant wheel to dehumidify, perform sensible ooling with a vapor-ompression aironditioning unit, and use relaimed heat or gas heat to regenerate the desiant wheel. Finally, ie-storage systems have been proposed for supermarket air-onditioning. While these systems would not redue energy onsumption they would allow energy use to be shifted to off-peak hours. Given an eletriity rate shedule that favors offpeak eletriity use, energy osts ould be redued. This study ompares onventional and humid-limate vapor-ompression supermarket air-onditioning systems for a representative store in Miami, Florida. omputer models of the store and the two systems were reated with the TRNSYS simulation program (Klein et. al., 1983). Using Typial Meteorologial Year (Hall et. al., 1978) weather data for Miami, annual simulations for both systems were ompared. Annual simulations were also made to determine the effets of varying limate, store humidity setpoint, ventilation and irulation flow rates, refrigerated ase apaity and internal sensible load. Sample energy rate shedules were used to ompare the operation osts of both kinds of systems in both Miami and in Madison, WI. A simplified eonomi analysis was then performed to determine how muh higher the first ost of a humid-limate system ould be assuming a paybak period of 2 to 3 years. 1.2: Literature Review At least two studies ompared onventional supermarket air-onditioning systems with hybrid desiant systems, using omputer simulations either in whole or in part. In addition to literature published by manufaturers of humid-limate systems, at least one professional paper explains the benefits of suh systems. In June 1984, the Gas Researh Institute (GRI) published the final report of its
3 two-year omparison of onventional and hybrid desiant systems (GRI, 1984). This study inluded both the experimental omparison of the systems installed in two supermarkets and the analysis of system omputer simulations. The field study onfirmed that redued store humidity levels dereased the refrigeration energy onsumption. It also showed that hybrid desiant systems allow the use of irulation flow-rates lower than those required by onventional vaporompression systems. This an result in a onsiderable savings in fan operating osts. Due to numerous tehnial diffiulties, the field tests did not demonstrate hybrid desiant system operating ost savings in both stores. A number of areas for improvement were suggested, however. omputer modelling in the GRI study showed that hybrid desiant systems were most energy effiient at onventional store humidity levels in spite of redued refrigeration energy onsumption at lower humidities and are most promising in areas of moderate and high humidity. For a modem Miami, FL store with a store humidity setpoint of 0.009 lbmwater/lbmdry air, an annual savings of $2400 to $4900 was predited. Rihard Bums, working with the Solar Energy Lab. of the University of Wisonsin, wrote his M.S. thesis on "An Analysis of Hybrid Desiant ooling Systems in Supermarket Appliations" (Burns, 1985). Burns' work relied exlusively on omputer modelling and was based, in part, on the supermarket and refrigerated ase models developed for the GRI study. He looked at a number of different hybrid desiant system onfigurations and predited a savings in air-onditioning osts of 50-70% (relative to a onventional vapor-ompression system). He also found that, given a hybrid desiant system, the air-onditioning energy onsumption required to maintain lower store humidities exeeds the resulting redution in refrigeration energy
4 onsumption. Bums found the greatest potential savings in regions with extended periods of high humidity, suh as the southeastern states and the Gulf oast. A paper by E.R. Whitehead (Whitehead, 1985) desribes the design and benefits of humid-limate air-onditioning systems. An experimental study detailed in the paper showed that a 39,000 ft 2 store with a humid-limate system used 83% of the energy used by a store of similar size equipped with a onventional air-onditioning system. Both stores were in Texas and the total annual store eletriity onsumptions were ompared.
5 HAPTER 2: SUPERMARKET AIR-ONDITIONING SYSTEMS Two kinds of vapor-ompression supermarket air-onditioning systems are onsidered in this study. This hapter desribes both of these systems, and then desribes the struture of the general vapor-ompression air-onditioning system model used for annual omputer simulations of both types of systems. 2.1: onventional Vapor-ompression System Most supermarket air onditioning is done with onventional vaporompression systems proessing all of the irulation air (a mixture of return and makeup air). Figure 2.1.1 diagrams suh a system. Make-up air enters the system, is mixed with return air, and is ooled and dehumidified by the air-onditioning oil. In order to ontrol store humidity, the proessed air must be ooled below the saturation temperature. Under many onditions, the sensible ooling resulting from adequately dehumidifying exeeds the spae sensible load. To maintain a omfortable store drybulb temperature, the air must be reheated. This is most often done with relaimed heat from the air onditioner or refrigerated ases ondensers. A supplementary heater may also be neessary in some ases if the relaimed heat is not suffiient for both reheating and for winter heating. Proessed air is often supplied at the front of the store, meeting the heavy load in this area due to hekout equipment, higher ustomer/employee density and infiltration through the entrane doors. The reommended plae at whih to remove the return air is underneath the freezer and refrigerated display ases (to be referred to olletively as "refrigerated ases"). This minimizes unomfortably old aisle temperatures near the ases. The resulting return air is ooler and drier than the
EXHAUST AIR STORE LOAD VENTILATION AIRI A B OOLING/DEHUMID. AIR-ONDITIONING OIL RELAIMED HEAT HEAT EXHANGER Fig. 2.1.1: onventional Air-onditioning System
7 store air. Beause ventilation requirements demand that a given amount of outside air be ontinuously added to the system, an equal amount of return air must be exhausted. Air states for a onventional air-onditioning system, given Miami design onditions, are shown on the psyhrometri diagram of Fig. 2.1.2. Air state labels orrespond to position labels in Fig. 2.1.1. Although the enthalpy differene per pound-mass of dry air aross the air onditioner evaporator oil is small (approx. 6.2 Btu/lbm), the large amount of air proessed (18,000 fm or 78,840 lbm/hr) makes the total enthalpy hange large (490,000 Btu/hr). At off-design onditions, it is often neessary to over-ool the air to ontrol humidity. If adequate relaimed heat is not available for reheating the over-ooled air, the system will be more ostly to operate. 2.2: Humid-limate Vapor ompression System A different kind of vapor-ompression air-onditioning system, sometimes known as a humid limate system, is shown in Fig. 2.2.1. The operation of this system is muh like that of the onventional systems exept that the return air is handled differently. Ventilation air only is ooled and dehumidified by a primary air onditioning unit. This proessed air is then mixed with the reirulated return air. Beause it handles only the ventilation air, the humid-limate primary air onditioner has a smaller apaity than the onventional system air-onditioning unit. Many humid-limate systems have seondary air-onditioning units to sensibly ool reirulated return air. This allows the sensible load to be met even when the primary air onditioner an not do all of the neessary sensible ooling. It also allows the sensible and latent loads to be more effiiently met: the primary air onditioner unit provides only enough ooling to meet the latent load while the seondary unit, whih has a higher evaporator temperature and is therefore more effiient, meets the remaining
-0.03 0.02 0.0 I i I I I 1 Io 30 40 50 60 70 80 90 100 OF Fig. 2.1.2: Psyhrometri Diagram for onventional System 00
EXHAUST AIR F OOLING AIR-ONDITIONING OIL STORE LOAD G VENTILATION A B RELAIMED HEAT HEAT EXHANGER Fig. 2.2.1: Humid-limate Air-onditioning System
10 sensible load. If make-up air intake is substantially redued at night, latent loads are high, or installed refrigeration apaity is low, it may be neessary to route some of the return air through the primary air onditioner to provide adequate humidity ontrol. Figure 2.2.2 shows the air states for a humid-limate system at Miami design onditions on a psyhrometri diagram. Again, the air state labels of Fig. 2.2.2 orrespond to the position labels of Fig. 2.2.1. The enthalpy differene per pound of dry air aross the humid-limate system primary evaporator oil is muh larger (approx. 24.2 Btu/lbm) than that for the onventional system. However, the amount of air proessed is onsiderably less (2700 fm or 11,826 lbm/hr) giving a muh smaller overall enthalpy differene (286,000 Btu/hr). The enthalpy differene per pound of dry air aross the seondary oil is approximately 3 Btu/lbm. With a 15,300 fm or 67,014 lbmlhr flow rate, this is a total enthalpy differene of 200,000 Btu. The ombined total enthalpy differene aross the two oils is equal to that of the onventional system. At off-design onditions, when the onventional system must overool and then reheat all of the irulation air, the humid-limate system only overools the ventilation air. The humid-limate system does muh less total ooling and is therefore more effiient. 2.3: General Vapor-ompression System Model One of the primary purposes of this study was to ompare the humid-limate air-onditioning system to the onventional air-onditioning system. It was important not to onfound the effets of the different methods of handling and proessing air with the effets of different air-onditioning equipment. For this reason, a general vaporompression air-onditioning system model was developed (Fig. 2.3.1). By properly setting the diverters, this model an emulate either a onventional system or a
-0.03 oil A -0.02 F 0.01 I I i i i l I I 40 50 60 70 80 90 30 100 of Fig. 2.2.2: Psyhrometri Diagram for Humid-limate System
EXHAUST AIR Fig. 2.3.1: General Vapor-ompression Air-onditioning Model k==
13 humid-limate system. The air-onditioning oils are sized aording to the design onditions and air-flow rates for eah individual system, however the rated oeffiients of performane and by-pass fators have been seleted to be idential for all aironditioning oils (see setion 3.3). In pratie, it is possible that the rated oeffiients of performane (OP's) will not be idential, as the flow rates aross the oils are quite different, and hene the temperature differenes between the entering air and the evaporator may be different. The assumption that the OP's are idential implies that the oils are equally "well-designed" for the system-speifi operating onditions. A number of non-essential features of vapor-ompression air-onditioning systems have not been modelled. Refrigerator fluid sub-ooling has not been modeled as it would improve both the performane of the onventional system and that of the humid-limate system. The option of an eonomizer yle for the onventional system (i.e. taking all irulation air from the outside when profitable) has not been inluded. It is expeted that suh an option's performane would be highly limate-dependent, however onsideration should be given to future testing of this onfiguration. Variations on the humid-limate system have also not been inluded. As an example, some humid-limate-type systems do not have a seondary (sensible ooling-only) aironditioning unit. These systems proess a fration of the return air, in addition to all of the outside air, through the primary air-onditioning oil if proessing only the outside air is not suffiient to meet the load. It is expeted that the performane of suh a system would fall between the performane of the onventional system and that of the humid-limate system as modelled, and that the equipment ost would be lose to that of a onventional system.
14 HAPTER 3: SYSTEM OMPONENTS The implementation of the general vapor-ompression air-onditioning system model desribed in setion 2.3 required the development of models for eah of the system omponents. This hapter desribes these models and their TRNSYS implementations. 3.1: Supermarket Building The supermarket building model used in this study is representative of many large modem supermarkets. It was initially based on the store models used in the GRI (GRI, 1984) and Burns (Burns, 1985) studies, however it was extensively reviewed and revised by supermarket onsultants. While it is important that the store model generate representative sensible and latent loads, it is not neessary, for the purposes of this study, to repliate a partiular store or lass of stores. Fig. 3.1.1 diagrams the energy and mass flows for this building. These quantities are explained below. The supermarket is assumed to be open 24 hours a day, 7 days a week. The sales floor area is 30,000 ft 2 and is onsidered to be square. The onditioned air volume is 480,000 ft 3. The building walls have a U-value of 0.103 Btu/hr-ft 2 -OF while the roof has a U-value of 0.092 Btu/hr-ft 2 -OF. The store front faes south and is assumed to be 30% window. The lighting and equipment load is assumed to be 2 W/ft 2. Internal moisture (from vegetable spraying, floor washing, et.) gain is taken to be 10 lbm/hr, exlusive of latent gain due to people. The irulation flow rate depends on the loads at design onditions but is onstrained to be no less than 15 fmlperson (the ventilation requirement). The normal store setpoint temperature is 75 F and the normal store absolute humidity setpoint is 0.01 lbmwater/lbnmdry air- For simulation
infiltration (sensible and latent) ventilation (sensible and latent) SUPERMARKET Internal Load Soures: people (sensible and latent) lighting and equipment (sensible) food spraying, floor washing, et. (latent) - refrigerated ases (sensible and latent redits) ondution (sensible) solar through window (sensible) exfiltration (sensible and latent) exaust (sensible and latent) U energy and mass flow M energy flow only Fig. 3.1.1: Store Model Energy and Mass Flows
16 runs in whih the humidity level is lowered, the temperature setpoint is adjusted to the equivalent ASHRAE omfort ondition (ASHRAE, 1981, 8.21). Refrigerated ases are divided into two ategories: medium-temperature ases and low-temperature ases. losed-door and open ases are treated identially. The installed refrigerated ase apaities and total linear lengths are: 42 tons medium-temperature apaity with 510 ft.; 15 tons low-temperature apaity with 300 ft. Table 3.1.1 summarizes the above information while Figures 3.1.2-3.1.5 desribe the ventilation, infiltration, number-ofpeople, and lighting and equipment gain shedules. This store model was implemented with a TRNSYS single zone omponent (Type 19) whih uses the ASHRAE transfer funtion approah to model walls, roofs and eilings. In addition to the store desription given above, the use of this omponent required assuming the following values, whih are summarized in Table 3.1.2: The store thermal apaitane is taken to be 7.8x10 5 Btu/OF, based on the estimate of 130 Ibm building material per ft 2 floor spae for heavy onstrution (ASHRAE, 1985, 26.3). The moisture apaitane of the building ontents is aounted for by multiplying the onditioned air volume by 20, giving an effetive air volume of 9,600,000 ft 3. The refletane of the inner building surfaes to solar radiation is assumed to be 0.5 while a value of 0.7 is used as the absorptane of solar radiation by the exterior surfaes. The onvetion oeffiient for all inside surfaes, is taken to be 1.4556 Btu/hr-ft 2 -OF, the value used in determining ASHRAE transfer oeffiients (ASHRAE, 1985, 26.11). The window diffuse solar radiation transmittane value used is 0.8 and the overall solar radiation transmittane is taken to be 0.87. The window U-value (not inluding inside or outside onvetion resistanes) is estimated as 14.16 Btufhr-ft 2 -OF based on 1/2 in. thik glass. The window dimensions are 10 ft x 83 ft, entered horizontally on the store front with the
17 Parameter Vau Hours of operation Peak hours onditioned air volume Sales floor spae 24 hours/day 7 am to 10pm 480,000 ft 3 30,000 ft 2 Building walls U-value 0.103 Btu/hr-ft 2 -F (ASHRAE 8" onrete blok, filled insulation: AO,A1,17,E1,EO) Building roof U-value 0.093 Btu/hr-ft 2 -F (ASHRAE #17, steel sheet with 2" insulation) Storefront window U-value Storefront window area (30% of south wall area) 0.917 Btu/hr-ft 2 -F 831.4 ft 2 irulation flow rate approx. 0.6 fm/ft 2 (based on design onditions with minimum of 15 fm/person) Internal moisture generation 10 lbm/hr Table 3.1.1: Store Model Desription (ontinued on p. 18)
18 Parameter Value Temperature setpoints (extremes) 75 0 F-75.4 0 F Humidity setpoints (extremes) 0.008-0.01 lbmwater/lbmai r (ASHRAE equivalent omfort guide will be used) Installed refrigerated ase apaity Medium temperature Low temperature 42 Tons 15 Tons Refrigerated ases linear length Medium temp. Low temp. 510 ft 300 ft Ventilation flow rate Infiltration flow rate Number of people in store sheduled (see Fig. 3.1.2) sheduled (see Fig. 3.1.3) sheduled (see Fig. 3.1.4) Internal sensible gain due to lighting and equipment sheduled (see Fig. 3.1.5) Table 3.1.1: Store Model Desription (ontinued)
3000 2500 2000 1500 1000 500 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 222324 hour of day Fig. 3.1.2: Store Model Ventilation Flow Rate Shedule
! 1000 900 800 700 600 S 500 400 300 200 100 0 I II -j JL 1I I - --Ir- WAI J.I. LIW JM all. Jl.I t~l~i WW. J i I II U I I I li I I I IlIIIIIII 7r I- I. 1-.1 1 2 34 567 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 hour of day Fig. 3.1.3: Store Model Infiltration Flow Rate Shedule O
180-160_ 140 120 100 6 80 40 20 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 hour of day Fig. 3.1.4: Store Model Number-of-People-in-Store Shedule
2-1.8-1.6-1.4-1.2-1- 0.8-0.6-0.4-0.2- o.! 4. 4-4J 4.1 4. L L4J 4J. 4-4- LI 4. LIP 41 I I I I I I I I I I I I I I I I I I I I L4J 4. Li 1234 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 hour of day w LI. I I-- Fig. 3.1.5: Store Model Lighting and Equipment Gain Shedule
23 Parameter Store apaitane 7.8x10 5 Btu/ F (ASHRAE heavy onstrution) Air volume used for moisture apaitane 9,600,000 ft 3 Walls and roof: Refletane of solar radiation by inner surfaes Absorptane of solar radiation by exterior surfae Inside onvetion oeffiient 0.5 0.7 1.4556 Btu/hr-ft 2 -OF Window: Diffuse solar radiation transmittane Overall solar radiation transmittane U-value (not inluding inside or outside onvetion) Inside onvetion oeffiient Window dimension Window position 0.8 0.87 14.16 Btu/hr-ft 2 -OF 1.4556 Btu/hr-ft 2 -OF 10 ft high, 83 ft long entered horizontally, up 3 ft from floor Surfaes struk by beam radiation floor Table 3.1.2: Additional Store Model Parameters Required by TRNSYS Type 19 (ontinued on p. 24)
24 Parameter People gain: Sensible Latent 315 Btu/hr-person 325 Btu/hr-person Floor U-value 0.0238 Btu/hr-ft 2 -OF ASHRAE 12" onrete slab with 12" insulation (AO, ll, B15, B15, EO) Table 3.1.2: Additional Store Model Parameters Required by TRNSYS Type 19 (ontinued)
25 lower edge 3 ft from the floor. Beam radiation transmitted by the window is assumed to strike the floor. The sensible and latent gains from people are taken to be 315 Btu/hr-person and 325 Btu/hr-person (ASHRAE, 1985, 26.21). The U-value of the floor is 0.02380 Btu/hr-ft 2 -OF. The floor is onsidered to be heavily insulated to simulate the minimal ondutive losses through an atual floor due to the ground temperature below the floor approahing, given adequate time, the store temperature. Sensitivity tests showed that store loads were not signifiantly effeted by hanging the following parameters: the store thermal apaitane, the refletane of solar radiation by inner building surfaes, absorptane of solar radiation by the exterior building surfaes, the window diffuse solar radiation transmittane value, the overall solar radiation transmittane value, the window U-value, and the number of surfaes struk by beam radiation. To avoid unneessary ompliations, only the sales area of the store was modelled. Storage and work rooms, offies, the bakery, et., have been negleted. As the heating and ooling needs of these areas are typially met by separate air-proessing systems, the effet on the analysis of the priniple air-onditioning system of negleting these areas is to inrease the ondution gains through the walls. However, the load due to ondution through the walls during the air-onditioning season was shown to be small ompared to the loads from other soures (see Fig. 4.1.1.1). Store temperature was not allowed to deviate from the temperature setpoint nor was the absolute humidity allowed to exeed the humidity setpoint. The TRNSYS Type 19 omponent apability to alulate infiltration loads was overridden to allow the use of separately alulated infiltration loads following the infiltration shedule of Fig. 3.1.2. It is assumed that 100% of the return air is taken from under the refrigerated ases, and that, as a result, one half of the sensible and latent refrigerated ase redit
26 diretly affets the store air state while the remaining half ats on the return air (argoaire, n.d., 6). It is also assumed that the exfiltration rate is equal to the infiltration rate, making the supply air flow rate equal to the return air flow rate. 3.2: Refrigerated ases The refrigerated ase model used for this study is based on that used by both GRI and Bums. It was modified to inlude anti-sweat heater energy onsumption: the energy required for heaters whih prevent ondensation on metal strips, glass doors, et. The refrigerated ases produe a negative load, or redit, on the surrounding air. The sensible and latent ase redits (Qs, and Ql,) are alulated as follows: Qs, - [med * fmed( ostore) + low * flow((ostore)] * (1 - j3) * (3.2.1) QL, = [med * fmed((ostore) + low * flow((ostore)] * * (3.2.2) where ostre is the absolute humidity of the store air in units of lbmwaterlbmdry air, med and low are medium temperature and low temperature installed refrigerated ase apaities, respetively; the latent-to-total redit ratio [3 is defined below, and y is the fration of time for whih the ases are on as indiated in Fig. 3.2.1. The modifying funtions fmed and flow adjust the ase redit for the absolute humidity level in the store and are defined below. These funtions are based on urve fits to Tyler Refrigeration ompany data (GRI, 1984, 108): fmed(estore) = 7.986 * 0.4 5 2 (3.2.3) 3store)
90 80 70 60 50 40 30 20 10 o. -2 I - t t - t It- I a I I It a I II IF I I ' I I II i i I I 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 hour of day Fig. 3.2.1: Store Model Refrigerated ase On-Time t'o -,,,I
28 flow(ostor e ) =-3.635 * restore 0.2 8 1 (3.2.4) 1-0.016 * exp( 245.0*ostor e) (3.2,5) Anti-sweat heater Power (qs) was alulated using a simple resistane model and representative manufaturer's data. Anti-sweat heaters must raise the temperature of the outside surfae of the ase above the dew point temperature of the store air (Trip). Td... iiili iiiiiiiiii~...... -.-.-... -... -...... ::::::::::::::::::::::::::: :...... :... '.. :.:.. '' Tdp TdstTeas Tstoe qs Rair = + Res e (air6
29 For both low and medium temperature ases it is assumed that Rai = 1 hr-ft 2 -OF/Btu and that the ase height (h) is 5 ft. Tyler refrigerated ase data suggests that 115 V antisweat heaters on medium temperature ases use approximately 0.035 amps per linear foot of ase, or 4 W/ft. If it is assumed that the heaters are ontinuously on when the store air has a temperature of 80OF and a dew point temperature of 70OF and the inside ase surfae temperature is 35 0 F, Rase med = 2.7 hr-ft 2 -OF/Btu. For low temperature ases, 115 V heaters are taken to require 0.2 amps per linear foot of ase or 23 W/ft. With the above store onditions and an inside ase surfae temperature of -10 0 F, Rase,low = 3.1 hr-ft 2 OF/Btu. Using the values for Rase alulated above, the anti-sweat heater energy onsumption is alulated using the following equation: + + qs dp Tstore + Td(p asernmd] L d* hnd + Rair RaseIr~d Tdp- Tstor e +Tdp2 Tase, ow Lo*ho 1t'sor + '~ oj* Li 0 w* h l 0 w (3.2.7) L airaselow(2 where the "+" subsript signifies that the quantity in brakets is used only when positive and is otherwise replaed with 0, and Lined and Llow are the linear ase lengths. The total power onsumption of the refrigerator ases (Pr) is defined by the following equation: ~r q~(3.2.8) (s~i~~)
30 where Qs, and Ql, are found from eqs. 3.2.1 and 3.2.2. Sine the evaporator temperature (Te) remains nearly onstant, the temperature of the air leaving the ase ondenser(s) (whih is approximately the ondenser temperature or T) will determine the OP. Using the arnot OP to approximate this dependeny, *(To - T~desig n OP = OPdesign * (T- TT (3.2.9) Table 3.2.1 indiates the "design" values assumed. (Design OP's are those used in the GRI study (GRI, 1984, 112).) Te T OPdesign medium temp. 0 OF 100OF 1.86 low temp -40 OF 100 OF 1.07 Table 3.2.1: Design Values Assumed for Refrigerated ase OP Determination A new TRNSYS omponent was written to implement the above model. In addition to alulating the total power onsumption, the exiting temperature and absolute humidity of the air passing under the refrigerated ases are alulated. It was assumed that 100% of the irulation air passes under the refrigerated ases before being either exhausted or reproessed and that 50% of the sensible and latent
31 refrigerated ase redits at on this air (argoaire, n.d., 6). (The other 50% of the ase redits is assumed to at on the store air.) It was also assumed that the refrigerated ases ondenser(s) heat is relaimed, and that the ondenser temperature is approximately the temperature of the store supply air when reheat is done and 15 0 F higher than the outside air temperature when reheat is not done. 3.3 Vapor-ompression Air-onditioning Units Both onventional and humid-limate air-onditioning systems use vaporompression air onditioners performing both sensible and latent ooling. A sensibleooling-only air-onditioning unit is sometimes used in humid-limate systems. These units are modelled in a manner similar to that used in DOE-2 (LBL and LANL, 1981, IV.7-IV.16 and IV.62-IV.74). The air state at the surfae of an evaporator oil (Tsurf, (Osurf) an be determined from the states of the air entering (Tin, Oin) and leaving (Texit, exit) the oil: Texit -Tin * BF Tf = 1 -BF (3.3.1) sirf of exit" '0in ("0 F(3 * BF su f = 1 - B.3.2) 1 -BF where BF is the oil by-pass fator (the fration of air that is unaffeted by the oil) alulated for the given onditions. For a oil on whih ondensation ours, Tsurf and 0 ) surf desribe an air state on the saturation line. Beause there is only one absolute humidity for eah temperature on the moist-air saturation line, the amounts of
32 sensible and latent ooling performed by the oil are oupled. This model requires the speifiation of oil performane at ARI rating onditions (67OF wet-bulb entering evaporator temperature, 95 0 F entering ondenser dry-bulb temperature, and a flow-rate equal to or less than 37.5 fm/1000 Btu/hr ooling apaity (ARI, 1966)). oil performane at off-rated onditions is determined by multiplying the various rated performane indiators by one or more modifying funtions: BF =BFrat d * fbf d4fmatm (3.3.3) rate d * f 1 ewbdb) * f (fifatua) (3.3.4) = (1*ed) f 1 /op jewb,db) * flopfr4fna=t)* O--i = Ora t fl/xop-pp,rh) (3.3.5) where ewb is evaporator entering wet-bulb temperature (OF), db is ondenser entering dry bulb temperature (OF), fmatual is the atual volume flow-rate (in fm), is the total apaity and plr is the part-load-ratio or (total load / ). The effet on the air onditioner OP of relaiming heat from the ondenser was not inluded in this study. Future studies should onsider modelling this effet. DOE-2 default funtions have been used for the modifying funtions (LASL, 1980, 72-73): fmaetml fbf--.~fm/atal) = fmrate d (3.3.6)
33 ft(ewb,db) = 0.418934 + 0.017421 * ewb - 0.00617 *db (3.3.7) f~f4firlatmai = 0. 6 9 7 17 19 9 + 0. 3 9 5 5 5 * (fma+tual + fllrated J 0.092727 * ( f - 2 (a3t3 ( fmrated ) (33.8) fl/opt(ewb,db) = 0.282094-0.005832* ewb + 0.01167 * db (3.3.9) f/p_.fmatu = 1.13318-0.13318* I n at I a (3.3.10) finrat d ]( fi/op.pir(plr) = 0.29333333 +0.70666667 * plr (3.3.11) It has been assumed that partial load apaity is ahieved exlusively by ylinder unloading (i.e., hot-gas bypassing and ompressor yling are not used). These default funtions were used to predit manufaturer's published performane data (arrier, 1986, 16). Predited and published results agree well (Figs. 3.3.1-3.3.3). It is expeted that the performane modifying funtions will have slightly different shapes for different air-onditioning units, however for the purposes of this study, a representative, not a speifi, air-onditioning unit is required.
34 250,000 200,000. 150,000 100,000 aity 50,000 " -..... 1 alulated 0 I 62 67 entering wet-bulb temperature (F) 0 qrptiivi r1 f %wu '" apaity Fig. 3.3.1 (a): Total apaity vs. Entering Wet-Bulb Temperature 72 0.4 0 0.3 Ask 0.2" 0.1 O- 1/OP atual -0 1/OP alulated 62 67 entering wet-bulb temperature (F) I 72 I Fig. 3.3.1 (b): 1/OP vs. Entering Wet-Bulb Temperature Fig. 3.3.1: Atual and alulated 18-Ton arrier Performanes (ond. dry-bulb: 95 F; flow rate: 8100 fm)
35 250,000 Amok 200,000- I.. 150,000 0 100,000-0 atual apaity * alulated apaity OL I I I 85 95 105 ondenser dry-bulb temperature (F) Fig. 3.3.2 (a): Total apaity vs. ondenser Dry-Bulb Temperature 115 0.4 0.4 0.3 0.21 0 50,000-0.1-0 1/OP atual * 1/OP alulated 04 I I i 85 95 105 115 ondenser dry-bulb temperature (F) Fig. 3.3.2 (b): 1/OP vs. ondenser Dry-Bulb Temperature Fig. 3.3.2: Atual and alulated 18-Ton arrier Performanes (entering wet-bulb: 67 F, flow rate: 8100 fm)
36 250,000 I 200,000; 150,000 100,000 50,000 > atual apaity * alulated apaity A' 5400 7200 fm 9000 Fig. 3.3.3 (a): Total apaity vs. Flow Rate 0.4 0.3 0.21 0.1t O 1/OP atual W l/op alulated AIa I I Ii10- V 1 5400 7200 fem Fig. 3.3.1 (b): 1/OP vs. Entering Wet-Bulb Temperature 9000 Fig. 3.3.3: Atual and alulated 18-Ton arrier Performanes (entering wet-bulb: 67 F; ondenser dry-bulb: 95 F) (ontinued on pg. 37)
37 0.07 0.06-- 0.05 0.04 S0.03.0.02 0atal apaity 0.01 0.0. alulated apaity 5400 7200 9000 Fm Fig. 3.3.3 (): Bypass Fator vs.fm Fig. 3.3.3: Atual and alulated 18-Ton arrier Performanes (entering wet-bulb: 67 F; ondenser dry-bulb: 95 F) (ontinued)
38 3.3.1: Sensible and Latent ooline Unit A new TRNSYS omponent was used to implement the vapor-ompression air-onditioner model desribed above. To avoid multiple system iterations for eah time-step of a simulation, the ontrol of the air onditioner was inluded in the omponent. For onventional systems the sensible load is first met. If this does not also meet the latent load, the latent load is met, sensibly over-ooling the air. For a humid-limate system, the latent load is met and it is assumed that the seondary aironditioner will meet any remaining sensible load. The TRNSYS psyhometeris routine was used to determine osurf given Tsurf, while a regression fit to moist air states on the saturation line was used to determine Tsurf given Osurf (see Fig. 3.3.1.1). Tsurf is onstrained to be no less than 33 0 F. The rated apaity was hosen to just meet the load at design onditions, the rated OP for all units was assumed to be 4 and the rated bypass fator for all units was assumed to be 0.05. Assuming that the temperature and humidity remain at or below the setpoints implies that the sensible and latent loads be met at all times. For this reason, for the few hours of the year that the load exeeds the apaity, the partialload-ratio is set to 1 and the load is onsidered to be fully met. 3.3.2 Sensible ooling Only Unit A TRNSYS omponent similar to that desribed above was reated to model the seondary (sensible-only) unit of the humid-limate system. Only the ontrol of the unit differs: as this oil is not expeted to do any latent ooling, Tsurf is first determined. If the orresponding osuf~ implies that ondensation ours, the resulting exiting absolute humidity will be alulated, otherwise the inoming and the exiting absolute humidities will be idential. Texit will be alulated in both ases. One again, the rated apaity was hosen to just meet the load at design onditions, the rated
75 70 0 85 - [i data points "0 - regression fit a 55 '.4, 50 0.- 45- '4 40' " Tsat : = 6.8123 + 1.1679 * (Wsat * 7000) - ~ 00.007985 * (wsat * 7000) 2 + 0.000023332 * (wsat "' 7000)3 30 -" 0.004 0.008 0.008 0.010 0.012 O.O14 O.0016 0.018 e.020 absolute humidity (lbmwater/lbmdry air) Fig. 3.3.1.1: Regression Fit for Saturation Temperature as a Funtion of Saturation Humidity
40 OP for all units was assumed to be 4 and the rated bypass fator for all units was assumed to be 0.05. If the load exeeds the apaity, the partial-load-ratio is set to 1 and the load is onsidered to be fully met. 3.4: Reheat Heat Exhanger To ontrol humidity, the primary air-onditioning unit in both the onventional and humid-limate systems often over-ools the air. This air must be reheated before returning it to the store. A typial 30,000 ft 2 store an produe 600,000-900,000 Btu/hr of relaimed heat from the air-onditioner ondenser (MQuay, 1979, 4). Heat may also be relaimed from the refrigerated ases ondenser(s). As reheat/winter heating requirements larger than 700,000 Btu/hr were not enountered in any Miami simulation, all heating was onsidered "free". A TRNSYS algebrai unit (Type 15) was used to alulate required heat and exiting temperature. 3.5: Fan Power The air flow rates through some omponents of the humid-limate system are onsiderably smaller than those in the onventional system. If these two systems are to be ompared, fan power ost must be inluded. Fan power is a funtion of volume flow-rate (mn/p) and stati pressure drop (APi) for eah omponent i of the system as given by: r i H 2 ] Pf * APi * (3.5.1)
41 where mairf is the flow rate at whih the pressure drop for the ith omponent was determine, rn is the atual flow rate through omponent i, and p is the density of the dry air. Fan ineffiienies whih heat the air have been onsidered negligible. Table 3.5.1 lists the assumed omponent pressure drops and referene flow rates. The dut pressure drops are those used by Burns (Burns, 1985, 54). Neither the onventional system nor the humid-limate system is a variable volume system. As a result, there are two onstant sets of flow rates for eah system: one for the peak hours period and one for the off-peak hours when the ventilation is ut bak. Fan power requirements were hand-alulated and then added to the omputeralulated power requirements for other omponents. 3.6: Misellaneous omponents Several standard utility TRNSYS omponents were used in system simulations. Solar radiation data, outside temperature and humidity, and windspeed were read from a Typial Meteorologial Year (TMY) data set with a Type 9 data reader. The radiation data from this omponent was proessed by a Type 16 radiation proessor to generate the solar energy reeived by the store walls, roof and window. Several units of the Type 14 time-dependent foring funtion omponent were used to generate hourly values for sheduled quantities suh as infiltration and ventilation rates. Finally, slightly improved Type 11 flow-diverter/mixer omponents were used to ontrol airflow. (See Appendix B.)
42 omponent APi(in. H20) mi,ref (ibm/br) duts 2.0 78,840 reheat oil: onventional (2 rows) 0.30 78,840 humid-limate (1 row) 0.24 78,840 onventional a.. oil (4-6 rows) 1.4 78,840 primary humid-limate a.. oil ( 8 rows) 1.7 11,830 seondary humid-limate a.. oil 1.2 67,010 ( 4 rows/sensible) indiret evaporative ooler 0.5 78,840 outside air filter 1.0 11,830 Table 3.5.1: Pressure Drops Aross Air-onditioning System omponents
43 HAPTER 4: SYSTEM SIMULATIONS The omponent models desribed in hapter 3 were ombined to form models of the onventional and humid-limate systems disussed in hapter 2. This hapter reports and analyzes the results of the annual simulations of these models. Unless otherwise speified, all simulations were performed using Miami, FL TMY weather data. 4.1: Base ases Design loads for the model supermarket were alulated for Miami. The irulation flow rate and air onditioner apaity were sized to meet these loads, and annual simulations of these "base ases" were run. 4. 1.1: Desien Loads and System Sizing Both the onventional and the humid-limate systems were sized using design loads alulated by the Trane ompany's Load Design program (Trane, 1986). Fig. 4.1.1.1 shows the supermarket loads and redits at design onditions. The lighting load and the load due to ondution through the roof are the major omponents of the sensible load. The total latent load is dominated by the people-generated load and ventilation latent loads. For design purposes, the refrigerated ase on-time was taken to have its lowest value (see Fig. 3.2.1). Even at this level, both the sensible and latent refrigerated ase redits are substantial. Given the internal (i.e. exluding ventilation) design loads of 381,600 Btu/hr sensible and 77,300 Btu/hr latent, the irulation flow rate was hosen to be 18,000 fm, the lowest round number flow rate whih gives a physially ahievable supply
ventilation (sensible) lights (sensible) infiltration (seisible) solar from window (sensible) ondution/roof (sensible) ondution/walls (sensible) ases (sensible) people (latent) ventilation (latent) internal (latent) infiltration (latent) ases (latent) -150000-100000 -50000 0 50000 100000 150000 200000 250000 Btu/hr Fig. 4.1.1.1: Supermarket Loads/redits at Design onditions
45 air state. The rated air onditioner apaities of 495,000 Btu/hr (41.3 Tons) for the onventional system and 235,000 Btu/hr (19.6 Tons) and 215,000 Btu/hr (17.9 Tons) for the humid-limate system primary and seondary air onditioners, respetively, were hosen to just meet the design loads. The onventional and humid-limate systems onfigured with these values formed the base ases. 4.1.2: Enere onsumption Breakdown Only those supermarket energy onsumptions that vary with air onditioner system type were onsidered in this study. Into this ategory fall the eletriity used by the air onditioner unit(s), the refrigerated ases and the fans, and the energy required by the reheat oil. Fig. 4.1.2.1 harts these quantities for eah month of the onventional system base ase simulation while Fig. 4.1.2.2 displays the same quantities for the humid-limate base ase simulation. The reheat energy is onsidered to be "free" in this study (see setion 3.4). Of the energy quantities whih must be purhased, the refrigerated ase ompressor energy onsumption is the greatest: between 5 to 230 times the air onditioner energy onsumption, depending on the month and on the system. For both types of systems, the refrigerated ase anti-sweat heater energy onsumption is very small (~2%) ompared to the ompressor energy onsumption This suggests that the Bums and GRI studies were justified in negleting the anti-sweat heaters. For the onventional system, the energy neessary for the refrigerated ases ompressor stays relatively onstant throughout the year as the result of two different effets: lower store humidities during the winter tend to derease this energy onsumption, while a higher demand for heat/reheat energy during the summer (and hene a higher refrigerated ase OP) also tends to derease this energy onsumption. The humid-limate system does not require as muh heat/reheat energy
2500-r 2000-f 1500! reheat "0 or-_ ref. ases: ompressor - G o 0 1000-500- air onditioner fans ref. ases: anti-sweat heaters 0 Jan. Feb. Marh April May June July Aug. Sept. Ot. Nov. De. month Fig. 4.1.2. 1: Annual Energy onsumption Breakdown (onventional System Base ase)
1800 ref. 1600 1400 1200 1000. reheat fan 0 m Jan. Feb. Marh April May June July Aug. Sept. Ot. Nov. De. month Fig. 4.1.2.2: Annual Energy onsumption Breakdown (Humid-limate System Base ase)
48 during the summer as it does not over-ool the proessed air to the same degree. The refrigerated ase ondenser is onsequently more often ooled by the warmer outside air, reduing the ase OP and inreasing the refrigerated ase energy onsumption. 4.1.3: onventional System vs. Humid-limate System Fig. 4.1.3.1 ompares the air-onditioner, refrigeration and fan power requirements for the two systems The substitution of a humid-limate system for a onventional system resulted in a 63% annual air onditioner energy savings. The differenes in the air handling of the two systems produes a 5% annual fan power savings for the humid-limate system. The humid-limate refrigeration energy onsumption is 5% higher than that of the onventional system as a result of the effet of relaimed reheat on the refrigerated ase OP. The total (i.e. air-onditioner units(s), fans and refrigerated ases) annual energy savings is 9,700 kwh, or 1%. If the hange in the refrigeration energy onsumption for the humid-limate system is ignored, the annual savings would be 37,100 kwh or 5%. learly the effet on the refrigerated ase OP of relaiming heat from the refrigerated ase ondenser is important. 4.2: Sensitivity Analyses A number of the base-ase model parameters were varied to study what effet, if any, different store onfigurations have on the potential energy savings of a humidlimate system. 42,1: limate omparing annual simulation results done with TMY weather data for a humid limate suh as Miami with those for a older, drier limate suh as Madison, WI,
2500 2000 - refi 1500 AM.. ref rig/onv 1000 500 fans 0 IMM Jan. Feb. Marh April May June July month Fig. 4.1.3.1: Energy onsumption for onventional and Humid-limate System Base ases Aug. Sept. Ot. Nov. De.
50 reveals the aptness of the name "humid-limate system". Fig. 4.2.1.1 is a bar graph showing the annual energy required for air onditioning, refrigeration and reheat/heating by both onventional and humid limate systems in Miami and in Madison. With the exeption of the weather data, the simulations for both loations are idential. As stated in setion 4.1.3, the humid limate system uses 63% less aironditioner energy in Miami than the onventional system. The perentage savings in Madison is similar: 65%. However, the limate in Madison requires only 5% of the air-onditioning energy neessary in Miami. learly the potential savings are muh greater in a humid limate. The refrigerated ases require 9% less energy in Madison in Miami. This is beause in Madison the average store humidity is lower due to the drier limate and the average outside temperature is lower. The inrease in refrigeration energy noted for the humid-limate system in Miami does not our in Madison, as the outside air (i.e., the air used to ool the refrigerated ase ondenser(s) when reheat is not done) is generally ooler than the store supply air. Finally, as expeted, the reheat/heating requirements in Madison are greater. In Miami, the humid-limate system requires notieably less reheat energy, whereas there is little differene between the onventional and humid-limate systems in Madison. This is due to the small air-onditioning requirements and large heating requirements in Madison. The heat/reheat requirements in Madison exeeded 1,000,000 Btu/hr during some portions of the year. It is unlikely relaimed heat ould meet this demand. 4.2,2: Humidity Setpoint It is well aepted that lower store humidity levels result in lower refr'igeration energy requirements (Tyler, 1978). As the refrigeration energy onsumption is muh
900,000-800,000-700,000-600,000-500,000-400,000-300,000-200,000-100,000-0- (2733) E im (947) M lim onventional/madison hum. lim./madison onventional/miami hum. lim./miami air ond. refrig. reheat/hea Fig. 4.2.1.1: limate Dependene
52 greater than that of either the fans or the air-onditioning units, it might be possible for a lower store humidity setpoint to result in net energy onservation. Fig. 4.2.2.1 shows the effet of lower store humidity setpoints on the total energy requirements of a onventional air-onditioning system. Only the store setpoints were hanged between simulations. The temperature setpoint was inreased as the humidity setpoint was dereased to give approximately the same omfort ondition (ASHRAE, 1981, 8.21). Although the refrigeration energy onsumption does derease, this is more than ompensated for by the inreased air onditioner energy use. The net effet is greater annual energy onsumption for lower humidity setpoints. A similar omparison was attempted for the humid-limate system, however lower store humidity levels ould not be maintained with the base-ase ventilation levels (see Fig. 3.1.2) as the ventilation air ould not be ooled and dried suffiiently to meet the load. To determine what would happen if the air-onditioning systems were sized for lower store humidity levels, design loads were alulated for the lower humidity setpoints and the irulation and air-onditioner apaities were realulated. Fig. 4.2.2.2 plots the air onditioner, refrigeration, fan and total annual energy onsumptions for resized onventional systems. Lower humidity setpoints do save energy, however this is the result of lower irulation flow rates (16,000 fm for a humidity level of 0.009 lbmwater/lbmdry air, and 13,000 fm for a humidity level of 0.008 lbmwater/lbmdry air). The lower humidity supply air an be ooler, allowing the sensible and latent load to be met with a lower irulation flow rate. Fig. 4.2.2.3 shows the energy requirements for resized humid-limate systems. The lower humidity setpoints allow the same lower irulation flow rates found for the resized onventional systems, however the ventilation rates must be inreased to provide adequate air to dehumidify. In the ase of the system sized for a humidity
800,000 700,000 total.0 600,000 500,000 '2) 400,000 300,000 200,000 100,000 air onditioner fan 75/0.01 75.2/0.009 Store Setpoint (OF / lbmwater/lbmdry air) 75.4/0.008 Fig. 4.2.2.1: Store Setpoint Dependene (onventional System)
800,000 T total 700,000 600,000 - _ refrigeration 500,000 400,000 300,000 200,000 100,000.-- aironditioner fan 75/0.01 75.2/0.009 75.4/0.008 Store Setpoint (OF 1 lbmwaterbmdr airy Fig. 4.2.2.2: Store Setpoint Dependene (Resized onventional System)
800,000 T 700,000 total 600,000-- refrigeration 500,000.. U 200,000 fan air onditioner 400,000-300,000-100,000-04- 75/0.01 75.2/0.009 75.4/0.1008 Store Setpoint (OF / lbmwaterbmdr airy Fig. 4.2.2.3: Store Setpoint Dependene (Resized Humid-limate System) ta.a
56 level of 0.009 lbmwater/lbndry air, trial and error showed that the off-peak ventilation rate needed to be 2400 fm instead of 1350 fm. For the system with a 0.008 lbmwater/lbmdry air setpoint, the same trial and error proess was used to determine that the peak ventilation flow rate must be 5300 fm with an off-peak rate of 4400 fm. Higher flow rates require more energy, making the total energy onsumption vs. humidity setpoint relationship more ompliated than that for the onventional system. The system sized for a 0.009 lbmwateribmdry air setpoint uses 5% less total energy annually while the system sized for a 0.008 lbmwater/bmdry air setpoint uses 4% more. It should be noted that the seondary air onditioner was used only a few hours of the year in the lower setpoint humidity humid-limate systems. In an atual installation, reproessing a small amount of return air through the primary air onditioner along with the minimum amount of ventilation air and eliminating the seondary air-onditioner would be more energy effiient than inreasing the ventilation flow rate. For the onventional system, if a lower setpoint humidity is to be used, the irulation flow rate should be reset. This would also hold true for humid limate systems modified as desribed above. onsideration should be given to the ost of the inreased air onditioner apaities neessary to maintain lower humidity levels. The inreased apaities are shown in Table 4.2.2.1. 4.2.3: Ventilation Flow Rate Annual simulations were run for varying ventilation flow rates. Figure 4.2.3.1 displays these results, showing the energy requirements as a funtion of the peak hours ventilation flow rate. In all ases, the off-peak hours ventilation rate was hosen to be
57 Humidity Setpoint bmwater/lbmdry air) 1 0009 0008 Air onditioner Type onventional 495,000* 510,000 555,000 Humid limate System: Primary 235,000 250,000 475,000 Seondary 215,000 200,000 135,000 *all apaities in Btu/hr Table 4.2.2.1: Air onditioning Unit apaities for Humidity Setpoint Tests
800,000 ]" tota!/onv -N 700,000 tota! '- *.*' 600,000 I refrigeration/h- A refrigeration/onv t 500,000 400,000 300,000 200,000 100,000.air onditioner/onvy I,- air onditioner/h- 13 -I I I I 0 1 1 29 00 3000 3300 3600 Ventilation Flow Rate (fm) Fig. 4.2.3.1: Ventilation Flow Rate Dependene 00
59 one-half that of the peak rate. The fan power in both types of systems inreased, as did the air onditioner energy requirements. The refrigeration energy requirements also inreased as the result of inreased annual average store humidity. For a 33% inrease in ventilation flow rate, the total energy onsumption inreased 3% for the onventional system and 2% for the humid-limate system. The ventilation flow rate does not appear to be a ritial fator, although the minimum should be used for maximum energy effiieny with either system. 4.2.4: irulation Flow Rate Inreasing irulation flow rate has little effet on air onditioner or refrigeration energy onsumption for either type of system, however fan power requirements are notieably inreased. These results are shown in Fig. 4.2.4.1. For a 44% inrease in irulation flow rate, the annual energy onsumption inreased 22% for the onventional system and 19% for the humid-limate system. learly the irulation flow rate should be set as low as possible, regardless of the system used. 4.2.5: Refrigerated ase apaity The latent and sensible redit due to the refrigerated ases aount for a substantial portion of the total ooling performed on the supermarket air. The refrigerated ase apaity was varied to study the resulting effet on the energy onsumption (see Fig. 4.2.5.1). As expeted, the refrigeration energy onsumption inreases as refrigeration apaity inreases. The air-onditioning energy onsumption dereases for both systems, with the differene between the two systems also dereasing. This ours beause the portion of the supermarket load not met by the refrigerated ases dereases, making the differenes between the two systems less
900,000-800,000- a. tota]h- 700,000 1 600,000 500,000 A L refrigeration/h- S0 0 refrigeration/ony 400,000 300,000 200,000.- _- 100,000 air onditioner/ony air onditioner/h--.. U I I I I * I I I d 18,000 19,000 20000 21,000 22,000 23,000 24,000 25,000 26,000 irulation flow rate (fm) Fig. 4.2.4.1: irulation Flow Rate Dependene
100,0000 900,000 800,000 700,000 600,000 500,000 400,000 7) 300,000 200,000 100,000 0 -I. 0.7 *air ond/onv...air ond/h- MIN,r I i i I a 0.8 0.9 1 1.1 1.2 multiple of base ase refrigerated ase apaity 1 1.3 1 1.4 Fig. 4.2.5.1: Refrigerated ase apaity Dependene
62 important. This implies that stores with very high refrigeration apaities will benefit less from humid-limate systems. Given a refrigerated ase apaity 0.5 times the base ase apaity, it was not possible to ool and dry the ventilation air suffiiently in the humid-limate system to meet the store loads. This suggests that humid-limate systems may not be feasible for stores with low refrigerated ase apaities. 4.2.6: Internal Sensible Load The internal sensible load was varied by hanging the lighting and equipment load. This variation also aused the sensible-to-latent-load ratio to hange. The results of annual simulations using 0.5, 1 and 1.5 times the standard lighting load are displayed in Fig. 4.2.6.1. The air onditioner energy requirements inrease only slightly, as the primary air onditioner is often already over-ooling in order to meet the latent load. The refrigeration energy requirements generally inrease very slightly (1-2%) with inreasing sensible load as less reheat is done (and hene the refrigerated ase OP dereases). At 1.5 times the base ase internal sensible load, the inrease in the humid-limate system refrigeration requirements is suffiiently larger than that for the onventional system to eliminate the energy onsumption differene between the systems.
800,000 700,000 600,000 - total/onv-, " totalh-- refrigeration/h- -. refrigeration/onv - 500,000 400,000 300,000 200,000 100,000 0 0 0.5 air onditioner/onv air onditioner/h-- 1 multiple of base ase intemal sensible load 1.5 Fig. 4.2.6.1: Internal Sensible Load Dependene
64 HAPTER 5: EONOMI ONSIDERATIONS This hapter onsiders two eonomi aspets of supermarket air-onditioning systems: the potential for operating ost savings, and the allowable inrease in first osts for an improved system given a speifi paybak period. 5.1: Annual Operating ost Savings In Miami, substituting a humid-limate vapor-ompression air-onditioning system for a onventional air-onditioning system resulted in a savings of 9,700 kwh (see setion 4.1.3). The peak demand (alulated hourly) was redued by 2.7 kw (121.8 kw to 119.1 kw) or 2%. Assuming a low flat rate eletriity shedule (see Table 5.1.1), this represents a $600 annual operating ost savings. A high flat rate shedule (again see Table 5.1.1) gives a $1,200 savings. Energy_ harge ($/kwh) Demand harge ($/kw) Low rate 0.04 5.50 High rate 0.076 13.00 Table 5.1.1: Flat Rate Eletriity Rate Shedules (Blatt, 1987) The GRI study reported a total energy savings of $2,400 to $4,900 for a hybrid-desiant system in Miami (see setion 1.2) using an eletriity prie of $0.0446/kWh with a demand harge of $6.25/kW and a gas rate of $0.00483/ft 3. This
65 rate shedule gives an annual savings of $600 for the base ase omparison of the urrent study. This suggests that a hybrid-desiant system offers a greater potential for energy savings than a humid-limate vapor-ompression air-onditioning system, however it is not lear if the GRI study aounted for the effet on the refrigerated ase OP of relaimig heat from the ondenser. Fig. 5.1.1 shows the total energy onsumption breakdown by hour of day for both base ase systems in Miami. (As an example, the annual sums of the refrigeration, air-onditioning and fan energy onsumed eah day between midnight and 1 am is represented by the first pair of bars.) As there is very little differene in the energy use profiles for the systems, a time-of-use eletriity shedule will not benefit one system to a greater extent than it will the other. Assuming on-peak hours from 8 am to 8 pm, rate shedule A of Table 5.1.2 gives an annual energy ost savings of $900 while rate shedule B gives a savings of $700. Energy harge ($/kwh) Demand harge ($/kw) Shedule A: on-peak 0.065 8.00 off-peak 0.05 4.00 Shedule B: on-peak 0.08 9.00 off-peak 0.08 0.00 Table 5.1.2: Time-of-Use Eletriity Rate Shedules (Blatt, 1987)
45000 40,000 35,000 30,000 25,000 20,000 15,000 10,000 5,000 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 hour of day Fig. 5.1.1: Hour-of-Day Energy onsumption Breakdown