1、202 2016 ASHRAEABSTRACTAs air-conditioning demand has increased significantlyduring the last decade, efficient energy use has become moreimportant due to large electric power demands and limitedreserves of fossil fuel. Electrical energy use fluctuates signifi-cantly during a 24-hour day due to varia
2、ble demand fromindustrial, commercial, and residential activities. In hot andcold climates, the dominant part of the load fluctuation iscaused by cooling and heating demands, respectively. If elec-tric loads could be shifted from peak hours to off-peak hours,not only would building operation costs d
3、ecrease but the needto run peaker plants, which typically use more fossil fuels thannon-peaker plants, would also decrease. Thus, shifting elec-tricity consumption from peak to off-peak hours promoteseconomic and environmental savings. This paper uses simu-lation and experimental work to examine 12
4、precooling strat-egies in three residential buildings in the Phoenix, Arizona,climate.Theselectedbuildingsareconsideredtorepresentthemajority of residential buildings in the area. Results of thisprojectshowthatprecoolingcansaveupto46%ofpeakenergydemand in a home constructed with concrete or cementit
5、iousblock and up to 35% in wood frame homes. Homeowners cansave up to U.S. $244/year in block construction and up toU.S. $119/year in wood frame homes.INTRODUCTIONMotives and ObjectivesBased on the Residential Energy Consumption Survey(RECS), the number of homes equipped with mechanical air-conditio
6、ning systems has increased in the U.S. from 68% in1993 to 87% in 2009 (EIA 2011). In Arizonaa hot, aridregion in climate zone 2Bcooling equipment consumesnearly40%oftheelectricityusedinhomes(EIA2009).More-over, the air-conditioning demand fluctuates significantlyduring a 24-hour day, which creates a
7、 challenge for electricalutilities. Many utility companies have introduced differenttariffprogramstomotivatecustomerstodecreasetheirenergyuse during peak hours (Herter and Wayland 2010). If electricloads could be shifted from peak hours to off-peak hours, notonly would building operation costs decre
8、ase but the need torun peaker plants, which typically use more fossil fuels thannon-peaker plants, would also decrease. Therefore, shiftingelectricityconsumptionfrompeaktooff-peakhourspromoteseconomicandenvironmentalsavings.Moreover,ifthepeakisreliably flattened or shifted, this may limit the need t
9、o buildexpensive new generating capacity.Solar shading, adoption of solar photovoltaics (PVs), andload shedding (reducing total electricity use) are examples ofother mechanisms of peak energy reduction (Turner et al.2015).Peakloadreductionhasalsobeenresearchedandtestedby improving insulation used wi
10、thin a wall (Al-Sanea andZedan2011)andbyaddingthermalmasstothebuildingenve-lope(Al-Saneaetal.2012;Burchetal.1982).Springer(2007)documents precooling as the most economical operationalstrategy for residential load shifting. This paper explores 12precooling strategies in the Phoenix, Arizona, climate.
11、Precooling aims to reduce the evening peak load of residentialbuildings by implementing optimal thermostat setpoints thatshiftpartoftheon-peakloadtooff-peakhours.Thisstudyalsoaims to assess energy and cost savings associated with the 12strategies modeled.Modeling and Testing Multiple PrecoolingStrat
12、egies in Three Residential BuildingTypes in the Phoenix ClimateReza Arababadi Kristen Parrish, PhDStudent Member ASHRAEReza Arababadi is a doctoral student and Kristen Parrish is an assistant professor at the School of Sustainable Engineering and the BuiltEnvironment, Arizona State University, Tempe
13、, AZ.ST-16-020Published in ASHRAE Transactions, Volume 122, Part 2 ASHRAE Transactions 203Other Precooling ResearchMore than three decades ago, the concept of precooling abuilding was published by Hartman (1980), who stated that abuilding can be cooled by free cooling during night and morn-ing hours
14、 and that cooling energy can be stored in the build-ings mass for release during warm hours. The stored energyin the building is limited by the capacitance of the building(i.e., the capacity of the building to store thermal energy or thebuildingsthermalmass)andbytheminimumindoortempera-ture at the o
15、nset of occupancy of the building, which deter-mines how cool the building can be before peak hours whilemaintaining occupant comfort. Results of precooling studies(e.g., Springer 2007; Herter 2012; Yin et al. 2010a, 2010b;Turner et al. 2015) show a reduction of electric energy cost byshiftingpartof
16、thedailycoolingloadstooff-peakhours,whenelectricity is cheaper. These studies also point out that mostprecooling strategies optimize for expected cooling loadrather than for cost savings. That is, consumers select aprecooling strategy that will promote comfort during the peakhours without necessaril
17、y considering energy cost savings(note many studies on precooling to date are part of a demandresponseprogram,wheredemandshedisrequired,whichmayexplain the secondary focus on cost). Under hot night condi-tions (e.g., in the Phoenix climate), when free cooling is notalways feasible, mechanical coolin
18、g is required, but this cool-ing is less expensive than on-peak cooling due to demandcharges. Precooling is an operational shift, rather than a tech-nological one, and is thus widely accessible to utilitiescustomer bases.The literature presents very limited studies of precoolingin residential buildi
19、ngs, and most of the previous precoolingstudies focus on commercial buildings, likely due to the factthat ventilation control is more common in these buildingtypes. Braun (2003) presented a review of research related totheuseofbuildingthermalmassforshiftingandreducingpeakcooling loads in commercial
20、buildings and provided specificresults obtained through simulations, laboratory tests, andfieldstudies.Morrisetal.(1994)studiedtwooptimaldynamicbuilding control strategies in a representative room in a largeoffice building. They observed a reduction of 40% in peakcooling load. Yin et al. (2010b) dev
21、eloped a methodology tooptimizeprecoolingstrategiesforbuildingsinahotCaliforniaclimatezonewithabuildingenergysimulationtool.Resultsoftheir work indicate that the optimal demand response strate-gies worked well for most of the commercial buildings testedin this hot climate zone. Keeney and Braun (199
22、7) developedand tested a cooling control strategy for a large office buildingnear Chicago. Their results showed reduction of cooling loadto 75% of cooling system capacity. Turner et al. (2015)focused on wood frame residential buildings and used energymodeling to evaluate the effectiveness of residen
23、tial precool-ing to reduce the on-peak energy demand. Their resultsshowed the best precooling results for most climates wereobtained using a medium (5 h) precooling period with a shal-low precooling setpoint temperature. Cole et al. (2014) usedan extensive data set including home energy audits, home
24、-owner interviews, and electricity use measurements to build asimulatedcommunityof900homes.Themodelwasthenusedto investigate the potential for coordinated control of a largenumber of residential air-conditioning systems.Booten and Tabares-Velasco (2012) used EnergyPlus tomodel cooling energy demand
25、during summer days in anactual home in Sacramento, California. Their study assessedthree cooling strategies, and their aims were to evaluate thecapability of EnergyPlus to accurately model cooling energyuse in a house by comparing simulations to empirical data andto investigate the potential impacts
26、 of these cooling strategiesover an entire cooling season. Their simulation results showthat EnergyPlus can capture the important features, such aswhen the cooling system was operating, and predicted peakload within acceptable ranges for the three studied measures.Thisstudyexaminesprecoolingstrategi
27、esindifferentresiden-tial building types in a hot climate and helps customers andutilities select the optimum strategy based on cost savings andon-peak energy reductions.PROJECT APPROACHThe typical methodology for assessment of peak loadshifting strategies of a building requires a difficult and expe
28、n-sive audit of the building to estimate the effects of applicablestrategiesandtechnologies.Someresearchhasusedmodelingandsimulationtodetermineabuildingsenergyuseanddiffer-ent energy saving measures that reduce the energy use of thatbuilding (Arababadi 2012; Christian 1982; Mata et al. 2013;Kosnyeta
29、l.2001;Kintner-MeyerandEmery1995;Robertson1985; Byrne and Ritschard 1985; Yin et al. 2010b). In theory,such building simulations can be used to evaluate peak loadshifting technologies, but even carefully constructed modelsmight deviate from real situations. Thus, this research lever-ages both experi
30、mental results and simulation modeling toevaluate different peak load shifting strategies.The project started with selection of sample buildings,which are expected to represent the majority of residentialbuildings in the Phoenix climate. Once the sample buildingswere selected, we collected the requi
31、red data to simulate themin EnergyPlus (DOE 2015). The energy models were used totest a total of 12 precooling strategies, and the optimum strat-egy was found. As the next step, the optimal strategy wasimplemented in each sample building and experimental datawere recorded. Experimental work provided
32、 improved reli-ability of the impact of selected strategies on the peak load.Field tests of optimal strategies were conducted on all samplebuildings from Monday, July 22, through Friday, August 5,2015.With access to the annual and daily energy use of thesample buildings (through the utility smart me
33、ters), we cali-brated our models. The calibrated energy models were used toquantify the cost and on-peak energy savings (if any) associ-ated with various strategies.Published in ASHRAE Transactions, Volume 122, Part 2 204 ASHRAE TransactionsSample Buildings SelectionThesimulationandexperimentalworkw
34、ascarriedoutfora number of buildings considered representative of the entireresidential building stock in the Phoenix area. The number ofsuchsamplebuildingswasdecidedbasedontheirconstructiontype. The number of sample buildings chosen was a compro-misebetweenaccuracyandfeasibility,becausethemoretypes
35、of buildings there are, the more precisely the stock is repre-sented, but also it becomes more difficult to make the simula-tions work and carry out the experiments.In this study, three sample residential buildings wereselected. The first simulated home (sample building 1) is asingle-story house wit
36、h wood frame construction located inMesa, Arizona. The second sample building (samplebuilding 2) is a two-story wood frame home located in Chan-dler, Arizona, and the third building (sample building 3) is aone-story home with block construction in Scottsdale,Arizona. These models provide a represent
37、ation of the mostcommon residential construction types in the Phoenix area.Figure 1 shows the three selected homes.Description of Sample BuildingsThe modeling approach requires detailed informationabout the physical characteristics of the homes, their air-conditioning systems, local weather, indoor
38、temperature, andinternalloads.Thisinformationwasgatheredbysitevisitsandextraction from various databases and previous case studiesand through interviews of the homeowners. The physical andthermal characteristics of the homes were modeled in Ener-gyPlus v8.1 (DOE 2015). Table 1 lists the initial simu
39、lationinputs for the studied sample buildings. Required data foroperation schedules such as lighting, electrical equipment,and air-conditioning systems as well as occupancy scheduleswere gathered through interviews with the homeowners.In this project, the building models were initially devel-oped us
40、ing the Legacy OpenStudio Plug-in for SketchUp,which creates an EnergyPlus input file. The detailed parame-ters such as the heat pump models, schedules, lighting, andequipment were added into the models by OpenStudio 1.5.0(NREL 2015).Precooling Strategies DescriptionOur previous work (Arababadi and
41、Parrish 2015) made itclear that simple precooling strategies are more appropriatefor residential buildings because they provide more economicand on-peak energy savings and are easy to implement withresidential thermostats. In the current work, 12 precoolingstrategies are examined on the sample build
42、ings. The strate-gies vary by precooling period, precooling setpoints, and on-peakhoursetpoints.Theprecoolingstrategieswiththelongestprecooling periods precool the buildings for eight hours by1.67C, 2.22C, and 2.78C (3F, 4F, and 5F). The secondsets of strategies precool the houses 1.67C, 2.22C, and2
43、.78C (3F, 4F, and 5F) for five hours. This project alsoexamines precooling strategies with three hours of precoolingwith and without higher setpoints for the on-peak hours.Lastly,weexaminedathermostatsettingthatdoesnotprecoolthe house but has a 1.67C (3F) higher setpoint for on-peakhours. Table 2 li
44、sts the examined strategies in this project.Note the base case assumes a constant temperature profilethroughout the day.Experimental WorkFollowing completion of simulation work to verify that thesimulationresultsareclosetorealvalues,experimentalworkwasdone on all three sample buildings. The experime
45、ntal workstarted by implementing the optimal precooling strategy, foundby simulation models, in the sample buildings. Once the thermo-stats were programmed, the houses were monitored for a periodof 10 weekdays to measure energy use and indoor/outdoortemperatures. We measured the energy consumption a
46、nd indoorand outdoor temperatures by means of utility smart meters andtemperature data loggers, respectively (see Figure 2). Tempera-ture data loggers have a measurement range of 35C to +80C(a) (b) (c)Figure 1 Selected sample buildings: a) one-story wood frame (sample building 1), b) two-story wood
47、frame (samplebuilding 2), and c) one-story concrete block (sample building 3).Published in ASHRAE Transactions, Volume 122, Part 2 ASHRAE Transactions 205Table 1. Simulation Model InputsInput Parameters Sample Building 1 Sample Building 2 Sample Building 3Total floor area 147 m2(1584 ft2) 183 m2(197
48、8 ft2) 143 m2(1540 ft2)Floor height 3.92 m (9.6 ft) 2.7 m (9 ft) 2.4 m (8 ft)Window-wall ratio, % 11.2 12.33 9.35Exterior wall U-factor0.75 (W/m2K)(0.1321 Btu/hft2F)0.429 (W/m2K)(0.0755 Btu/hft2F)0.85 (W/m2K)(0.1497 Btu/hft2F)Floor U-factor0.31 (W/m2K)(0.0546 Btu/hft2F)0.51 (W/m2K)(0.0898 Btu/hft2F)1.17 (W/m2K)(0.2060 Btu/hft2F)Roof U-factor0.21 (W/m2K)(0.0370 Btu/hft2F)0.12 (W/m2K)(0.0211 Btu/hft2F)0.15 (W/m2K)(0.0264 Btu/hft2F)Window U-factor6.50 (W/m2K)(1.1447 Btu/hft2F)2.72 (W/m2K)(0.4790 Btu/hft2F)1.57 (W/m2K)(0.2765 Btu/hft2F)Infiltration 0.53 ach 0.53 ach 0.53 achHVAC coeffici
