AIVC #13,715 1
El'JEI\.GY 1-\ND BUJLDJNGS ELSEVIER Energy and Buildings 33 (2001) 151-166 www.elsevier.com/locate/ enbuild
Predicting air-conditioning energy consumption of a group of buildings using different heat rejection methods * a b F.W.H. Yik , , J. Burnetta, I. Prescott
"Department of Building Services Engineering, The Hong Kong Polytechnic University, Hong Kong SAR, China b Scott Wilson (HK) Ltd., Hong Kong SAR, China Received 19 August 1999 ; accepted 2 August 2000
Abstract
Water-cooled air-conditioning systems (INACS) are, in general, more energy efficient than air-cooled air-conditioning systems (AACS).
The Laws of Hong Kong prohibit the use of fresh water from town mains for comfort air-conditioning, which rules out the use of cooling towers to most air-condi.ti.oned buildings. In the absence of lakes or major rivers in Hong Kong, only those buildings that are situated near the harbor can adopt WACS. As a measure to cut down greenhouse gases emissions, the Hong Kong Government is cun-ently exploring the feasibility and viability of widening the use of WACS by the provision of centralized, district-wide water supply systems, which include seawater supply systems for once-through condenser cooling or for making up of water losses at cooling towers, and district cooling systems. Evaluation of the likely energy, economic and environmental benefits of such capital-intensive infrastructure developments requires estimation of the simultaneous cooling demand of a large group of buildings on a district scale, and the energy use for air conditioning such buildings. This paper describes the method used in these estimations. Comparisons of the results estimated by this method with results obtained by detailed simulation, and with building energy use data obtained from surveys and audits are presented.
© 2001 Elsevier Science B.V. All rights reserved.
Commercial buildings; Energy; Hong Kong; air-conditioning systems; Cooling load profiles; Electricity consumption Keywords: Simulation; Water-cooled profiles;Surveyed data consumption
1. Introduction WACS. For buildings that are remote from the harbor, WACS could still be adopted in conjunction with using Being intensive electricity consumers, buildings world cooling towers. The water supply needed for replenishing wide share a substantial responsibility for the consumption the water losses at cooling towers would be 1-3% of the of nonrenewabl�;energy resources and the greenhouse gases condenser water flow rate only, which could be coped with emissions consequent to electricity generation. For modem by the town mains for potable water supply. Unfortunately, cities in the tropical and sub-tropical regions, air-condition the Laws of Hong Kong [1] prohibit the use of fresh water ing systems are the dominant electrical loads in buildings. In from the town mains for comfort air-conditioning, as fresh such cities, improving the efficiency of using electricity for water is regarded as a scarce resource. Consequently, the air-conditioning in buildings is, therefore, a key measure for majority of buildings in Hong Kong are restricted to AACS. conservation of energy resources and protection of the Such a restriction is uncommon amongst modern cities environment. worldwide. Water-cooled air-conditioning systems (WACS) are, in In view of the substantial reduction in electricity con general, more energy efficientthan air-cooled air-�ondition sumption that can be made possible through facilitating ing systems (AACS). In consideration of the long-term buildings that are currently restricted to using AACS to saving in energy cost, many buildings in Hong Kong adopt or convert to WACS, the Hong Kong Government is that are situated near the harbor and have pump house(s) exploring the feasibility of widening application of WACS in at the seashore available for their use are already using Hong Kong [2]. This would require the development of centralized systems for supplying seawater to buildings for the purpose of heat rejection, or chilled water that can be * Corresponding author. Tel.: +852-2766-5847; fax: +852-2774-6146 used directly for air-conditioning. The schemes being inves E-mail address: [email protected] (F.W.H. Yik). tigated include the following:
0378-7788/01/$ - see front matter© 2001 Elsevier Science B.V. All rights reserved. PTI:S0 378-778 8( 00)0 009 4-3 152 FW.H. Yik et ol.IE11ergy and Buildings 33 (2001) 151-166
1. Centralized piped seawater supply for condenser cooling and DCS schemes, are capital-intensive infrastructure 3. District cooling systems (DCS). environmental benefits of each of the schemes, and thf capital investment and the degree of difficulty that will bt A CPSSCC system will comprise a centralized seawater involved in the· implementation. Greater difficulties an pumping station and a supply and return piping network for anticipated in the implementation of the CPSSCC anc distribution of seawater to the connected buildings for once DCS schemes in developed districts, but much less so ii through condenser cooling, and for the used seawater to new districts, as there will be much less inte1rnption from th1 return to the harbor. A CPSSCT system is similar to a CPSSCC engineering work and will not involve replacement o system, but the flowrate to be delivered will be much smaller equipment in existing buildings. The marketability am and there is no need for a return pipe, as the majority of the potential of attracting private investments, the scale am seawater will be lost at the cooling towers, and the relatively stages of implementation etc. will also need to be carefull: small amount of water bled-off from cooling towers for assessed. water quality control can be disposed off through the public sewers. Consideration is also being given to allowing build Comparison of air-conditioning electricity consump ings to use fresh water from the town mains for making up 2. of cooling tower water losses, but this scheme can be imple� tion of existing buildings using AACS and WACS mented only in those districts that have water supply systems capable of coping with the additional water demand. In order to obtain an indication of the energy saving th� A DCS is a centralized plant that supplies chilled water to would be achievable by converting AACS in buildings t a group of buildings. When connected to a DCS, a building WACS, the elect1icity consumption data of 23 existin will no longer require a major chiller plant of its own, but office/commercial buildings and 16 existing hotels in Hon will still require a pumping system for circulation of the Kong were analyzed and compared. These data were co· chilled water to the airside equipment. A DCS is able to take lected in recent energy surveys and audit studies [5,6]. Tb advantage of the diversity in cooling demands amongst the electricity consumption data shown in Table 1 are tb connected buildings, and can be run more efficiently, parti landlords' electricity consumption of the surveyed offic1 cularly during the light load periods (e.g. between late night commercial buildings, whilst those summarized in Table and early morning in cool winter days). are the annual air-conditioning energy consumption of tt There is no evidence that CPSSCC systems have been audited hotels. widely used on a large scale in any major cities. CPSSCT The landlord's consumption refers to the total electrici1 systems that are dedicated for air-conditioning purposes consumption of central services systems in a building, whic would also be rare, since few places would have the restric includes air-conditioning and mechanical ventilation sy tion on the use of potable water supply for making up of terns, lighting installations in public areas, vertical tran cooling tower water losses. However, district heating and portation systems, plumbing and drainage systems, ar cooling systems have already been widely used for decades various other systems connected to the landlord's electtic in many countries, particularly in North Ametica, Europe power system. Te nants in a building are supplied wi and Japan, for providing steam, hot water and/or chilled electricity through feeders that are separate from tho: water to buildings. More recently, DCS plants can also be for the landlord's services and their consumption is indi> found in Southeast Asian countries such as Singapore and dually metered by the power company, and are in general n Malaysia. available for audits. Combining a DCS with plants that provide other energy The annual electricity consumption data summarized services, such as heating and electricity, can further increase Tables 1 and 2 have been normalized by the gross floor are the overall efficiencyof providing the energy services. It has (GFA's) of the respective buildings, and are referred to been claimed that, through district cogeneration systems, energy use intensity (EUI) or air-conditioning energy u efficiencies can be boosted to 70-80%, whilst conventional intensity (NC EUI). The "mean" values shown in Table power plants have an average efficiency of about only 30% are the area-weighted mean electricity consumption of t [3]. Encouraging investments in cogeneration plants, how landlord's services amongst the surveyed office/commerc: ever, would require deregulation of electricity supply [4], buildings. but this will remain impossible in the near future in Hong It can be seen from Tables 1 and 2 that even though t Kong, albeit the power companies that have the monopoly to same heat rejection medium is used, the electricity cc generate and sell electricity could also be the providers of sumption for air-conditioning can vary· largely from o cooling service for buildings. building to another. Such variations could arise from t The centralized systems under consideration for differences in the building envelope designs, maintain implementation in Hong Kong, especially the CPSSCC indoor conditions and ventilation rates, hours of operatic (2001) F.WH. Yik et al. I Energy and Bitildings 33 151-166 153 Table J Characteristics and the landlord's EUI of 23 office/commercial buildings in Hong Kong" Description No. of buildings Minimum Maximum Mean Tot<\] (A) Buildings with AACS Direct air-cooled - (1) 10 Area per floor in office tower (m2) 1 1200 213 2 10 Total GFA in office tower (m2) 4050 10200 29800 298 000 Area per floor in podium (m2) 0 5530 1830 Total GFA in podium (m2) 0 33200 11900 54 300 Total GFA of building (m2) 5270 105000 35200 352 000 Landlord's EUI (kW h/m2) 110 250 206 (2) Indirect air-cooled - 2 Area per floor in office tower (1112) 1180 1710 1530 Total GFA in office tower (m2) 50800 145000 98000 196 000 Area per floor in podium (m2) 1300 6890 4740 Total GFA in podium (m2) 13000 110000 61600 123 000 Total GFA (m2) 63800 256000 160000 319 � 000 Landlord's EUI (kW h/1112) 195 315 219 (3) All buildings with AACS - 12 Area per floor in office tower (1112) 213 2 110 1310 Total GFA in office tower (m2) 4050 145000 41100 494 Goo Area per floor in podium (m2) 0 6890 3190 Total GFA in podium (m2) 0 110000 20200 l 7 8Goo Total GFA (m2) 5270 256000 56000 6n 0o0 Landlord's EUI (kW h/ni2) 110 315 212 (B) Buildings with WACS (1) Indirect SW cooled with cooling towers - 17 Area per floor office tower (m2) 4 9 1860 in 6 3450 Total GFA in office tower (m2) 7500 122000 59400 4160 oo Area per floor in podium (m2) 750 5100 2230 Total GFA in podium (rn2) 2250 20400 8930 62 s00 Total GFA (m2) 9750 138000 68300 478 Landlord's EUI (kW h/m2) 134 316 177 0oo (2) Indirect SW cooled with heat exchangers - 4 Area per floor in office tower (m2) 975 22 10 1440 Total GFA in office tower (m2) 21200 39600 28500 l 14oOo Area per floor in podium (m2) 1730 2920 2430 Total GFA in podium (m2) 3690 14600 9700 38 900 Total GFA (m2) 34900 46500 38200 1530 0o Landlord's EUI (kW h/m2) 182 249 211 (3) All buildings with WACS - 1 1 Area per floor in office tower (m2) 469 3450 1750 !. Total GFA in officetower (m2) 7500 48200 530 122000 o0o Area per floor in podium (m2) 750 5100 2300 Total GFA in podium (m2) 3690 20400 57400 10100() 63 Total GFA (m2) 9750 138000 57400 10oo Landlord's EUI (kW h/m2) 134 316 185 "Energy use intensity (EUI) means the annual electricity consumption per square meter of the floor area in the building; GFA means gross floor a te�. intensity of internal loads and occupancy densities in the The differences in electricity consumption are quite sub. different types of premises in the buildings, and in the stantial and be regarded as the "average" saving may that efficiency of the air-conditioning systems. Nonetheless, it would be achievable by using WACS in lieu of AACS, fora can be seen that there is a significant difference in the mean group of office/commercial and hotel buildings, impleme . n EUI values between buildings that are equipped with AACS ted on an indiv:idual building basis. Even greater saving may and WACS. For office/commercial buildings, the difference be expected if a DCS is employed to provide chilled Water U in the average landlord E I is 27 kW h/m2 (=212-185, for a group of buildings. The difference for hotel is larger Table 1), and for hotels, the difference in the average Al becau e their air-conditioning ystems are continuously 49 C EUI is kW h/m2 (=217-168, Table 2). operated, whereas sy terns are only iotenuittently operated 151-166 154 F.WH. Yiket al.IEnergy and Buildings 33 (2001) Table 2 Characteristics and NC EUI of 16 hotel buildings in Hong Kong" Description No. of hotels Minimum Maximum Mean Total (A) Hotels with AACS -6 Grade (no. of stars) 3 5 3.8 Year built 1988 1993 1990 No. of guest rooms 216 500 372 2232 No. of restaurants 6 2.5 15 Total GFA of building (m2) 12900 32600 21400 129000 Annual energy consumption per m2 (MJ/m2) 1570 2950 2100 EUI (kW h/m2) 320 439 384 NC EUI (kW h/m2) 166 257 217 (B) Hotels with WACSb -10 Grade (no. of stars) 4 5 4.7 Year built 1969 1994 1983 No. of guest rooms 450 862 591 5909 No. of restaurants 8 5.5 55 Total GFA of building (m2) 28800 66900 44800 448000 Annual energy consumption per rn2 (MJ/m2) 1538 3371 1980 EUI (kW h/m2) 250 844 412 NC EUI (kW h/rn2) 128 231 168 (C) All hotels in the sarnpleb - 16 Grade (no. of stars) 3 5 4.4 Year built 1969 1994 1986 No. of guest rooms 216 862 509 8141 No. of restaurants 8 4.4 70 Total GFA of building (m2) 12900 66900 36000 577000 2 Annual energy consumption per m2 (MJ/m ) 1070 3370 2030 EUI (kW h/m2) 250 844 406 AIC EUI (kW h/m2) 128 257 177 " AIC means air-conditioning; energy use intensity (EUI) means the annual electricity consumption per square meter of the floor area in the building; GFA means gross floor area. b Envelope design: (A) curtain wall -1, window wall -5; (B) curtain wall -6, window wall - 4; (C) curtain wall -7, window wall -9. in office/commercial buildings. Also, many hotels in the capacity required of the central cooling water supply system sample use direct seawater-cooled chiller plants, which are or the district cooling system, which would ensure that the more energy efficient than the indirect seawater-cooled plant capacity so sized would be sufficient to cope with the plants and plants with seawater cooling towers that are more cooling demand for a sufficiently high confidence level commonly used in the surveyed office/commercial buildings (97.5% in the cooling months is recommended by American (Table 1). Society of Heating, Refrigeration and Air Conditioning These results are prima facie evidence that promoting Engineer (ASHRAE) [7]). The realistic cooling load wider use of WACS would be very effective in reducing predictions would allow the simultaneous demands for electricity use for air-conditioning buildings in Hong Kong. electricity and for condenser cooling water or chilled water, at different times in the year, to be estimated. The relative benefits of the WA CS scheme options can then be Cooling load prediction methods 3. assessed. Three simulation programs, namely HKDLC [8], HTB2 Detailed evaluation of the energy, economic and environ [9] and BECON [10], were used as prediction tools in this mental benefits that could be realized by adopting each of study. HKDLC is a design cooling load calculation program, the WACS scheme options requires the ability to predict the whereas HTB2 is a detailed building heat transfer simulation simultaneous cooling load of a variety of buildings on a program that can predict the realistic cooling loads of a "district" scale. Such buildings would comprise a mix of building at different times of the year. BECON is an air uses, such as offices, retail shops, restaurants and hotel conditioning system simulation program, which can predict guestrooms. Both the simultaneous design peak cooling the hourly power consumption of air-conditioning equip load and the simultaneous realistic cooling load of the ment on the basis of the cooling loads predicted by HTB2. buildings in a district need to be predicted. The design More detailed descriptions of these programs are given in cooling load prediction would be needed to determine the Appendix A. et al.I Energy and Buildings F.WH. Yik 33 (2001) 151-166 155 Although HKDLC, HTB2 and BECON could be used particular time in the year can be determined as follows: repeatedly to generate cooling load and air-conditioning electricity consumption predictions for a large number of Qo,cotat(!) = LQo,k(t) (3) buildings, their use would require the preparation of rather k detailed input data for adequately describing the buildings Qcotal(t) = Qk(t) (4) and the air-conditioning systems. Given the large number of Lk buildings that would be involved in studying the perfor mance of district-wide systems, the time and effort required From the hourly results, the simultaneous peak cooling load for preparing the input files and to perform the simulation on a district cooling system can also be determined. work are prohibitive. Therefore, a simplified method has In order that the models would be able to provide good been developed for faster determination of the profiles of predictions for a large number of buildings, the cooling load design and realistic cooling loads, and the air-conditioning intensity data in the models should be representative of the electricity consumption in buildings. "average" cooling load intensities of the corresponding types of premises in buildings in Hong Kong. Cooling load profile rnodels 3.1. Office premises 3.2. The approach adopted was to establish a design and a ,,• realistic cooling load profile model for each major type of The survey on the 23 office/commercial buildings (sum- premises that can be found in typical air-conditioned build marized in Table 1) did not aim at obtaining comprehensive ings. Since offices,retail shops, restaurants and hotel guest information about their building construction, which rooms are the major types of premises in buildings in a restricted the use of these buildings in detailed cooling load commercial district, models were established for each of simulation studies. Instead, the design and realistic cooling these premises, which were based on detailed simulation load profile models for office premises were established predictions using HKDLC and HTB2. In order to make the based on the office towers of six existing office/commercial models applicable to premises of different floor areas, the buildings, five new building designs and one hypothetical predicted cooling loads were normalized into intensities of office tower design, of which the required data were avail cooling loads per square meter of the GFA of the premises. able. Characteristics of the 12 buildings and the internalload Having established these cooling load profile models, the criteria adopted in the simulation studies are summarized in cooling load of an office/commercial building complex can Table 3. The cooling load profiles were developed fromthe be estimated from the office, retail shop and restaurant sum of concurrent cooling loads of the 12 buildings, as models. Likewise, the guestroom, retail shop and restaurant predicted by HKDLC and HTB2, and normalized by the models will allow the cooling load of a hotel to be estimated. total GFA of these buildings. The method is expressed mathematically as follows: The patterns of occupancy and lighting loads used in the simulation study are summarized in Tabl s of e 4, in fraction = qo ; t Qo,k(t) L , ( )Ai,k (1) their design intensities. Appliance loads in office buildings, 2 i=l,2,3,4 however, were �ssumed to remain at 25 W/m during the occupied hours. The lighting load intensities shown in Qk(t) = q;(t)Ai,k (2) L Table 3 are actual figures used in designing air-conditioning i=l,2,3,4 s1stems forthe buildings. The average design lighting load , (t) 2 where Qok is the design cooling load of building k at intensity of these 12 office buildings is 20.1 W/m . time t (a given month and hour in the day) (kW), q0,;(t) It was found in a recent energy end-use survey [11] that the normalized design cooling load of type i premises at lighting energy use in offices in Hong Kong amounted to 2 time t (kW/m ), Qk(t) the cooling load of building at time 1906 TJ in 1994. The size of this building segment was k 2 t (a particular hour among the 8760 h in a year) (kW), 6,895,000 m . This reported energy use figure however q;(t) the normalized cooling load of type premises at time includes energy use of lighting installations both inside 2 i t (kW/m ), and where A;,k is the area of type premises and outside air-conditioned office spaces, such as lighting 2 i in building k (m ). in staircases, lift lobbies, toilets, etc. The calculated average 2 As shown in Eqs. (1) and (2), each cooling load profile figure of 20.1 W/m would lead to an annual electricity model (q0,;(t) or q;(t)) comprises a time series of the year consumption of 1738 TJ for the same total area of officesas round hourly cooling load intensities of a specific type of in 1994 over a year-round total of 3484 operating hours of premises. This is an essential feature to allow the cooling (the total hours of operation in a year used in the simulation). load variation patterns of different types of premises to be This accounts for 91% of the total energy use for lighting in taken into account in the determination of the simultaneous the office building segment in 1994, which is considered a cooling load of a building. reasonable figure for lighting installations inside air-condi The simultaneous design cooling load (Qo,totai(t)) and tioned office spaces. Therefore, the average value of the ( w1a1(t)) gs realistic cooling load Q of a group of buildings at a design lighting load intensities of the 12 office buildin ,.... "' 0\ Table 3 Summary of characteristics of office towers" Building B G (model) A c D E F H I J K L 2 GFA(m ) 61600 34300 23500 1904 36860 70140 51840 48960 21000 55480 61440 32160 � 1.6 1.5 1.4 1 4.4 1.1 1 1.1 1.8 1.5 1.6 1.5 � Aspect R � Major exp SW N E,W SE N and S N, E, S and W E,W N, S N, S N and W NE, w w ,.,.� WWR 0.16 0.8 0.49 0.4 0.53 0.7 0.5 0.58 0.18 0.35 0.29 0.42 2 � Wall U(Wlm 1.4 2.8 0.2 1.9 2.2 0.5 0.9 0.7 1.2 0.6 0.7 0.4 2 K) f2- 0.7 0.5 0.5 0.7 3.5 l 0.5 0.4 NIA NIA 0.35 0.58 '- Roof U(Wlm K) 2 2.7 5.3 5.5 3.1 5.3 4.9 5.4 5.3 5.1 5.3 5.3 5.2 ;:sl"rJ Glass U(W/m K) "' 0.45 0.5 0.25 0.63 0.53 0.73 0.45 0.2 0.3 0.34 0.22 0.24 Glass SC �"' 2 OTTV (W/m ) 11.8 38.8 18.1 15.5 39.1 75.2 34.7 20.7 15.8 19.2 17.8 15.7 "' Indoor (°C) 25.5 23 24 23 24 23.5 25.5 24 25.5 25 24 24 5. T � Indoor RH(%) 54 55 55 55 55 55 54 55 55 54 55 55 " 2 Occup (m per person) 14 9 9 9 9 14.2 9 9.4 9 5.9 8 9 2 °"� Lighting(W/m ) 16 25 24 25 18 14 25 21.3 25 21.6 25 "' 2 25 "" Appliance(Wlm ) 25 25 25 25 25 25 25 25 25 25 25 25 "" V. rate(!Is per person) 14.2 13.4 7 8 5.1 12 10 6.7 10 6.3 6.5 10 "Na Inf. Rate(A CH) 0.1 0.l 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 a .'.:::: HRM DSWC InDSWC DSWCCT Aircool Aircool InDSWC Aircool Aircool Aircool JnDSWC DSWC DSWCCT '-"..... COP 4.1 5.1 4.6 2.5 2.4 4 2.8 2.6 2.9 4.1 5.2 4.8 I...... Airside S fcu vav vav cav vav cav +vav vav vav vav vav vav· vav Cl\ Cl\ "GFA: gross floor area; Occup: occupancy; DSWC: direct seawater cooled; Aspect R: aspect ratio of floor plan; Lighting: lighting load intensity; InDSWC: indirect seawater cooled; Major exp: major exposure; Appliance: appliance load intensity; DSWCCT: DSWC with cooling tower; WWR: window to wall area ratio; V. rate: ventilation rate; Aircool: air cooled; Wall U: average U-value of wall; Inf. rate: infiltration rate; fcu: fan coil unit system; Roof U: average U-value of roof; HRM: heat rejection method; cav: constant air volume system; Glass U: U-value of glass; COP: rated cop of chiller; vav: variable air volume system; Glass SC: shading coefficient of glass; Airside S: air- side system design; cav vav: dual conduit system; OTTV: overall thermal transfer value; ACH: air change per hour; Indoo1· T: indoor dry- + bulb temperature; Indoor RH: indoor relative humidity. F.H'.H. Yik el al.IEnergy and Buildings 33 (2001) 151-166 157 Table 4 Occu ation, lighting air su ly system operating for office premises p load profiles and fresh pp schedule Day in the week Hour Occupancy Lighting (perimeter) L' ::--_ 1sh1in . up y �terior) Fresh air s pl Weekdays 1-7 0 0.05 0.05 7-8 0.05 0.1 0.1 Off 8-9 0.4 0.5 0.5 Off 9-10 0.95 0.9 On 10-11 0.95 0.9 On 11-12 0.95 0.9 On 12-13 0.95 0.9 On 13-14 0.95 0.8 0.9 On 14-15 0.95 0.9 On 15-16 0.95 0.9 On 16-17 0.95 0.9 On 17-18 0.5 0.8 0.8 On 18-19 0.25 0.5 0.5 On 19--20 0.1 0.3 0.3 On 20-21 0.05 0.2 0.2 Off 21-24 0 0.05 0.05 iOff Off Saturdays 1-7 0 0.05 0.05 7-8 0.05 0.1 0.1 Off 8-9 0.3 0.5 0.5 Off 9-13 0.6 0.75 0.8 On 13-17 0.1 0.2 0.2 On 17-18 0.05 0. 1 0. 1 Off 18-24 0 0.05 0.05 Off Off Sundays 1-9 0 0.05 0.05 9-17 0.05 0.1 0.1 Off 17-24 0 0.05 0.05 Off Off may be considered representative of office buildings in Hong patterns of variations in . . occup _ appliances Kong. loads used in the si l ancy, tmg and mu ation l ligb . S 'lldy ummanze. d . Ta bl e 5 In the energy end-use survey [11), the energy consump The simulated cooling load are . m · Pro normalized by the tion of equipment in office buildings in Hong Kong in 1994 GFA of the retail and restau files were . tan11ll floors, respect1ve1 y, to was found to be 2192 TJ. When converted into electricity yield the cooling load prol1J Odel _ � e lllo premises. consumption per unit area, this amounts to about 88 kW h/ On the basis of tbe load dels for these ' lactors . . eqm -pment m2. Assuming the annual operating hours to be also 3484 h, loads in retail and for hgh.tmgand restau t the annual the steady rate of electricity consumption of the equipment operating hours the n tan premises, and an ua i . , l ee<:" consumption o fl'hig t- would be about 25.3 W/m2. It can be seen that this value ing installations and equi lt1c11y and restaurant compares well with that used in the simulation of cooling premises have been calcul Pl11eot in retail · ate<1 ummanze· d m Ta ble 5 · load for the 12 buildings. The load factors (equivaletit fua n d . � h s as a fracti�n of the total operating houxs) er l-Joad �ur W det on the basis of Retail shops and restaurants h ermmed 3.3. the assumed operatina ou s intensily profiles of light- incr and equipment lo ds r:, band . . "' ; \ la . d eqmp- · le 5 . The lightmg an The cooling loads of retail shops and restaurants are ment energy consum tion ) p n t · xestaurant b m·1d· mg usually dominated by the internal loads from the occupants, segments given in the ener re1a1 1 and survey report [11) is lighting and appliances, and by the ventilation loads, summarized in Table 6. gy end-u se whereas the building envelope design will have relatively The consumption figures sh . Table .6 are the tot l little influence. Therefore, representative cooling load pro values of all retail shops su own m � h include those m files can be generated as long as the heat gain profiles and shopping malls and t whic a stree�:Yed, exa Ho�ever, clos�r � ventilation rates used in the simulation studies can reflectthe in a ti on of the data revealed lha�tdcs. lheti_ loads of shops typical conditions arising in retail shops and restaurants. commercial buildincrs witli ghtrng m "' · g were Cent I w·-con· d'ti 1 omn The cooling load profile models for retail shops and sicrnificantly hicrher than in ra buildi gs. T�e restaurants were each derived on the basis of a model floor, a:erage lightin; load of retaj�thcr type of � in centrally mr-cond1- with layout and envelope designs conforming to a retail shop tioned buildings wa found shops 2 . lO b him , hic h is floor and a restaurant in an existing commercial building. close to that on which the 111. e 297 kW w kW h/m2, The �el wa based (292 internal load and ventilation criteria adopted and the Table 5). Moreover ' the e qu1Plllem consumption in retail ...... U\ 00 Table 5 Occupancy, lighting and appliance loads and ventilation rates for retail shops and restaurants !>'l Time (h) Pattern in the day (fraction of design intensity) ::::::: � Retail Restaurant ""� Occupancy- Lighting- Appliances - Ventilation rate- Occupancy- Lighting- Appliances - Ventilation rate - � "' 2 2 2 2 2 2 :-- 5 m per person 70W/m 30 W/m 7 l/s per person 1.5 m per person 30W/m 30W/m 7 l/s per person ' gi 0 0 0 0 0 0 0.05 0 "' 0--6 o;i 6-10 0 0 0 0 0.6 0.9 0.75 1 '< "' 10--11 0.25 0.95 0.75 1 0.6 0.9 0.75 1 "-:. 11-12 0.25 0.95 0.75 1 0.9 0.9 0.75 l b:i 12-14 0.75 0.95 0.75 1 0.9 0.9 0.75 1 14-18 0.25 0.95 0.75 0.05 0.5 0.6 1 � 1 "";:. "' 18-22 0.75 0.95 0.75 1 0.75 0.95 0.75 1 '""' 22-24 0 0 0 0 0.05 0.25 0.1 1 '""' ;:,: w llA\JI a Occupied hours per day 12 12 12 12 18 18 18 18 a ,:::: Load factor 0.95 0.75 0.75 0.5 '-"...... Annual electricity 292 99 148 131 .!... °' 2 consumption (kW h/m ) °' (2001) 151-166 F. W.H. Yik el ell. IEner gy and Buildings 33 159 Table 6 consistent values would result . . Lighting and equipment load in shops and restaurants if h load intensities were retail expressed on per umt. area basis. t e d" . . . ' ue to vaned room sizes Unit Retail Restaurant the amongst diueren,,., th o t e 1 s. ? Total area m- 5214008 3092287 Normalized cooling load 3.5. pr,'"' Lighting consumption TJ 4327 1890 0J zles 2 kW h/111 231 170 2 c kW h/m 297" 144" Fig. l(a) shows the hourly o01 1· . ds per square mete 1 of GFA for the four types of no loa Equipment consumption ° . .· hi h TJ 2581 4110 pretri·1Ses day 1 w c 2 were extracted from the m. a m Ju y' kW h/m 138 369 • • houtI" J 2 coolm }l,.. .; 160 F.WH. Yik et al. /Energy and Buildings 33 (2001) 151-166 Table 7 Characteristics of the model for guestroom floors in a hotel building Type % of Time Fraction Fraction of Fraction of Indoor guestrooms of full maximum maximum small temperatu occupancy lighting load power load (oC) (A) Indoor conditions adopted in design load estimation for energy simulation 1 72 0-8 1.00 0.17 0.54 24 8-20 0.00 0.20 0.54 28 20-24 1.00 0.97 0.99 24 2 20 0-8 1.00 0.17 0.55 24 8-20 1.00 0.85 0.65 24 20-24 1.00 0.97 0.99 24 3 8 0-8 0.00 0.17 0.47 28 8-20 0.00 0.20 0.47 28 20-24 0.00 0.17 0.47 28 Overall 100 0-8 0.92 0.17 0.54 8-20 0.20 0.33 0.56 20-24 0.92 0.91 0.95 (B) Indoor conditions adopted in design load estimation 0-8 1.00 0.17 0.54 24 8-20 0.25 0.33 0.56 24 20-24 l.00 0.91 0.95 24 (C) Characteristics and indoor design conditions for guestrooms 2 Gross floor area - 20,000 m No. of sto1ies - 25 No. of guestrooms - 500 Aspect ratio - 2.44 Major exposure - SE/NW Window to wall area ratio - 0.5 for the maj or facade 2 Wall U-value - 1.74 W/m for the maj or facade 2 K Roof U-value - 0.4 W/m K2 Glass U-value - 5.83 W/m K Glass shading coefficient - 0.56 2 OT TV - 33.8 W/m Indoor temperature - 24°C reset to 28°C when unoccupied Indoor relative humidity - 50% Occupancy - two persons per room Lighting load - 600 W per room Appliance load - 100 W per room Ventilation rate - 34.1 l/s per room Infiltration rate - 0.1 ACH sumption profile, therefore, needs to be decomposed into (Wk(t)) or a group of building complexes (Wtotai(t)) as show two components: one representing the consumption that below: varies with the cooling load on the chiller plant (wvarj), and the other representing the minimum rate of electricity c: that the system will consume under zero load The (wminJ)· minimum consumption component has been expressed as a Wtotai(t) = LWk(t) (E fraction of the rated power demand of the chiller plant in the k models. The variable consumption models were de1ived where Wk(t) is the electricity consumption for air-condition from the difference between the predicted total air-condi ing building k at time t (kW), Wvar,j(t) the variable ai1 tioning system consumption (using BECON) and the mini conditioning system electricity consumption at time t, fc mum consumption at each hour of operation, which were heat rejection method j per kilowatt of cooling load on th then normalized by the corresponding cooling load on the system (kW/kW), Wmin,j(t) the minimum cou-stant air-con system. ditioning system electricity consumption at time t, for hea The air-conditioning electricity consumption profile mod rejection method j per kilowatt of installed plant capacit: els can be used to predict the electricity consumption for air (kW/kW), and where PCap is the installed cooling capacit: conditioning a building comprising a mix of premises types of the central air-conditioning plant (kW). '4!-L£ .4 F.WH. Yik et al. I Energy and Etti/dings 33 (2001) 151-166 161 0.4 demand from a group of buildings CVwia1(t)) was estimated 0.35 based on the following equation: 0.3 --+-Office 0.25 E --&-Shops ll 0.2 (7) � 0. 15 ---.!>-- Restaurants 0.1 --*-Guest Rm 0.05 where C is the conversion factor for determining the con 0 denser water flow rate from the cooling load (=0.0529 -0.05 l/s kW), Vk.inin the minimum water flow rate demand from (a) Hour the kth building (l/s), Co = 1.0 for CPSSCC scheme, C0 = = 0.35 0.017 for CPSSCT using fresh water, and Co 0.03 0.3 for the CPSSCT scheme using seawater. The total condenser water flow rate in a water-cooled air 0.25 --+-Ornce -s-Shops conditioning plant was estimated on the basis of a C value of E 0.2 ! ---.!>-- Restaurant 0.0529 l/s kW of cooling output of the chillers, in conjunc ""':: 0.15 --*-Guest Rm tion with the hourly cooling load of the building (Q1/t)), as 0.1 predicted by the cooling load profile modeL This condenser 0.05 4.5 and . I flow rate conesponds to a chiller with a COP of a 0 temperature difference of 5.5°C across the chiller condenser. 2 3 4 5 6 7 8 9 10 11 12 The condenser water flow rate so estimated would also be (b) Month taken as the water demand of a building when its plant is Fig. Cooling load profiles of offices, retail shops, restaurants and hotel 1. connected to a CPSSCC system (therefore, Co = 1.0). guestrooms: (a) hourly values in day July; (b) monthly mean vahies in a in For a chiller plant comprising multiple chillers, the con the year (based on occupied hours for the respective type of premises). denser water flow rate will vary in steps according to the number of chillers that are put into operation at a particular time. Assuming there are four chillers in each building, the By expressing the air-conditioning electricity consump condenser water flow rate will vary as shown by the dotted tion profile model as a time series of normalized hourly line in Fig. 3. values throughout the year, the effects of variations in the For simplicity in calculations, the condenser water flow outdoor air or seawater conditions in the year on the rate was assumed to be directly proportional to the cooling efficiency of the air-conditioning plants can be preserved. load (=CQk(t), as shown by Line 1 in Fig. 3). However, this Fig. 2 shows the monthly mean hourly chiller COP for obviously would lead to under-estimation of the water flow four selected months, as well as the monthly mean COP rate, except at those conditions where the running chillers values over the year, as predicted by the power consumption are fully loaded. To correct for this underestimation, Line 1 profile models for the four heat rejection methods. Fig. 2(e) is shifted upward by a constant flow rate (V1c,min,Eq. (7)) that shows that the monthly mean chiller COP would peak in equals half of the condenser flow rate for the chiller that April-May. This reflects the effect of lowering in the cooling would remain operating in the smallest cooling load range. medium temperature (air or seawater) during the mild Thus, the flow rate predicted by Eq. (7) would be related to weather months, leading to the improved chiller perfor the cooling load in the manner as shown by Line 2 in Fig. 3. mance. It would become slightly lower during the hottest This implies that the assumption that there are equal chances months of June-October. However, the monthly mean COP of over- and under-estimation has been made. would drop substantially during the cold months, which is a Water losses at cooling towers that need to be replenished result of the large reduction in cooling load, leading to by water supply from a CPSSCT system include losses due chillers running under unfavorable part-load conditions. to evaporation, drift and blow down. The required make-up flow rate is usually determined on the basis of an evaporation loss and a drift loss equivalent to 1.0 and 0.2% of the condenser water fl.ow rate, respectively [14]._The blow dow 5. Condenser water flow rate and cooling tower n make-up rate profiles rate, however, is dependent on the tolerable concentration of impurities in the system water, quantified as a ratio to the Apart from the cooling load and the air-conditioning impurity concentration in the supply water by the parameter electricity consumption, the required water supply flow rates "number of concentration" [14]. for condenser cooling and for cooling tower make-up for a The Carrier Handbook [15] recommends that the blow group of buildings are essential variables that need to be down rate should be 50-200% of the evaporation rate. As the evaluated for estimating the required output and power evaporation rate is about 1 % of the condenser water flow rate consumption of the centralized pumping systems of the and the drift rate is about 0.2%, the make-up rate would be in CPSSCC and the CPSSCT schemes. The cooling water the range of 1.7-3 .2% of the condenser water flow rate. The F. WH. Yik 162 et al. / Energy and Buildings 33 (2001) 151-166 4 3.5 -- 3 _,.__ Feb ...... _ Feb 2.5 n � - I Cl. 2 --May Q. --May 0 "'-s..,,. a -Ji "'--- __p- .... 0 3 � _.._Aug _.._ Aug u 1.5 u 2 1 -+-Nov -+-Nov 0.5 0 0 0 3 6 9 12 15 18 21 24 0 3 6 9 12 15 18 24 21 u) Hour b) Hour 6 5 5 _,.__ Feb 4 _,.__ Feb 4 Cl. 3 --May Cl. 3 --May 0 0 u u _.._ Aug 2 _.._Aug 2 -+-Nov -+-Nov 0 0 0 3 6 9 12 15 18 21 24 0 3 9 12 15 18 21 24 c) Hour d) Hour 6 5 D.. --+-Aircool 0 4 u... -G-WCCT 3 � _,.._ 1swc :E u 2 --oswc 11 2 3 4 5 6 7 8 9 10 12 e) Month Fig. 2. COP of chillers in the electricity consumption profile model: (a) monthly mean hourly COP of air-cooled chillers in selected months in the year; ( monthly mean hourly COP of water-cooled chillers with cooling towers selected months in the year; (c) monthly mean hourly COP of indirect seawatt in cooled chillers in selected months the year; (d) monthly mean hourly COP of direct seawater-cooled chillers in selected months in the year; (e) month in mean COP of air-cooled chillers, water-cooled chillers with cooling towers and direct and indirect seawater-cooled chillers. Condenser water flow rate Actual flow rate Line 2 - used in estimation Line 1, flow rate proportional to of 50% cooling load condenser water flow rate of one chiller Cooling load Fig. 3. Condenser water flow rate in a chiller plant. r 2 r: WH. Yik et al. I Energy and Buildings 33 (2001) 151-166 163 Table 8 lower end is recommended forcases where there is no severe Comparison of predictions by the simple model and b water problem, e.g. fresh water. This corresponds to a y detailed simulation: office-only building number of concentration of about 2.5 and a flow rate of 0 86 10-3 Heat rejection method AJC EUI of the office o l . x l/s kW of the cooling capacity. However, no - n y l 0 2 bui d100' explicit guideline on the blow down rate required for (kW hlm ) systems using seawater is given. ---Simple Detailed % On the basis of plant operation records of an existing model simulation Deviation building [16], where seawater cooling towers are used, the Air-cooled 104 104 -0.J annual seawater consumption in that building was found to Water-cooled cooling towers 95 91 3.5 be about 2.8% of the circulation rate. Based on a number of Ind irect seawater-cooled 92 89 2.8 1.5, Direct seawater-cooled 86 85 concentration of the required blow down rate will be l.6 1.8% and the total cooling tower water make-up rate will be approximately 3% of the condenser flowrate, which is close 2.8%. 3% to the empirical figure of The figureof is equiva involved in i the imultaneou · coo pred cting ling load e c lent to a flow rate of 1.587 10-3 l/s kW of the cooling ' le x tricity consumption and water demand of a bu.ilding capacity. com prising a mix of premises and fora group of buildings w (7) ould Therefore, the values of C0 in Eq. corresponding to a become relatively straightforward. The procedures av h e al 0 CPSSCC scheme and CPSSCT schemes using fresh water been computerized into a spread beet program, W th 1.0, 0.017 0.03, i the and seawater were set as and respectively. profile model incorporated into the program as its database. Note that the above method will be applicable only to the The input required would involve only the var . GFA of iou dete1mination of the supply flow rate for once-through premises in the buildings lo be modeled. cooling of chiller condensers (CPSSCC) or for making up he AJC EUI of the mod ! office-only building (B d. ! � . uil ing of cooling tower losses (CPSSCT). For the prediction of the G m Ta ble 3) has been predicted both by detailed iin n ulatio chilled water flowrate demand on a DCS, a differentmethod using HTB2 and BECON and b using the simp y Jjfied model. will have to be used, as the relation between the chilled water Four cases were studied the heat rejeclion with method used flow rate demand and the cooling load is nonlinear, and is in the air conditioning system varied in each cas - e. lt can dependent on the characteristics of the cooling coils in the be een from the results ummarized in Ta ble 8 that the NC air-handling units. It will be predicted by using a cooling coil values predicted by the si mplified ode EUJ m l are in good model in conjunction with the cooling load predicted by the agreement with the predictions of the detailed models. above-described simplified model for individual connected The simplified have been models also applied to predict buildings. The prediction method has been incorporated into the air-conditioning electricity consumption of the surveyed/ a special version of BECON that can simulate the perfor audited exi ·ting office/commercial buildings and hotel mance a DCS, the details which will be described in a buildings ummarized in Tables l and 2, based o of of n the total separate paper [17]. u l GFA's ofvariou types ofpremi es in these b i ding s. Table9 bows a comparison of the predicted A/C BUI Values for these buildings with estimates based on th surveyed land Trial runs and verification of predictions total con umption of the exi ling office/co111 6. lord' rnerc ial building . It also includes the metered electricity consump Having established the cooling load, electricity consump tion data of an existing offi.ce/commercial building, where tion and water demand profile models, the procedures detailed sub-meter readings were available [16]. That bu.ild- Table 9 Comparison of air-conditioning electricity consumption (office/commercial buildings) No. of Mean landlord 's Predicted AJC AJC to land lord Estimated AJC D d 2 2 b 2 2 % eviation buildings EUI (kW h/m ) EUI" (kW hlm ) EUi (kW hlm ) EU:f (kW him ) (A) Buildings with AACS Direct air-cooled 10 206 159 0.77 159 o.o (B) Buildings with WACS Seawater cooled with cooling towers 7 177 143 0.81 133 7.6 Indirect seawater cooled with heat exchangers 4 211 160 0. 6 158 7 0.9 (C) Metered data for a office/commercial build ing [16] 0.75 126d 173 130 3.0 • Predictions of the simplified model. b Ratio of pred icted AJC EUI to surveyed landlord total EUI. c Assumed to be 77% of mean landlord 's EUI for air-cooled systems, and 75% for water-cooled systems. d Audit result. 164 F.WH. Yik et al. / Energy and Buildings 33 (2001) 151-166 Table 10 40 �------Comparison of air-conditioning electricity consumption (hotel buildings) 20 -!------.-.-; 1------1 " No. of NC EUI 0 buildings � -20 ·;; Surveyed Model % Deviation 2 2 i3 (kW h/m ) (kW b/m ) � -60 Hotels with AACS 6 217 217 -0.09 �O +w----""------1 Hotels with WACS 10 168 179 6.67 -100 �------� 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 Building No. NC EUJ vnlues predicted by 1he Fig. 5. Percentage devintious between ing is air-conditioned by direct seawater-cooled chillers with eslirnntes Sllrvey/audit data - A/C EUJ simplified model and from cooling towers . The EUI for air-conditioning in that build system, assumed to be 77% of landlord's EUl for thenir·coo!ed and 75% ing, including the consumption of waterside and airside forwater-cooled systems. 2 equipment, was found to be 126 kW h/m [16]. The pre dicted A/C EUI values' and the data obtained from the audit studies for the existing hotels are shown in Table 10 for excluded). Buildings 1-10 have direct air-cooled chiller comparison. plants, 11-17 have eawater-cooled plants with cooling It can be seen from the comparisons summarized in towers and 18-2 1 have indirect seawater-cooled plants. This Tables 9 and 10 that the model predictions compare well comparison show that the model may not provide accurate with the measured data for the audited office/commercial predictions when compared to an individual building, but the building and with the mean values of the surveyed hotel predictions are close to the mean values amongst the sur buildings. Since only the landlords' total EUI values were veyed/audited buildings. available for the other office/commercial buildings, the predictions were compared with A/C EUI values estimated from the landlords' total EUI values. The deviations are 7. Conclusions within a few percent of the estimated values, although such comparisons would be subject to large uncertainties. Never Analysis of the energy consumption data of 23 existing theless, it can be seen that fue simplified model can provide office/commercialbuildings and 16 existing hotel buildings, estimates that are reasonably close to the mean values of which included buildings that use AACS and WACS, existing buildings. showed that a significant difference existed in the average Fig. 4 compares the ·profiles of monthly electricity use for annual electricity consumption between buildings using the air-conditioning as predicted by the· simplified model and two types of air-conilitiuuiug ysten:iheat rejection methods. obtained from metered readings in the audited office/com Hence, facilitating buildings that would otherwise bt mercial building described in [16]. It can be seen from this restricted to using AACS to adopt or convert to WACS figure that the predicted monthly electiicity consumption is would potentially lead to significantly reduced electricit) close to the metered readings. use for air-conditioning buildings in Hong Kong. However Fig. 5 shows the percentage deviation between the model it would require the development of centralized systems fo1 predictions and the audit/survey results for 21 of the 23 supplying water to such buildings, for condenser coolin! existing office/commercial buildings, on individual building (CPSSCC) or for making up of cooling tower losse: basis (the two with indirect air-cooled plants were {CPSSCI'). District cooling systems (DCS's) would alsc 16 �------� 1 E 4 +-���--:::�a;2:���---l � 12 -t---��-?;---���"'="'""""r--�� � 10 ..------,7'"'------=-�-� 0 -Audited E B 'F"''->-""-c�------=>-c. --Q- :i:� 6 Predicted 0 4 _,______, < 2 -1------0 +---.--,--�.----.--�--..-�---.---.-�c---i 1 2 3 4 5 6 7 8 9 10 11 12 Month Fig. 4. Comparison of monthly power consumption profiles predicted by the simplified model and obtained from metered data in an office/commerci building. F.WH. Yik et al.!Ene1gy and Buildings 33 (2001) 151-166 165 be feasible and may yield the greatest saving, but would of DCS to replace c ille . · · · h r s in the interest . . . Plants bUI id mg involve the highest capital investment. of ach1evmg higher over m. . a]J of air-conditioning Evaluation of the energy, cost and environmental benefits bm . 1d mgs. e ffic1ency of such schemes requires an efficient method for determin ing the simultaneous cooling load, water demand and air� conditioning electricity consumption of a large number of Acknowledgements buildings using different heat rejection methods. A simpli fiedmethod has been developed to allow these estimations to The w0rk presented . in th· ;; d part a con- be accomplished speedily. The method is based on cooling s t study is paper Lorme of ul ancy �Ommissi Kon oned b the Hong g SAR load profilemodels for different types of premises and power rnment, which y Gove w consul- as uud ertak:e nyb a group of consumption profile models for different heat rejection tants 'inc I ud h i ug t e lead consu S (HK) Ltd. methods. The cooling load patterns of different types of and h Hong ltant cott Wilson t e �011g oly Un P technic iversity. The authors premises are taken into account in this method, which is an thank the Electnca] and . Mechamca 1 erv1· ce s Department of essential feature predicting the cooling load of buildings the Hong Kong S for SAR Gove · . . inment for th ei 1· permission to comprising a mix of premises for different uses. These publish this work. profile models were established from detailed simulation studies. Reference was made as far as possible to data of existing buildings obtained from building energy surveys and audits to ensure the input parameters used in the Appendix A. simulation calculations would match with actual situations. The model can also account for the effectof variations in the HKDLC [8] is de�ig a n cooling load prediction program outdoor weather conditions and seawater temperatures on developed by the .Departrnen . t of Building Services Engi the efficiency of air-conditioning plants. neermg of The �ong Kong Po ec University. It was l t hnic Predictions obtained using this simplified model have devel ped acc rd11\g h y � � to t e CLTD/SHGF/CLF method of the been compared with detailed computer simulation results Amencan Society ,of H . Con- . . . . . eating, R · ,tion and Air• and electricity consumption data obtained from surveys d1t10 mg Eng 1 eer e ·fn· ger Heating, Refrigerating and Air-conditioning Engineers Inc., USA, [24] F.W.H. Yik, T.K. Chan, Air-conditioning system modelling in 1989. BECON, in: Proceedings of the CIBSE National Conference'98, [8] F.W.H. Yik, HKDLC - A Design Cooling Load Calculation Bournemouth, 18-20 October 1998, The Chartered Institution of Program for Buildings in Hong Kong, User Manual, Department of Building Services Engineers, UK, 1998, pp. 40-49.