1、VOLUME 10, NUMBER 2 HVAC Walton 1993). These models represent the thermal zone with both surfacesand air nodes in a single network and present a single representation of the thermal zone to anHVAC component simulation. Researchers and engineers have long had the ability to formulatedetailed network
2、models of a thermal zone and solve them using a variety of software tools suchas SPARK (1997) or IDA (EQUA 2002). On the other hand, computational fluid dynamics (or CFD) has been used to model buildingroom airflow for nearly 30 years. Chen and van der Kooi (1988) and Negrao (1995) coupledCFD to a b
3、uilding load/energy simulation program and, later, Beausoleil-Morrison (2000)expanded these capabilities. Other coupling work between CFD and a load/energy programincludes those from Srebric et al. (2000) and Zhai et al. (2002). As pointed out by Srebric et al.(2000), a direct coupling of CFD with a
4、n energy simulation program for hourly simulation ofbuilding performance over a year is too demanding computationally. This investigation tries to systematically build up a framework that allows an easy combina-tion of different air models with load and energy models. Figure 1 diagrams the classific
5、ation ofroom air models used in this paper. Such models have been developed for more than 30 yearsand are plentiful. The goal of the framework is to allow using all such air models with theASHRAE toolkit (Pedersen et al. 2001). Although the terms nodal and zonal are used inter-changeably in the lite
6、rature, for the purposes of this study a distinction is made between them.The distinction is basically one of how strictly and how resolved the geometry of the controlvolumes is defined. A “nodal” model treats the building room air as an idealized network ofFigure 1. Classification of room air model
7、s. 2004. American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc. (www.ashrae.org). Published in HVAC Haghighat et al. 2001;Inard et al. 1996; Griffith and Chen 2003) and have been compared by Chen and Griffith (2002).FORMULATIONIn formulating a framework for combined room air
8、 and load models, it is desirable that theframework be simple and applicable to a variety of room air models. The starting point for thiseffort is the original heat balance model in the ASHRAE toolkit (Pedersen et al. 1997; Pedersenet al. 2001). The new formulation alters the heat balance model wher
9、ever its model equationsinclude variables for the zone air temperature. The air heat balance of the original model is con-sidered an aggregate assessment of the air systems change in enthalpy and is referred to as the“ -equation.” The air and surface domains are modeled separately. In addition to th
10、e single control volume for room air, additional subdivisions of this controlvolume are allowed for the purpose of modeling distributions of air temperature within theroom. The air in the room is assumed to be a collection of separate, essentially well-mixed con-trol volumes, where each is modeled a
11、s having the following:1. uniform state conditions such as temperature and pressure,2. constant properties such as density and specific heat,3. transparency to radiation, and4. uniform distributions of heat and mass transfer at each control volume boundary.In aggregate, the room air control volumes
12、must agree with an overall air system heat balance.Note that the assumptions for room air control volumes parallel those for surfaces:1. uniform surface temperatures2. uniform irradiation3. diffusely emitted radiation4. one-dimensional heat conduction withinThere are five distinct processes:1. outsi
13、de face heat balance2. wall conduction heat transfer3. inside face heat balance4. air system heat balance5. air convective heat transportEach inside face interacts with a specified control volume rather than all of them interactingwith a single air control volume. The term “inside face” refers to th
14、e inside face of an enclosuresurface (such as windows, walls, floor, ceiling) that faces the room air. The near-surface air isreferred to as the adjacent air control volume. The additional fifth heat transfer process accountsQsys 2004. American Society of Heating, Refrigerating and Air-Conditioning
15、Engineers, Inc. (www.ashrae.org). Published in HVAC however, the film coefficient,hc, and the effective air temperature,Ta, are perhaps deceptively simple and are discussed inmore detail below. The sign convention here is that positive surface convection heat transfer, indicates net heat flow from t
16、he surface to the air and therefore adds to the cooling load.Equation 5 is a mean relation for an individual surface, so values for Ta, Ts, and hcare averagesappropriate for the surface. In the event that the resolution of the air model is finer than the sur-face model, the data should be averaged s
17、o that they conform to the surface area. This averagingis necessary since the underlying surface is treated as one-dimensional. Spatial Location for Determining Adjacent Air TemperatureBuilding rooms are enclosures and not semi-infinite fluid regions. Considering that Ta is avariable and not a const
18、ant, a framework for coupling air models to load/energy routinesrequires clear specification of how values for are to be determined. This air temperature isalso known as the reference temperature for convective heat transfer calculations, but the term“reference temperature” is avoided here because i
19、t implies fixing the value. For the well-mixedmodel, one obvious selection is that Tashould match the one available value (considered a goodmodel of the average air temperature). With nodal models, each surface is associated with a par-ticular node, and the result for temperature at that node is use
20、d for Ta. (Although all surfacesneed to have a node associated with them, some nodes might be associated with interior controlvolumes and not directly affect surfaces.) For zonal and CFD air models with a grid of interiorair control volumes, the basic question is what distance scale to use when dete
21、rmining Ta. A dis-tance of 0.1 m (4 in.) into the air away from a building surfaces inside face is selected as anappropriate geometrical scale for a point at which to determine Ta. The sensitivity of tempera-ture results to this distance scale is presented below. This value is chosen in view of the
22、follow-ing points: Tai TaiTstatDB() ,=TaiTsetpointTai,+=QcQc,iAhc,iTs,iTa,i() .=QcTai 2004. American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc. (www.ashrae.org). Published in HVAC these were used for both the zonal modelfrom Rees (1998) and Rees and Haves (2001) and the m
23、omentum-zonal model (Griffith andChen 2003). For testing a cooling load calculation, we can imagine the simulation is of a variable air vol-ume system, and the model “finds” the flow rate to meet TSetpoint that was actually fixed in theexperiment. A value for TSetpointwas extracted from the measured
24、 air temperature data by inter-polating between measured air temperature locations to obtain a value at 1.1 m from the floor.The surfaces are modeled as resistance constructions with Lis values for a U-factor of 0.36(W/m2K) for all surfaces except the west wall, which had a U-factor 0.15 (W/m2K). Ta
25、ble 1shows the “outside” boundary conditions for testing the coupled air and loads models.Table 2 lists overall results for this case using the coupled air and load routines. Figure 5 com-pares the results using the indirect-coupling method. Agreement between predicted and mea-sured air temperatures
26、 is fairly good for all the air models, especially in the occupied zone. Thewell-mixed model performs adequately. In general we find that the coupled models have only asmall effect on the result for but do have a significant effect on the air system temperaturedifference (Tleaving Tsupply). The resu
27、lts indicate that (for this case) the Mundt model overpre-dicted the temperature at the outlet leading to a system flow rate that is too low. The Rees andTable 1. Coupled Air and Surface Model Boundary ConditionsValue UnitsQconv,s (estimated splits) 225 (W)Qrad,s (estimated splits) 75 (W)Tsupply18.0
28、 (C)TSetpoint23.8 (C)Outside face (“TG”), east, north, west, and floor 22.63 (C)Outside face air (“TB”), south and ceiling 19.9 (C)Qsys 2004. American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc. (www.ashrae.org). Published in HVAC it is essential to explore how to optimize
29、 the condensing temperature toimprove their efficiency. With the simulation program TRNSYS, a chiller model is developedand experimentally verified to ascertain all parameters within chiller components. An algo-rithm, which can be used to determine the number of staged condenser fans by a setpoint o
30、f con-densing temperature, is first introduced. The model with this algorithm serves to illustrate theextent of decrease in the condensing temperature, which has rarely been studied. When the set-point of condensing temperature varies linearly with outdoor temperature, maximum chillerefficiency is o
31、btained. The condensing temperature also becomes controllable, resulting in lessfluctuation in chiller efficiency in different operating conditions.INTRODUCTIONAir-cooled chillers, one of the common refrigeration machines, are required to provide cool-ing for the indoor environment in the subtropica
32、l climate of Hong Kong. These chillers tend tooperate even in the cool season, as the cooling load of indoor activities usually exceeds the con-duction and ventilation losses. In this circumstance, the operation of these chillers is energyintensive, constituting as much as 40% of the year-round elec
33、tricity consumption of a commer-cial building (Chan and Yu 2002). Chiller efficiency is of paramount concern for buildingenergy effectiveness and is usually expressed by kW/ton: chiller power in kW over coolingcapacity in tons of refrigeration (one ton of refrigeration is equal to 3.52 kW). It shoul
34、d be notedthat a lower value of kW/ton means a higher chiller efficiency. Chan and Yu (2001) have indi-cated that the efficiency of air-cooled chillers in existing chiller plants is quite low: it exceeds1.4 kW/ton most of the time and rises from 2.3 kW/ton when the chillers are operating at halfload
35、. There has been little investigation into how a change in controllable parameters influenceschiller efficiency. With regard to compressor staging and condenser fan cycling, we tend to bereluctant to reform their control algorithms in the microprocessor control. We also assume thatthermostatic expan
36、sion valvescontrol devices of refrigerant flowinside chillers require ahigh differential between the evaporating pressure and condensing pressure to function properly.As a result, most existing air-cooled chillers operate at a high condensing temperature of around50C based on an outdoor temperature
37、of 35C, irrespective of weather-load conditions. Thiskind of head pressure control hinders the decrease in compressor power. Under local weather,outdoor temperature is below 25C for half the year. There is considerable room to reduce thecondensing temperature to improve chiller efficiency.There has
38、been little research on how to modulate the condensing temperature. Smith andKing (1998) have carried out an experimental study on a reduction in the condensing tempera-ture for an air-cooled reciprocating chiller with a cooling capacity of 35 kW. Their experimentalK.T. Chan is an associate professo
39、r and F.W. Yu is a Ph.D. student in the Department of Building Services Engineering,The Hong Kong Polytechnic University, Hong Kong. 2004. American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc. (www.ashrae.org). Published in HVAC&R Research, Vol. 10, No. 2, April 2004. For p
40、ersonal use only. Additional reproduction, distribution, or transmission in either print or digital form is not permitted without ASHRAEs prior written permission.114 HVAC&R RESEARCHfindings illustrate that chiller power consumption can be reduced by 10% by driving the con-denser fan harder when the
41、 ambient temperature is below 25C. According to another experi-mental study by Roper (2000), electronic expansion valves with sophisticated control algorithmsare prerequisites to lower the condensing temperature. Chiller power can be saved by as much as20% by modulating condensing temperature instea
42、d of maintaining a fixed condensing pressure.However, the interaction between condensing temperature and outdoor temperature has rarelybeen studied. It remains to be learned how the condensing temperature interacts with the stagingof condenser fans and heat rejection airflow. Equipment and system mo
43、deling is the best way tocontend with the above uncertainties and to understand the parameters of chillers (Reddy andAnderson 2002). Many simulation models for chiller systems exist (e.g., Browne and Bansal1998, Gordon and Ng 1994, Khan and Zubair 1999, Koury et al. 2001, Leverenz and Ber-gan 1983,
44、Yuill and Wray 1990), but models characterized as air-cooled reciprocating chill-ers, which are widely installed in local chiller plants, are not readily available. An air-cooledreciprocating chiller model is developed here by using the simulation program TRNSYS (SEL2000). This model is experimental
45、ly verified and serves to simulate the operation of chillers indifferent loading conditions.REVIEW OF HEAD PRESSURE CONTROLFor air-cooled chillers, different numbers of condenser fans are staged to control the head(condensing) pressure and to achieve heat rejection. Heat rejection varies with outdoo
46、r tempera-ture and heat rejection airflow for a given chiller load. Under head pressure control, the controlaction is based on some settings of condensing temperature. All condenser fans are staged whenthe condensing temperature exceeds its maximum level of 51.6C. Condenser fans are consecu-tively t
47、urned on when the condensing temperature rises from 45C and turned off when it dropsto below 22.8C. This means that the condensing temperature will float between 22.8C and51.6C, depending on outdoor temperature. This range, in fact, may be narrowed down to 33Cto 51.6C. The lower limit of the condens
48、ing temperature of 33C is due to a situation wherethermostatic expansion valves require a minimum differential of 690 kPa for proper operation.CHILLER SIMULATIONModel Configuration and Basic AssumptionsThe features of the model chiller are the same as the experimental chiller. It uses the refriger-a
49、nt R-22 and has a nominal cooling capacity of 120 kW (35 tons of refrigeration). This chillerrepresents the basic module of medium chillers with a cooling capacity of 422 to 704 kW (120to 200 tons of refrigeration). For the shell-and-tube liquid evaporator, the evaporating tempera-ture is designed to be 3C. The temperature of chilled water supplied is set at 7C, with a tem-perature rise of 5.5C at full load. The flow of chilled water is maintained at 5.2 kg/s in alloperating conditions. There is one refrigeration circuit where four recipr
