1、2010 ASHRAE 57This paper is based on findings resulting from ASHRAE Research Project RP-1312.ABSTRACTThe validation process and result of a dynamic air han-dling unit (AHU) model (referred to as 1312 model hereafter)are discussed in this paper. The development of the 1312 modelis summarized in a com
2、panion paper. Strategies to validate themodel using experimental data mostly from common systemoperations are designed. If problems were identified usingsystem operation data, follow up component model calibrationis used to modify and improve the model. A series of experi-ments were designed and imp
3、lemented to obtain pressure resis-tance parameters for the supply duct system and mixing boxdampers. Building operation data from winter, summer, andspring seasons were used to validate the 1312 model. Goodagreements were achieved between experimental data andsimulation outputs for the 1312 AHU mode
4、l, especially forsummer and winter seasons. When using 1312 model to sim-ulate AHU operation for spring season conditions, simulatedoutdoor and supply airflow rates and supply air temperature,while tracking experimental data, showed certain level ofoscillation.INTRODUCTIONThe objectives of this stud
5、y are to develop and validate adynamic air handling unit (AHU) model. The developmentprocess of the dynamic system model (referred to as 1312model hereafter) that includes a single duct dual fan variableair volume (VAV) AHU system and four building zones servedby the VAV AHU system are introduced in
6、 the companionpaper (Li and Wen 2009). The 1312 model is developed inHVACSIM+ environment and is based on the E51 modeldeveloped from ASHRAE research project RP825 (Norfordand Haves 1997). The test facility used for this study is alsodescribed in Li and Wen (2009).In this paper, the validation strat
7、egy, process, and finalresults for 1312 model are described. Validation of HeatingVentilating, and Air Conditioning (HVAC) and building zonesimulation dynamic models is not a trivial issue. Detailedreview about simulation code verification and validation isprovided by Reddy et al. (2006). While veri
8、fication deals withdetermining whether the equations are solved correctly, vali-dation involves solving the right Equations and comparingsimulation results against field or experimental data.There are publications in the literature that discuss HVACsystem dynamic model verification and validation su
9、ch asthose focus on (a) component models (Clark et al. 1985, Zhouand Braun 2007); (b) primary systems (Henze et al. 1997,Wang et al. 2000); and (c) air conditioning process and itsinteraction with building zones (Brandemuehl et al. 1990,Ahmed et al. 1998).Two papers were found that specifically disc
10、ussed AHUdynamic model validation: Chen and Deng (2006) developed a dynamic simulationmodel for a direct expansion VAV air conditioning sys-tem consisted of a VAV air distribution subsystem and aDX refrigeration plant. AHU model was part of theoverall model. A test rig was developed for validatingth
11、e model. However, the validation process only in-cluded comparing model outputs and experimental dataunder an open loop step change of compressor speedDevelopment and Validation of a Dynamic Air Handling Unit Model, Part 2Shun Li Jin Wen, PhDStudent Member ASHRAE Associate Member ASHRAEXiaohui Zhou
12、Curtis J. KlaassenMember ASHRAE Member ASHRAEShun Li is a PHD student in the Department of Civil, Architectural, and Environmental Engineering, Drexel University, Philadelphia, PA. JinWen is an assistant professor of the Department of Civil, Architectural, and Environmental Engineering, Drexel Unive
13、rsity, Philadelphia, PA.Xiaohui Zhou is an assistant scientist and Curtis J. Klaassen is the manager in Iowa Energy Center Energy Resource Station.OR-10-008 (RP-1312) 2010, American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc. (www.ashrae.org). Published in ASHRAE Transacti
14、ons 2010, Vol. 116, Part 1. For personal use only. Additional reproduction, distribution, or transmission in either print or digital form is not permitted without ASHRAEs prior written permission. 58 ASHRAE Transactions(one speed adjustment). No real weather conditions orinternal loads was applied t
15、o the model.Nassif et al. (2008) developed a series of simplifiedcomponent models for an AHU, a VAV terminal unit,building zone, and their control systems. Real operationdata collected from the system control system were usedto obtain model parameters. Model outputs were com-pared with system measur
16、ements. However, the compo-nent models were not connected to each other and it wasunknown how well the entire system model would per-form if all component models were connected.The above literature review indicates that there is a lack ofa comprehensive validation study for AHU dynamic modelsthat co
17、mpares the entire system model predictions with realoperation data. In this study, three perspectives of the 1312AHU model are to be validated:1. Parameters: during the model development process, allparameters are obtained from either nominal designvalues or from manufacturer catalogs. Those values
18、oftendo not reflect the true parameters for a real system. Animportant part of the validation process is to first “tune/calibrate” the parameters in the simulation model fromsystem measurements; 2. component models: component models used in HVAC-SIM+ may not be able to simulate the test facility AHU
19、performance satisfactorily because HVACSIM+ compo-nent models generally represent new and ideal compo-nent behaviors; and 3. system performance: even after all component modelsperform satisfactorily, the system performance may stillnot be satisfactory due to error propagation and numericalcalculatio
20、n stability.STRATEGYThe key for a validation process is to separate differentcomponent dynamics and parameters from each other. Ideally,before a system level validation, experiments should bedesigned and executed first for each component in the AHUwithout other components involved. Dynamic operation
21、 datafrom these experiments can be used to validate componentmodels, including model structure and model parameters.However, because of the limited time and budget, detailedcomponent model validation is difficult to perform for a realsystem. System level validation using data collected fromnormal op
22、eration period is more realistic. In this paper, vali-dation strategies that utilize normal operation data and datafrom two easy-to-configure system tests are introduced.During a system level validation, if a component model wasfound to be unacceptable, experimental data specifically forthat compone
23、nt are then sought to modify the componentmodel.For system level validation, again, the process needs to bedesigned to separate as many subsystems as possible to iden-tify unacceptable subsystems, or components, or parameters.Considering available operation data at the test facility, thefollowing va
24、lidation process is designed:System Level Steady State ValidationDuring a steady state experiment, controllable variablesand parameters associated with the AHU remain unchanged.The simplest steady state operation which is routinelyexecuted at the test facility to self-examine the facility is theAir
25、Loop Operational Test (ALOT). During an ALOT, theAHU is operated at 100% recirculation mode, cooling andheating coil valves are at 100% closed position. The return fanis manually operated at its maximum speed, while the supplyfan is off. Temperatures in the zones are maintained at around70F (21.11C)
26、. Therefore, ALOT data can be used as a firstcheck to examine the simulation model, especially the airnetwork at maximum fan speed. Data from four ALOTs areprovided by the test facility, with each test lasting about 6hours.Another special test, Heating Coil Test (HCT), is alsoused to validate the AH
27、U model. In a heating coil test, theAHU is operated at 100% outdoor air mode with fans operatedat maximum speed, and heating coil valve is at 100% openposition. Used after ALOT, HCT can be used to examine theimpacts of heating coil at 100% open valve position andoutdoor air path at 100% open damper
28、position. One data setfor HCT that lasted for about 6 h is provided by the test facility.System Level Dynamic ValidationAn experiment, Normalization Test (NT) is used forsystem level dynamic validation. During a NT, the entire testsystem is operated as a real commercial building. The testfacility ha
29、s conducted NTs in different seasons: winter,summer, and transition seasons (spring or fall).If available, winter NTs with heating coil operationshould be used firstly to validate the model. When the heatingcoil is used during a winter season, the outdoor air damper ismaintained at a minimum positio
30、n during the occupied period.Furthermore, during a winter season, the fans are mostly oper-ated at a constant speed because the VAV dampers are typi-cally maintained at their minimum positions. Therefore, thedynamic behaviors of mixing box model and fan model do notimpact the simulation results exte
31、nsively. Winter NTs can beused to validate the simulation model with a focus on heatingcoil model, fan models at constant speed, and other relevantcontrol models.Summer NTs with cooling coil operation should be usedsecondly to validate the model. During the summer season,when the cooling coil is in
32、use, the outdoor air damper is eithermaintained at a minimum position when outdoor air condi-tions do not satisfy the economizer control conditions; or theoutdoor air damper is maintained at 100% open position whenoutdoor air conditions satisfy the economizer control condi-tions. Under either situat
33、ion, the dynamic behavior of mixingbox model does not impact the simulation results significantly. 2010, American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc. (www.ashrae.org). Published in ASHRAE Transactions 2010, Vol. 116, Part 1. For personal use only. Additional reprod
34、uction, distribution, or transmission in either print or digital form is not permitted without ASHRAEs prior written permission. ASHRAE Transactions 59Summer NTs can be used to focus on cooling coil dynamicmodel, fan dynamic model, and relevant control model vali-dation.Winter or summer NTs with eco
35、nomizer operation shouldbe used thirdly. During such tests, outside air temperaturesatisfies the economizer operation conditions and thereforethe coils are not operated. Mixing box dampers are used toregulate the supply air temperature. The fans are mostly oper-ated at a constant speed. The differen
36、ce between winter dataand summer data when economizer is used is that the supplyairflow rate is generally lower during the winter. Therefore,data from such tests can be used to validate the system with afocus on the mixing box and relevant control models.Transition season NTs should be used finally.
37、 During atransition season NT, the coils are not operated for most of thetime. However, the mixing box and fans are often operatedacross a large range of operation conditions. Therefore, datafrom such tests can be used to validate the system model witha focus on the mixing box and relevant control m
38、odels.During a NT, test settings are designed to be similar tothose used in a commercial building. For the AHU settings,1. The system operation is scheduled to be occupiedbetween 6:00 to 18:00 and unoccupied otherwise. Theentire system is turned off during the unoccupied period.2. The minimum outdoo
39、r air damper position is 40%(summer or transition season) or 47% open (winter).3. The economizer control is enabled when outdoor airtemperature is less than 65F (18.33C) (for winter andsummer tests). For the transition season case used in thisstudy, the economizer control is enabled when outdoor air
40、temperature is 30F (1.67C) lower than the return airtemperature.4. The supply air static pressure set point is at 1.4 in.w.g.(348.34 pa).5. The supply air temperature set point is 55F (12.78C) forsummer and transition seasons and is 65F (18.33C) forwinter season.For a zone,1. The zone heating set po
41、int is 70F (21.11C) during theoccupied time period. The zone cooling set point is 72F(22.22C) during the occupied time period.2. The maximum VAV airflow rate is 1000 cfm (0.472 m3/s)for exterior zones and 450 cfm (0.212 m3/s) for interiorzones. The minimum VAV airflow rate is 450 cfm (0.212m3/s) for
42、 exterior zones and 250 cfm (0.118 m3/s) forinterior zones.3. There are no window coverings.Table 1 summarizes some key settings for normalizationtests.Component Model CalibrationDuring the steady state and dynamic system validationprocesses, simulation models for several components, such asfan, coi
43、l, damper, and valve, did not provide accurate predic-tions of the real component performance. Data from specialexperiments, which only focus on the dynamics of a specificcomponent, are sought to modify existing component modelor to calibrate model parameters. This process is referred to asa calibra
44、tion process. The fan and valve models calibrationprocesses are introduced in the companion paper (Li and Wen2009). The duct and damper model calibration process isintroduced in the section on Pressure Resistance for Damperand Duct Models of this paper. Figure 1 demonstrates the over-all validation
45、process and the interaction between validationand calibration processes. More detailed discussion about thevalidation process and its interaction with the calibrationprocess is provided in the section on Validation Process.AHU Model SeparationThe scope of project ASHRAE RP-1312 only includesvalidati
46、ng the AHU dynamic model. The validation of build-ing zone model and VAV terminal unit models, which are alsoparts of the 1312 model, are out of the scope for this project.Therefore, during the validation process, it is desired to sepa-rate the AHU model from the building zone and VAV terminalunit m
47、odels. All models are inter-connected in HVACSIM+code (Norford and Haves 1997). And HVACSIM+ solves alldifferential equations (including those for AHU, buildingzone, and VAV terminal units) together using an iterationprocess (Park et al. 1985). Building zone and VAV terminalunit models affect the AH
48、U model simulation by providingvalues of the return air temperature and supply air pressure. Toseparate AHU model, outputs from building zone and VAVterminal units are replaced with experimental data during theFigure 1 Flow chart for validation and calibration pro-cesses used for 1312 model. 2010, A
49、merican Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc. (www.ashrae.org). Published in ASHRAE Transactions 2010, Vol. 116, Part 1. For personal use only. Additional reproduction, distribution, or transmission in either print or digital form is not permitted without ASHRAEs prior written permission. 60 ASHRAE TransactionsAHU model validation process, i.e., return air temperature andsupply air pressure.ANALYSISTo compare simulation results with experimental data,performance indexes need to be defined. Two types of perfor-mance indexes, namely energy i
