1、2010 ASHRAE 45This paper is based on findings resulting from ASHRAE Research Project RP-1312.ABSTRACTA dynamic air handling unit (AHU) simulation model thatis capable of producing operational data for commonly usedAHU configurations will assist further research in AHUcontrol and operation, as well a
2、s fault detection and diagnosis.In this study, dynamic behaviors of an AHU and four buildingzones that were served by the AHU were modeled using HVAC-SIM+ software developed by the National Institute of Stan-dards and Technology. The model (called 1312 modelhereafter) was developed based on two prev
3、ious ASHRAEprojects (RP-825 and RP-1194). However, significant modifi-cations, which included new parameters, new control strate-gies, and new component models, were made in this study todevelop the 1312 model. The 1312 model structure, modelparameter development process, and two new componentmodels
4、, namely a new coil valve model and a new fan energymodel, are introduced in this paper. The new coil valve modelconsiders nonlinear behaviors of a three way valve. The newfan energy model outputs fan energy consumption that includesenergy consumptions for fan, belt, motor and VFD. Coeffi-cients for
5、 the new fan energy model can directly be estimatedfrom the total fan energy measurement. The 1312 model wasalso validated using real building operation data obtainedfrom a large scale building laboratory facility. The validationprocess and results are introduced in a companion paper.INTRODUCTIONAn
6、air handling unit (AHU) connects primary heating andcooling plants with building zones, controls building ventila-tion intake, and greatly affects the energy consumed for heat-ing, cooling, and ventilating, as well as supply air temperatureand humidity levels. An AHUs operation significantlyimpacts
7、building energy use, health, and comfort aspects.Nevertheless, only limited experimental studies under restric-tive scopes were available to evaluate AHU automated faultdetection and diagnosis methods (Norford et al. 2000, Carling2002, Castro et al. 2003). A dynamic AHU simulation modelthat is capab
8、le of producing fault free and faulty operation datafor commonly used AHU configurations, and control andoperation strategies is thus needed. Moreover, developeddynamic AHU simulation models need to be properly vali-dated systematically with experimental data for both fault freeand faulty operation
9、before any credibility can be placed on theprediction accuracy and usefulness. Therefore, developingand validating a dynamic AHU simulation model that is capa-ble of producing fault free and faulty operation data are objec-tives for an ASHRAE research project (RP-1312). In this two-paper series, the
10、 fault free model development and validationprocesses are summarized. Fault model development and vali-dation process will be reported in the future.The objectives of this study are to: (1) identify the propersimulation program and develop a full scale dynamic AHUand building zone system model that
11、has the capability toproduce fault free operational data; (2) identify a proper testfacility and gather experimental and/or field data to validatethe developed system model; and (3) develop a validationstrategy and validate the developed AHU model (reported inLi et al. 2009).Development and Validati
12、on of a Dynamic Air Handling Unit Model, Part IShun Li Jin Wen, PhDStudent Member ASHRAE Associate Member ASHRAEShun Li is a PHD student in and Jin Wen is an assistant professor of the Department of Civil, Architectural, and Environmental Engineering,Drexel University, Philadelphia, PA.OR-10-007 (RP
13、-1312) 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 reproduction, distribution, or transmission in either print or digital form is not permitted witho
14、ut ASHRAEs prior written permission. 46 ASHRAE TransactionsSIMULATION MODEL STRUCTUREVarious building heating, ventilating, and air condition-ing (HVAC) simulation models have been developed duringthe past decade for different purposes (Reddy et al. 2006): (1)Simplified Spreadsheet Programs, such as
15、 BEST (Waltz2000); (2) Simplified System Simulation Method, such asSEAM and ASEAM (Knebel 1983, ASEAM 1991); (3) FixedSchematic Hourly Simulation Program, such as DOE-2(Winkelmann et al. 1993), and BLAST (BSL 1999); (4)Modular Variable Time-Step Simulation Program, such asTRNSYS (SEL 2000), SPARK (S
16、PARK 2003), ESP (Clarkeand McLean 1988), Energy Plus (Crawley et al. 2004),ASHRAE Primary and Secondary Toolkits (Bourdouxhe et al.1998, Brandemuehl 1993); and (5) Specialized SimulationProgram, such as HVACSIM+ (Park et al. 1985), GEMS(Shah, 2001), and other CFD programs (Broderick and Chen2001). D
17、etailed building and HVAC simulation model reviewscan also be found in Kusuda (1999 and 2001), Bourdouxhe etal. (1998), Shavit (1995), Ayres and Stamper (1995), and Yuilland Wray (1990). Based on available reviews, several simu-lation software can be used for dynamic AHU model devel-opment and are d
18、iscussed in further detail below.HVACSIM+ (Park et al. 1985) developed by NationalInstitute of Standards and Technology (NIST) uses a uniquehierarchical variable time step approach in which componentsare grouped into blocks and blocks into super-blocks. Theactual breakdown of the system is left to t
19、he user. Each super-block is an independent subsystem, whose time evolution isindependent of other super-blocks. The only exception is thebuilding envelope, which uses a fixed, user-specified timestep. The time step in a super-block is a variable, which is auto-matically and continuously adjusted by
20、 the program to main-tain numerical stability. HVACSIM+ is especially appropriatefor simulating secondary systems and control strategies, andhas been undergoing experimental validation and improve-ments for several years (Dexter et al. 1987).TRNSYS (SEL 2000) developed by the Solar Energy Lab-orator
21、y, University of Wisconsin Madison, uses a componentbased methodology in which: (1) a building is decomposed intocomponents, each of which is described by a FORTRAN sub-routine, (2) the user assembles the arbitrary system by linkingcomponent inputs and outputs and by assigning componentperformance p
22、arameters, and (3) the program solves the result-ing non-linear algebraic and differential equations to determinesystem response at each time step.SPARK (SPARK 2003), which is similar to a general dif-ferential/algebraic equation solver, is an object-oriented soft-ware system that can be used to sim
23、ulate physical systems thatare described by differential and algebraic equations. InSPARK, components and subsystems are modeled as objectsthat can be interconnected to specify the model of the entire sys-tem. Models are expressed as systems of interconnectedobjects, either created by the user or se
24、lected from a library. AnHVAC tool kit library is supplied with SPARK. An on-goingproject (Xu and Haves 2001) conducted by the LawrenceBerkeley National Laboratory extends the current SPARKHVAC library to include more equipment models, such asAHUs and chillers, as well as models related to control s
25、ystems.SIMBAD (SIMBAD 2004) is a family of HVAC systemtoolboxes developed for the MATLAB/SIMULINK environ-ment (MATLAB 2006). It provides 11 modules which simu-late building and zones, production and storage devices suchas boiler and storage tank, hydronic and airflow distribution,heat emission devi
26、ces such as zone terminal units, controldevices, and weather and occupancy profiles. EnergyPlus (Crawley et al. 2004) is a building energysimulation program developed from BLAST and DOE-2. Itincludes many innovative simulation capabilities such as timesteps of less than an hour, modular systems and
27、plant inte-grated with heat balance-based zone simulation, multizoneairflow, thermal comfort, and photovoltaic systems (Energy-Plus 2005).ASHRAE 825-RP (Norford and Haves, 1997) extendedthe ability of HVACSIM+ and TRNSYS in the followingareas: (1) new models such as those for controller, sensor, and
28、airflow path were developed; (2) component models of thebuilding fabric and mechanical equipment were enhanced; (3)a real building (building E51), including the AHU system, wassimulated and documented in detail to demonstrate the use ofthe models. This model is referred as the E51 model hereafter.Al
29、though the E51 model was not validated, it provided a goodframework and model structure for other system model devel-opment.Among several qualified HVAC dynamic programsdescribed above, HVACSIM+ was selected for this project forthe following reasons: (1) HVACSIM+ already has most of thecomponent mod
30、els needed for this project; (2) HVACSIM+has been widely used and several studies have pointed out itssuitability for fault modeling (Bushby et al. 2001, Dexter1995, Peitsman and Soethout 1997); and (3) HVACSIM+demonstrated less computational difficulties compared toTRNSYS (Norford and Haves 1997).
31、Furthermore, the E51model can serve as a starting point for this project.The overall software structure, referred to as the 1312model hereafter, is illustrated in Figure 1. It consists of a userinterface, simulation models (including AHU and buildingzone models), and an I/O interface. The user can i
32、nput systemparameters, select AHU configurations, provide weather infor-mation and zone interior loads, and select faults to be modeled.Figure 1 ASHRAE 1312 Simulation model structure. 2010, American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc. (www.ashrae.org). Published i
33、n 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. ASHRAE Transactions 47TEST FACILITYThe same test facility, which was used by two prev
34、iousstudies (Norford et al. 2000, Castro et al. 2003) was chosen forthis project.Details about the test facility were provided by Price andSmith (2003). The test facility was built to compare differentenergy efficiency measures and to record energy consumption.To perform side-by-side testing, the fa
35、cility is equipped withthree AHUs (Figure 2). AHU-1 serves the common areas ofthe building. The remaining AHUs serve two identical sets oftest systems (the A- and B-Test Systems). AHU-A and B areidentical, with each AHU serving four zones. Of the fourzones, three have external exposures and one sees
36、 only internalconditions. The A and B zones are mirror images. The zoneshave identical construction and identical exposures yieldingidentical external thermal loads. The simulation models forthis project are based on the A-Test System.Major components of the Test System AHUs, shown inFigure 3, are t
37、he supply air and return air fans; preheat, heat-ing, and cooling coils; heating and cooling control valves;recirculated air, exhaust air, and outdoor air (OA) dampers;and the ducts to transfer the air to and from the conditionedspaces. The preheat coil is not used in this project. Instrumen-tation
38、consisting of humidity, pressure, and temperaturesensors, airflow stations, and electric power meters are avail-able to monitor the operational characteristics of the AHUs.The temperatures in the zones are controlled by variable airvolume (VAV) units with hydronic reheat coils. The testsystems are c
39、ontrolled by a commercial energy managementand control system.MODEL DEVELOPMENTAHUCommon AHU configurations include: (1) single ductconstant air volume (CAV) AHU; (2) single duct VAV AHU;(3) dual duct CAV AHU; and (4) dual duct VAV AHU. Foreach type of AHU, many different component configurationsand
40、 operation strategies exist, such as: (a) Single or Dual Fan;(b) With or without pre-heating coil; (c) With or without heat-ing coil; (d) a minimum OA damper that can be controlledusing minimum position, minimum flow rate, mixed airtemperature, or CO2 strategies; (e) a Return fan that can becontroll
41、ed using speed tracking, flow rate tracking, or buildingpressure difference strategies; (f) using various supply airtemperature control strategies such as separated mixed airtemperature control and supply air temperature control, finitestate control using multi Proportional-Integral-Derivative(PID)
42、loops, or split range control using single PID loop; and(g) using various economizer control strategies.Obviously, there are numerous possible AHU configura-tions and operation strategies (far more than what can be studiedwithin this research project). Furthermore, none of the currentHVAC system sim
43、ulation software provides an easy manner toallow the users to freely configure the system including hard-ware and control/operation strategy configurations. Similarly,the 1312 model, restricted by the HVACSIM+ configurationcapability, can only provide limited AHU hardware and control/operation confi
44、guration choices for simulation.The test facility is equipped with a single duct dual fanVAV AHU system. The AHU can also be operated like a CAVsystem. Consequently, the 1312 AHU model hardware config-uration will be a single duct dual fan AHU system that is simi-lar to the actual AHU system at the
45、test facility. Either VAV orCAV operation mode can be selected with VAV operationmode as the default choice. Although the AHU componentconfiguration cannot be changed, the user is allowed to altercomponent parameters based on specific needs. However, it isunderstood that only the default AHU simulat
46、ion model whichhas parameters similar to those at the test facility can beconsidered as a validated model.Differences between E51 Model and 1312 ModelThe E51 simulation model developed in ASHRAE 825(Norford and Haves 1997) serves as the base model here. Thereare three major differences between the E
47、51 model and the1312 model, which intends to represent the test system: differ-ences in components, parameters, and control systems. Thecomponent differences between the two systems include the:(1) OA duct: the E51 AHU has a minimum OA damper system.Because of its popularity among commercial buildin
48、gs, singleOA duct AHU configuration is adopted for the AHU simula-tion model developed in this project; (2) Heating coil: the testfacility AHU contains a heating coil which is not present in theE51 AHU; and (3) Building Zone: the test facility AHU servesfour building zones while the E51 AHU serves s
49、ix buildingFigure 2 Energy resource station layout (Price and Smith2003). 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 reproduction, distribution, or transmission in either print or digital form is not permitted without ASHRAEs prior written permission. 48 ASHRAE Transactionszones. AHU and building zones in the E51 building are muchlarger than those at the test facility. Hence, values for nearly allparameters used in the simulation models
