ASHRAE LV-11-017-2011 Applications of a Simplified Model Calibration Procedure for Commonly Used HVAC Systems.pdf

1、2011 ASHRAE 835This paper is based on findings resulting from ASHRAE Research Project RP-1092.ABSTRACTA calibration procedure for simplified building energysimulation models for commonly used HVAC systems has beendeveloped through an ASHRAE-sponsored project (ASHRAERP-1092). The procedure is applied

2、 to five buildings. Thesefive buildings include a 44-story office building, a low-riseoffice building, a church, a university teaching building, anda university research building. Three of the five buildings arelocated in Omaha, Nebraska (cold climate). Two buildings arelocated in College Station, T

3、exas (hot and humid climate). Thispaper presents the calibration procedure and summary resultsof five case studies. A two-level calibration procedure providesa good approach for model calibration. The case studiesstrongly indicate that the simplified model calibration proce-dure developed can be use

4、d to accurately calculate long-termenergy consumption data using short-term field energymeasurement data for different types of buildings with differentsystems. The first-level calibration procedure is very importantand improves the model accuracy significantly. The averageabsolute value of NMBE dec

5、reased from 41% to 6%, andCV(RMSE) decreased from 63% to 26% for these buildingmodels after the first-level calibration. The second-level cali-bration procedure further reduced these values to 2.5% and22.4%, respectively. After calibration of the simulation usingfour weeks or less of measured hourly

6、 data, the average of theabsolute values of the prediction error for annual consumptionvalues was 2.5%, with a maximum error of 9% for the casestudies examined. General information on the building andHVAC systems are the most critical input parameters for thesimplified model calibration. Short-term

7、hourly chilled-waterconsumption and hot-water consumption are the most criticalenergy data for calibration. INTRODUCTIONWith the increased use of building energy simulation forevaluating the effectiveness of energy conservation retrofits,calibration of simulation programs to measured data has beenre

8、cognized as an important factor in substantiating how wella model represents a real building. This procedure for calibrating simplified simulationmodels of commonly used HVAC systems has been developedthrough an ASHRAE research project (ASHRAE RP-1092).The development of a step-by-step simplified mo

9、del calibra-tion procedure will allow building professionals to ascertainthe annual cooling and heating energy use of buildings withmultiple HVAC systems from short-term field measurements.This project validates the step-by-step procedure using fivecase study buildings with an existing simulation pr

10、ogramdeveloped using the ASHRAE simplified energy simulationprocedure (modified bin method and ASHRAE tool kits).These five buildings include a 44-story high-rise building, alow-rise office building, a church, a university teaching build-ing, and a university research building. Three of the five bui

11、ld-ings are in relatively cold Omaha, Nebraska, and the other twobuildings are in hot and humid College Station, Texas. Thispaper presents a detailed description of the calibration proce-dure and results of the five building case studies. It also pres-ents the conclusions of this project and discuss

12、es the methodsapplicability to various building types and systems and theimpact of data availability.Applications of a Simplified Model Calibration Procedure for Commonly Used HVAC SystemsGuopeng Liu, PhD, PE Mingsheng Liu, PhD, PEMember ASHRAE Member ASHRAEGuopeng Liu is a senior research engineer

13、at Pacific Northwest National Laboratory, Richland, WA. Mingsheng Liu is a professor in theDepartment of Architectural Engineering, University of Nebraska Lincoln, Omaha, NE.LV-11-017 (RP-1092)2011. American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc. (www.ashrae.org). Pub

14、lished in ASHRAE Transactions, Volume 117, 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.836 ASHRAE TransactionsPROJECT OBJECTIVESA calibration procedure suitable for use

15、 with simplifiedengineering simulation models for commonly used HVACsystems has been developed through an ASHRAE-sponsoredproject (ASHRAE RP-1092). The objectives of the projectwere to do the following:Develop a step-by-step simplified model calibration pro-cedure to allow building professionals to

16、project annualcooling and heating energy consumption of buildingswith multiple HVAC systems from short-term field mea-surement data.Validate the step-by-step procedure using five case studybuildings with an existing simulation program devel-oped using the ASHRAE simplified energy simulationprocedure

17、s (modified bin method and ASHRAE toolkits). These five buildings include a 44-story high-risebuilding, an office building, a church, a university teach-ing building, and a university research building. Threeof the five buildings are located in Omaha, Nebraska(cold climate). Two buildings are locate

18、d in College Sta-tion, Texas (hot and humid climate).The project also documents the scientific and engineeringfoundations of the simplified model calibration procedures.This paper presents the step-by-step calibration proceduredeveloped and illustrates its use with a detailed case study. PROCEDURETh

19、e procedure developed adopts the definitions of Clar-idge et al. (2003) for calibration signatures and characteristicsignatures. It builds on the procedure developed by Wei et al.(1998) but uses building-specific characteristic signaturesinstead of generalized characteristic signatures, and uses atw

20、o-level calibration procedure. It includes a much moredetailed specification of data collection and calibration proce-dures, and is shown to be suitable for calibration to muchshorter periods of measured data. The simplified model cali-bration procedure consists of three steps. In the descriptiontha

21、t follows, it is assumed that the procedure is applied to abuilding with cooling supplied by chilled water (CHW) andheating by hot water (HW). For other systems, simply treatCHW as cooling and HW as heating in the methodology.1. Information CollectionGeneral building information, occupancy schedule,

22、mechanical system information and control sequence, energyconsumption data, and bills are collected in the first step. 1.1 General building information: the total condi-tioned area, number of floors, and occupancy schedule1.2 Mechanical system information1.2.1 Primary system Source of heating and co

23、oling (remote centralplant or plant in building).If electricity consumption model calibration isneeded, the rated pump power, chiller and boilerefficiency, and chilled-water/hot-water tempera-ture control sequence should be collected.1.2.2 Secondary systemInformation for each AHU: type of area serve

24、d(e.g., office or classroom), the system type (seeappendix for details), one-line diagram, ratedfan power, areas for envelope, windows andfloor. Terminal boxes: the terminal box type(s) (con-stant airflow boxes or VAV boxes, electricalreheat or hot-water reheat), and the minimumairflow ratio, if ava

25、ilable. 1.3 Energy consumption data and weather dataThe hourly outdoor air temperature and energyconsumption data for CHW, HW, whole-build-ing electricity (WBE), and HVAC electricity, ifavailable. Monthly energy bills can be obtained from thebuilding owner or the energy supplier. Last 12 months util

26、ity bills and energy con-sumption data are recommended.2. Site Visit and Short-Term MeasurementIn this step, the information obtained from step 1 is veri-fied, and the short-term measurement data should be recorded.These measurements include one-time site measurements,usually made using hand-held in

27、struments, and short-termmeasurements of energy consumption data, for which instru-ments with data-logging capabilities are set up and left in placeto collect data for at least two weeks.2.1 Short-term measurement2.1.1 Measurement periodThe hourly energy consumption data (chilled water, hotwater, an

28、d electricity) are typically measured for about two tofour weeks. The period from end of March to the end of Aprilis recommended for most locations because sufficient heatingand cooling energy consumption data can be obtained for cali-bration. 2.1.2 Measurement method Local weather data (dry-bulb te

29、mperature anddew-point temperature or relative humidity)should be measured for the simulation period.Alternatively, National Weather Station (NWS)data from a weather station as close to the site aspossible may be used. When no NWS weatherdata for this site are available, the outdoor airdry-bulb temp

30、erature should be measured at aminimum.Electricity: measured by a true-power meter orobtained from the utility.Hourly energy consumption during the calibra-tion period should be measured. The short-term2011 ASHRAE 837in-situ measurements will include the entirebuilding level measurements. 2.1.3 Mete

31、r and sensorsThe chilled-water and hot-water temperatureswill be measured using temperature sensors,combined with portable data loggers. Site measurements of whole-building electricityconsumption will be measured using a truepower meter or obtained from the utility sup-plier.Dry-bulb temperature and

32、 relative humidityshould be measured using temperature andhumidity data logger.The water flow rate can be measured using anultrasonic flowmeter or the water flow meters inthe HVAC system. If the system only has one or two AHUs, the air-flow can be measured by a fan airflow station(see RP-1092 report

33、, Appendix A-1, for details).For detailed measurement procedures, refer to ASHRAEGuideline 14-2002, Measurement of Energy and DemandSavings. 2.1.4 Measurement resultsThe following data plots are made and examined.Outdoor air temperature: time series of outdoorair temperature (dry bulb and dew point)

34、.Cooling/heating/electricity energy consumptiontime series.Hourly energy consumption as a function of out-door air temperature for the measurement timeperiod.2.2 A site visit is recommended to verify the mechan-ical system and schedule information.2.2.1. Verify the information for each AHU: the area

35、it serves, the system type, one-line diagram, and rated fanpower.2.2.2. Terminal box: verify the terminal box type andthe minimum airflow ratio if available. 2.2.3. Verify the operation schedule from the buildingoperator. 3. Model Calibration3.1 Determine the initial simulation inputs for thebuildin

36、g energy simulation model. 3.1.1 Consolidate similar AHUsGroup the AHUs.Four basic types of AHUs are commonly con-sidered: dual-duct (DD), single-duct with termi-nal reheat (SDRH), single-duct cooling +heating (SDCH), and single-duct heating + cool-ing (SDHC). In each of group of basic systems,separ

37、ate the VAV and constant volume systems. Divide the building into two zones to obtain theinterior zone ratio.Typically, the building is divided into two zones:an exterior or perimeter zone, and an interior orcore zone. Therefore, the interior zone ratioshould be calculated after dividing the buildin

38、gspace into two zones.Consolidate all similar AHUs.Separate each group of AHUs identified above intosingle-zone or two-zone units. A single-zone unit has either aninterior zone or an exterior zone. The two-zone systems haveboth interior and exterior zones. Users can have potentially 16different kind

39、s of systems. Then summarize the informationfor each group: floor area, envelope area, airflow, etc. Summa-rize the information into tables based on the mechanical draw-ings, spot field measurements, and EMCS printout for eachAHU.3.1.2 Determine the initial simulation input values foreach AHU group.

40、For each consolidated AHU, selection of the proper initialparameter values can significantly reduce calibration effort,although modifications to these values are essential during thecalibration process. The RP-1092 report (Liu et al. 2006)contains guidelines for initial value selection.3.2 Develop t

41、he calibration signatures and character-istic signatures.3.2.1 Generation of calibration signaturesRun the simulation using the initial set of inputs.The residuals, the root mean square error(RMSE) and the mean bias error (MBE) are cal-culated for the initial simulation (see appendixfor details on R

42、MSE and MBE) as(1)where Residual = Simulated consumption Measured consumption. See the RP-1092 report, Appendix B-1, (Liuet al. 2006) for calibration signature examples.Plot the measured consumption data, simulatedresults, and residuals in the same chart as a func-tion of outside air dry-bulb temper

43、ature, and plotthe calibration signatures on the same or sepa-rate charts. 3.2.2 Develop the characteristic signatures for thecorresponding system type and climate.The characteristic signature is defined as a nor-malized plot of the difference between two dif-ferent sets of simulated energy consumpt

44、ionvalues as a function of outdoor air temperaturewhere the two simulations have identical inputparameters except for one that differs. The char-acteristic signatures for heating and cooling arethen defined asCalibrationsignature =ResidualMaximummeasuredenergy- 100%838 ASHRAE Transactions(2)where th

45、e change in energy consumption is thedifference between the simulated consumption(e.g., cooling) at a particular point in time and themaximum energy consumption is the maximumvalue of the cooling consumption in the simula-tion period, typically one year. See Liu et al.(2006), Appendix A-5, for detai

46、led instructionson how to create the characteristic signatures. The input parameters that are recommended forgenerating the characteristic signature include:internal heat gain, building envelope heat-trans-fer coefficients, outside air intake for interiorzone and exterior zone, solar radiation load,

47、 airinfiltration, cold and hot deck temperatures,space temperature, and maximum and minimumairflow rates. The calibration parameters thathave a significant influence on energy consump-tion are found to be the most sensitive inputparameters (Mottillo 1999), or are those inwhich the authors have frequ

48、ently seen errors.Building specific signatures, rather than generalsignatures, are more effectively used in calibra-tion, based on the authors experience. 3.3 Simplified model calibration using characteristicsignatures and 24-hour daily pattern. Based on the energy signatures and physical rules, a t

49、wo-level calibration method was developed. The first calibrationlevel focuses on the weather dependent parameters of themodel. The second level focuses on the time schedule depen-dence.3.3.1 The first calibration level compares themeasured and simulated energy consumption plotted in x-yscatter plots, where the outside air temperature is plotted onthe x-axis. Compare each cooling/heating calibration signa-ture with the characteristic signatures. In com-paring the cooling/heating calibration signaturewith the