CALIBRATION OF A BUILDING ENERGY PERFORMANCE SIMULATION .

2y ago
57 Views
8 Downloads
1.74 MB
8 Pages
Last View : 3d ago
Last Download : 2m ago
Upload by : Laura Ramon
Transcription

2018 Building Performance Analysis Conference andSimBuild co-organized by ASHRAE and IBPSA-USAChicago, ILSeptember 26-28, 2018CALIBRATION OF A BUILDING ENERGY PERFORMANCE SIMULATION MODELVIA MONITORING DATABasak GucyeterDepartment of Architecture, Faculty of Architecture and Engineering, EskisehirOsmangazi University, Eskisehir, TURKEYABSTRACTEnergy performance gap is considered as one of the mostsignificant issues associated with the assessment ofenergy consumption in the built environment. In order tonarrow this gap, simulation approach for energyperformance assessment are requires comprehensivecalibration procedures. Calibrated energy performancemodels facilitate a baseline representation of existingbuilding performance patterns, thus, further accuracy indiagnosis, operation and energy conservation measures(ECMs) become possible through the use of calibratedmodels. The present study presents an iterative approachfor calibrating a building energy model using full yearmonitoring data. The methodology focuses on disclosingthe steps in calibrating the simulation model and therelative sensitivities of the assumed and monitoredparameters used in calibration. The magnitude of thealteration in different levels of calibrated simulationmodels are evaluated with Mean Bias Error (MBE) andRoot Mean Square Error (RMSE) and the modelaccuracy is controlled through benchmarks defined byASHRAE Guideline 14, International PerformanceMeasurement and Verification Protocol (IPMVP) andFederal Energy Management Program (FEMP).INTRODUCTIONGiven the fact that 40% of the world energy consumptionoriginates from energy use in buildings for spaceconditioning, ventilation, hot water, lighting andappliances (DoE 2008), providing environmentallysensitive and efficient measures became a priority forresearchers and professionals involved in the productionof the built environment. Assessing building energyperformance, decreasing fossil fuel resourceconsumption and endorsing the utilization oftechnologies that support integration of non-renewableenergy sources became significant emphases (Fumo2014; Ahmad and Culp 2006). Furthermore, regulatoryapproaches encouraged the decrease in energyconsumption on building level through energyconservation measures (ECMs) for existing buildingenvelopes such as insulation, better-performing glazing,solar shading, etc. and integration of renewable or cleanenergy technologies within the building services (Henset al. 2010; Diakaki et al. 2010). In such a framework,evaluating the effects of ECMs became crucial inachieving decreased levels of energy consumption inbuildings. Simulation software, machine learning,compliance systems (De Wilde 2014) gained importancedue to their capability to replicate real world phenomena,and especially simulation tools were considered reliablewhen results were within error margins that were set viastandards such as International Measurement andVerification Protocol (IPMVP 2001), ASHRAEGuideline 14 (2002) and the Federal EnergyManagement Program Monitoring and VerificationGuide (FEMP 2008). Simulation modeling, the ology, was distinguished with its capability toreplicate the thermal behavior and energy performanceof a building (Crawley et al. 2008). Validation andtesting became of utmost importance to accurately assessthe realistic energy performance of the buildings.However, starting from the mid 1990s, strong indicationsof a “performance gap” were evident between thepredicted and actual energy consumption, and theexhibited discrepancies, in some cases, were more than100% (De Wilde 2014; Bordass et al. 2004; Menezes etal. 2012). Consequently, building energy performancegap turned out to be one of the widely-discussed issuesassociated with energy use in the built environment.Despite the national/international standards thatrecommend accurate assessment of building energyperformance, the discrepancy between the designpredictions and as-built energy performance of buildingswas still significant due to an array of reasons related tofactors affecting energy consumption (such as occupantbehavior, simulation model simplifications, poor542 2018 ASHRAE (www.ashrae.org) and IBPSA-USA (www.ibpsa.us).For personal use only. Additional reproduction, distribution, or transmission in either print or digital form is not permitted withoutASHRAE or IBPSA-USA's prior written permission.

assumptions etc.). In addition to efforts that facilitatepost-occupancy evaluation for buildings to bridge theperformance gap, simulation modeling was as welldesignated as an assessment methodology that requires acertain degree of confidence. Hence, to holisticallyaddress whole-building energy performance assessmentthrough the utilization of a simulation model, it becamesignificant to implement a calibrated building energysimulation approach. Although intended to function as adesign phase tool, building energy simulation (BES)models were developed into tools that allowed complexcalculation of the energy performance of existingbuildings mainly to evaluate the effects of ECMs(Royapoor and Roskilly 2015; Coakley et al. 2014). Theforward approach in modeling and simulation brieflyemphasizes the importance of acquiring (1) climate datafor the case building, (2) building design, (3)geographical data (location, orientation, obstructionsetc.), (3) construction data, (4) building installationcharacteristics, (5) building operations, occupancy andschedules (Harish and Kumar 2016), yet inadequacy inabovementioned data could result in a discrepancybetween the simulation results and actual thermalbehavior of the building. Ahmad and Culp (2006)established that uncalibrated simulation models producediscrepancies between the monitored and calculatedconsumption levels in the range of 30% and suggestedthat the discrepancies even rise to a range of 90% forend uses such as chilled water, hot water, and electricityconsumption. Therefore, it is possible to assert thatemployment of uncalibrated simulation models is animportant factor in the emergence of buildingperformance gap and simulation models should becalibrated in order to decrease the effect of modelingerrors, insufficient inputs, imprecise assumptions, anduncertainty related to design and operation on thesimulation outcomes.Calibrating building energy models based on monitoringdata for existing buildings and from feedback data fromvarious field studies for new designs could facilitateperformance predictions with high accuracy (Raftery etal. 2011; Zhao and Magoulès 2012). In this framework,this study focuses on disclosing six distinct steps incalibrating the simulation model of an existing buildingthrough employment of monitored indoor temperatures,calculated/assumed infiltration rates, monitoredoccupant presence within the simulation model with aniterative approach. The outcomes are expected toprovide sensitivities of the assumed and monitoredparameters in calibrating building energy simulationmodels. The magnitude of the alteration in presentedcalibration steps are evaluated through Mean Bias Error(MBE) and Root Mean Square Error (RMSE) values andthe model accuracy is inspected with respect to thebenchmarks defined in ASHRAE Guideline 14 (2002),IPMVP (2001) and FEMP (2008). The present study,therefore, both underscores the significance ofcomprehensive calibration procedures in building energyperformance simulation and interprets the researchoutcomes in terms of their impact on buildingperformance gap.METHODOLOGYBuilding Information and the Monitoring ProcessThe case building, located in the main campus ofEskisehir Osmangazi University, Eskisehir, Turkey,predominantly accommodates office functions (Figure1). Further information on the building is presented inTable 1. Situated on a flat and open lot, the spaces areoriented towards a central corridor aligned to thenorth/south. The building has a reinforced concretestructure with filled in brick walls and no insulationdespite the cold/snowy climate in winters. Measuredthermal characteristics of opaque building envelopecomponents are presented in Table 2. Transparentenvelope parts of the building consist of aluminum andPVC frames without thermal break and double-paneclear glass with U-values of 3.0 W/m2K and 3.2 W/m2K,respectively (TS2164, 2000). The building isconditioned only with an old non-condensing boilerusing natural gas as the primary energy source. Indoortemperatures for office and classrooms were designed as23 ºC and 20 ºC for circulation spaces during the heatingseason. Approximate discrepancies of 1 to 3 C inindoor temperatures were observed during themonitoring of the building. The building is used foradministrative and teaching purposes between 8AM and5PM on workdays.Figure 1 (a) Typical floor plan of the case building (b)South façade of the case building543 2018 ASHRAE (www.ashrae.org) and IBPSA-USA (www.ibpsa.us).For personal use only. Additional reproduction, distribution, or transmission in either print or digital form is not permitted withoutASHRAE or IBPSA-USA's prior written permission.

Table 1 Building informationBuilding InformationFloor Area (m2)Floor Height (m)Volume (m3)Façade Surface Area (m2)Roof Area (m2)Glazing Area (m2)Glazing Ratio (%)Compactness (Atot/Vtot)34023.50 to 4.50132612678561666250.23Table 2 Building Elements and U-valuesBuilding ElementsReinforced Concrete WallsBasement Retaining WallsExterior WallsConcrete Floor on GroundFlat RoofSloped RoofRoof SlabInterior WallsInterior FloorsU-Value 66Indoor temperature and humidity, gas and electricityconsumption and weather data were measured in thebuilding during 2016. Electricity consumption wasmonitored with a power analyzer data logger on 10 minintervals. Hourly gas consumption for 2016 wasretrieved from the remote monitoring system of the gasprovider company. Outdoor temperature, outdoorhumidity, global horizontal solar radiation, wind speed,and wind direction were monitored with 10 min intervalwith a weather station. Cloudiness (0-1) data wasretrieved from the macro-climatic weather station inEskisehir, Turkey. Heating installation efficiency and Uvalue measurements for the opaque building envelopewere completed during the monitoring process (Table 3).Modeling and CalibrationThe calibration approach employed in the present studyintends to adjust simulation parameters iteratively, untilcertain degrees of accuracy between the monitored andthe simulated hourly indoor temperatures and themonitored and simulated monthly heating consumptionpatterns were achieved. The model accuracy iscontrolled through benchmarks provided by the IPMVP(2001), ASHRAE Guideline 14 (2002) and FEMP(2008). EDSL Tas was used for energy performancemodeling of the case building. A multi-zone simulationmodel was developed with respect to the spatialdivisions of the building, since the calibration of themodel would be conducted with hourly comparisons ofmonitored and simulated data for 37 zones. Figure 2presents the steps in the iterative process in calibratingthe energy simulation model. R01, the initial model,was created with basic information that was collectedthrough building audit including as built information,measured envelope characteristics throughthermocouple U-value measurements, monitored fullyear micro-climatic data, calendar and schedules foroccupancy, heating season design temperatures andheating installation properties.Table 3 Monitored building energy performanceparametersMonitored Building EnergyPerformance ParametersIndoor Temperature ( C)Indoor Relative Humidity (%)Occupant Presence (%)U-value (W/m2K)Gas consumption (m3/h)Electricity consumption (kWh)Outdoor temperature ( C)Outdoor relative humidity (%)Global horizontal solar radiation (W/m2)Wind speed (m/s)Wind direction ( )Cloudiness (0–1)Boiler Performance (CO2)MeasurementInterval10 min.10 min.1 min.Multiple1 h.10 min.10 min.3 h.OnceFigure 2 The iterative calibration process (Partlyadopted from Raftery, Keane and Costa 2009)Once the simulation outcomes were retrieved, hourlyindoor temperature results of the R01 simulation modelwere compared to the hourly monitoring data, inaddition to the comparison of monthly simulated andmonitored heating consumption data. The initial modelwas not expected to yield an acceptable accuracy;however, the results were extremely discrepant from the544 2018 ASHRAE (www.ashrae.org) and IBPSA-USA (www.ibpsa.us).For personal use only. Additional reproduction, distribution, or transmission in either print or digital form is not permitted withoutASHRAE or IBPSA-USA's prior written permission.

actual monitored indoor temperature and monthlyheating consumption data. Therefore, the calibrationprocess was initiated to match the simulation outcomeswith the monitored data as accurately as possible. Aspresented in Figure 2, the process in running iterativesimulation models, followed the procedure of obtainingrun results, comparing these results with monitoreddata, identifying the discrepancy and the possiblesource of discrepancy, adjusting relevant parameters,and running the next iterative model. This process wasrepeated until the model calibration was completed onthe 15th run.Figure 3 presents the integrations/adjustments in theabovementioned process of iterative runs. In the presentstudy, simulation outcomes of the 6 of 15 runs arecompared to monitored hourly indoor temperatures andmonthly heating energy consumption data and theresults of these comparisons are presented in detail todisclose the effect of the integration of monitoredindoor temperatures, occupant presence, and adjustmentof calculated/assumed infiltration rates within theenergy performance simulation of an existing building.(RMSE) and the mean bias error (MBE). Equations (1)and (2) present the formulas employed for RMSE andMBE, where, n is the number of observations, Tm,av. isthe average of the monitored data for n observations, Tsis the simulated data for n observations, and Tm is themonitored data for n observations.𝑅𝑀𝑆𝐸(%) )* ,-,/0.* 𝑀𝐵𝐸(%) (,-,/0.*;2 [ (𝑇8 𝑇: ) ] . ) 5 (,?@,-)(1)(2)5The results obtained with the linear correlation (R)analysis and the RMSE and MBE analyses were used toevaluate the accuracy of the iterative simulation runswith respect to the benchmark values provided by theIPMVP (2001), ASHRAE Guideline 14 (2002) andFEMP (2008). In Results and Discussion section, theiterative model characteristics, the nature ofintegrated/assumed parameters and the magnitude ofthe alterations in the outcomes due to the calibrationattempts are discussed in detail.RESULTS AND DISCUSSIONR01, created as the initial model with informationcollected through building audit and partly during themonitoring process, was setup with heating seasondesign temperatures with the intention to demonstratethe effect of monitored indoor temperatures on thesimulation outcomes. Simulation outcomes for R01 werecorrelated with the hourly monitoring data (37 spaces x8760 hours 324120 hours) and the comparison ofmonthly simulated and monitored heating consumptiondata was carried out using RMSE and MBE analyses.The indoor temperature errors between the monitoredand the simulated hourly data (Tm – Ts) were found to benormally distributed as presented in Figure 5. However,heating energy consumption comparison yielded adiscrepancy of -58.55% (MBE) (Table 5), which ishighly unacceptable when compared to the hourlycalibration benchmarks provided by ASHRAEGuideline 14, IPMVP and FEMP (Table 4).25Monitored and Simulated Temperatures for G0320Figure 3 Iterative runs and calibration attempts1050-500:00 12:0000:00 12:0000:00 12:0000:00 12:0000:00 12:0000:00 12:0000:00 12:0000:00 12:0000:00 12:0000:00 12:0000:00 12:0000:00 12:0000:00 12:0000:00 12:0000:00 12:00Temperature ( C)The comparison between monitored and simulated datawas carried out with two analyses. The first approach isa linear correlation (R) analysis based on hour-to-hourcorrespondence of simulated and monitored indoortemperatures for a full year, for each of the 37 zonesmonitored in 2016. Second approach is an erroranalysis that intends to check the deviation of simulatedhourly temperatures and monthly consumption patternsfrom the monitored data with root mean square error15-10Exterior Temperature ( C)HoursMonitored Temperature ( C)R01 Simulated Temperature ( C)Figure 4 Monitored and simulated temperatures for theunoccupied and unconditioned space G03(February 1st to 16th)545 2018 ASHRAE (www.ashrae.org) and IBPSA-USA (www.ibpsa.us).For personal use only. Additional reproduction, distribution, or transmission in either print or digital form is not permitted withoutASHRAE or IBPSA-USA's prior written permission.

Calibration BenchmarksASHRAE 14 (2002)IPMVP (2001)FEMP (2008)MBERMSEMBERMSEMBERMSECalibration TypeHourlyMonthly 10% 5%30%15% 20%10-20% 10% 5%30%15%R07 is presented as the next disclosed iterative run, sinceit could represent a level of mid-calibration with respectto the accuracy achieved for simulated indoortemperatures. Integrating monitored indoor temperaturedata in the simulation model resulted in improvedcorrelation and error values in comparison to theoutcomes of the initial run R01, with 7.32% RMSE,0.16% MBE, 1.62 C absolute average error (Eav) and0.91correlation coefficient (R) (Figure 5, Table 5). Inaddition, integrating monitored indoor temperature datain the simulation model resulted in a 38% improvementin annual heating energy consumption in comparison tothe initial run. However, R07 could not predict themonthly heating energy consumption with an accuracythat would meet the calibration benchmarks, rather theprediction was inaccurate with a MBE of -42.66%. Inorder to accept a simulation model as calibrated, bothindoor temperatures and consumption patterns should bewithin the acceptable calibration values presented inTable 4. Hence, the calibration process was continuedwith the integration of infiltration rates in the nextsimulation run, R08. Since blowerdoor tests could not becompleted during the monitoring period, the infiltrationrates were calculated based on the effective leakage areaand volume of each zone, and to the ATTMA standardTSL2 (2010) benchmark for normal levels of building airpermeability 0.7 m3/h.m2 @50Pa.Table 5 Case building calibration resultsRUNSSimulation model R01 was checked for modeling errorsand building envelope characteristics via indoortemperature comparison of unoccupied andunconditioned spaces, since such data could easilyreveal the errors in thermo-physical characteristics ofthe building envelope in simulation. Figure 4demonstrates the exterior, monitored, and simulatedtemperature fluctuations for ground floor entrancespace G03, which is an unoccupied and unconditionedspace. This evaluation indicated that building envelopecharacteristics were modeled with a certain level ofaccuracy since the average errors for eight unoccupiedand unconditioned spaces between the monitored andthe simulated hourly data (Tm– Ts) were found to bebetween 0.85 and 1.13 C. Such discrepancy did notnecessarily have to be the result of errors in the thermophysical characteristics of the building envelopeintegrated in simulation. Evaluation of the buildingenvelope characteristics would be more substantialconsequent to the calibration of other parameters thatmight be causing the model discrepancies. Identifiedsources of discrepancies for R01 were interpreted asfollows: (1) absence of realistic indoor temperatureprofiles for the heating season, (2) omitted infiltrationrates and (3) assumed occupant presence instead of theactual presence of occupants. In this respect, firstmonitored indoor temperature data was used tocalibrate the model. Indoor temperature profiles wereintegrated in the simulation model in seven runs andrequired changes to schedules, set point temperatures,calendar days and the heating system operationschedule.Table 4 Calibratio

Modeling and Calibration The calibration approach employed in the present study intends to adjust simulation parameters iteratively, until certain degrees of accuracy between the monitored and the simulated hourly indoor temperatures and the monitored and simulated monthly heating consumption patterns were achieved. The model accuracy is

Related Documents:

Calibration (from VIM3) Continued NOTE 1 A calibration may be expressed by a statement, calibration function, calibration diagram, calibration curve, or calibration table. In some cases, it may consist of an additive or multiplicative correction of the indication with associated measurement uncertainty. NOTE 2 Calibration should not be .

Host action Target needed * : Timings are given for information only, they can vary depending on the Host capabilities Calibration Data result Calibration data Calibration data Calibration data Calibration data Device initialization SPADs calibration Temperature calibration Offset calibrat

calibration. The vertical comparator was built during the year 2003. The calibration facility is designed to calibrate up to 3-m-long invar rods, both for system calibration of digital levels and for tradi-tional rod calibration. SLAC System Calibration Facility The procedure of system calibration of digital levels is described

1 Manual calibration introduction 3 1.1 Purpose 3 1.2 Manual calibration software 3 2 Manual calibration overview 3 3 Manual calibration process 4 3.1 Setup MCR program 4 3.2 Calibration pattern positioning 9 3.3 Capturing pattern image 14 3.4 Using the Synergy Calibration tool 18 3.5 Fine tuning side images 24

Calibration C-0942, Mechanical(Dimensional, Pressure, Mass & Volume) Calibration C-0943 & Thermal Calibration C-0944. Our company is also ISO 17025:2005 Certified for Mechanical Calibration, Elctro Technical Calibration & Thermal Calibration. The firm is proud to have earned distinction in its various services and we are performing as a

Quantify – Multivariate Calibration Overview Calibration Model Development Steps Load calibration spectra Merge all spectra into one data view Create a project (and a folder) Add labels with concentrations (Label Editor) Start the calibration wizard Create and optimize a calibration

Other calibration ranges available on request 6. Unit A B K P Z Calibration in MPa Calibration in bar Calibration in kgf/cm² Calibration in psi (Standard) Other calibration units available on request 8. Output signal C X 4 20 mA Current output signal Other Signals on request 9. Option 0 1 None (Standard) Accessories 4. Connection size

January 12, 2013 G450 / G460 Calibration Slide 15 Single sensor calibration procedure (part 5) To perform a span Calibration: Choose the sensor to be calibrated Choose "Calibrate" from the menu Make sure the calibration adapter, calibration gas and regulator are attached to the instrument