Evidence-based calibration of a building energy performance model with uncertainty and sensitivity analysis

• Autor(a) principal: Scortegagna, Elisa de Moura
• Data de Publicação: 2020
• Tipo de documento: Dissertação
• Idioma: eng
• Título da fonte: Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
• Texto Completo: http://hdl.handle.net/10773/28272
• topic: Evidence-based calibration process; Uncertanty analysis; Sensitivity analysis; Building energy performance model; Internal environmental quality

Dynamic simulation of buildings is an increasingly common practice in engineering and architecture. Although the simulators currently used are increasingly powerful, the complexity of buildings leads to the need for simplifications that can have a relevant impact on the quality of the results obtained. Several studies performed recently have emphasized considerable discrepancies between measured and simulated building energy performance. As buildings usually do not present the same performance during their operation as the one predicted in the design phase, a broad interest in building real-monitoring has been instigated and the gap between measured and simulated energy consumption data has thus become an elementary concern in the building simulation domain. For this reason, the calibration of building simulation models is of growing interest. The objective of this work is to use the data from the Indoor Environmental Quality (IEQ) measurements and the actual energy use measurements, electricity and heating consumption for one year of an in-use low-carbon newly built residential apartment in West London - UK, for the calibration and validation of an energy performance model. To achieve this, performances of some typical apartments were analyzed, and an energy model was created in DesignBuilder using design stage documentation, postoccupancy measurement and metering along with IEQ data from typical zones. A systematic, evidence-based methodology was used for calibrating one representative apartment energy model-based to monthly energy consumption data. The outcomes from the calibrated model energy simulation were compared with the actual measured energy data, then the causes of discrepancies between the two results were elaborate and the gap between these two performances was analysed to predict possible determinants. Uncertanti and Sensitivity Analysis (UA/SA) were conducted after the completion of the calibrated model in order to verify and quantify the degree of uncertainty for and the most influential and determinants variables in an energy performance model. The calibrated model created was validated by monthly calibration criteria as per IPMVP/ASHRAE Guideline 14, of CV(RMSE) <15% and NMBE<±5%. The energy performance remaining gap between the actual measurements and the calibrated model simulation results are than point out and explained. The work also reflects on practicalities of data collection such as shortcomings in the metering, monitoring and observations that could be addressed for model calibration more accurate. Some improvements in the limitations found in this work are recommended: more rigidity in the validation standards of calibrated models and existing methods for calibration; and the reduction of uncertainty in the model's input parameters.