Calibration of Building Performance Simulation (BPS) Models to Measured Data: Selection of Measurement Points and Measurands

Building performance simulation (BPS) models are increasingly being used throughout a whole building’s lifecycle. In order to ensure model reliability and accuracy, calibration is an integral part of the overall modeling process, providing the modeler with increased confidence. Although BPS models have hundreds of output parameters, typical calibration approaches use only one or two of these [1]. The calibration problem is thus generally over-parameterized and under-determined. Choice of measurement points, measurement sensors and corresponding accuracy requirements for model calibration has been given limited attention in the literature. Research highlights: We explore calibration of BPS models with a large number of input & output parameters. Most influential input & output parameters relevant for calibration are found in a screening process. Sensors with accompanying measurement uncertainty are associated with each output parameter. Accuracy of calibrated model is investigated as a function of sensor uncertainty & sensor configuration.