Stepwise calibration for residential building thermal performance model using hourly heat consumption data

Abstract

Model calibration plays an important role in reducing the discrepancy between the simulation value and the measured data. The calibrated building thermal performance model will better guide the building retrofit and optimized operation. At present, however, researchers often calibrate building thermal performance model based on monthly data, and there are a few studies on hourly model calibration. In this paper, a stepwise calibration method for residential building thermal performance model based on hourly heat consumption data is proposed. This method mainly includes the following two key steps: (1) Sensitivity analysis. Morris sensitivity analysis method is adopted to reduce the number of calibrated parameters and improve the efficiency of model calibration. (2) Stepwise calibration. The building thermal parameters and the occupant behavior related parameters are distinguished, and proposing a calibration method for the schedule of occupant behavior related parameters. The effects of different processing methods for occupant behavior related parameters on the accuracy of calibration results are analyzed. A case study is demonstrated in a residential building with satisfying results. The Morris sensitivity analysis method reduces the number of calibrated parameters from 14 to 6. On the basis of the calibration for building thermal parameter, three calibration schemes for different occupant behavior related parameters are compared. Compared with before calibration, the MBE and CV(RMSE) of the best calibration scheme are reduced from −81.71% and 83.89% to −2.91% and 15.90%. The results meet the limits of model calibration accuracy in ASHRAE Guideline 14, IPMVP and FEMP standards.