Paper Conference

Proceedings of eSim 2014: 8th Conference of IBPSA-Canada


Calibrating Dynamic Building Energy Models using Regression Model and Bayesian Analysis in Building Retrofit Projects

Wei Tian, Qiang Wang, Jitian Song, Shen Wei

Abstract: In building retrofit analysis, the whole-building calibrated simulation approach can be used to calculate potential energy savings by developing the physical simulation model of a building. The most difficult part in this process is to calibrate simulation outputs of an energy model to measured energy data. A new methodology of combining statistical regression models and Bayesian analysis is introduced in this paper to infer the distributions of unknown inputs based on measured energy data and other available information on input data. This method consists of three steps: (1) create and run EnergyPlus models to construct an original data set; (2) construct full linear models (or other variations) to approximate the relationships between model inputs and simulated energy consumption; (3) perform Bayesian analysis to estimate unknown inputs based on regression models, available input information, and measured energy data. Two case studies indicate this method can provide fast and reliable inputs.
Pages: 151 - 163