| Title |
Bayesian Calibration of Energy Model for Existing Buildings |
| Authors |
Yoon, Seong-Hwan ; Kim, Young-Jin ; Park, Cheol-Soo |
| Keywords |
Bayesian calibration ; Markov chain Monte Carlo ; model calibration ; existing building ; ISO 13790 |
| Abstract |
Energy consumption by existing buildings accounts for a significant portion of domestic energy use. To improve energy performance of existing buildings, it is necessary to have a first principle based energy model for building energy performance assessment as well as rational decision making for energy retrofit. Over the past several decades, building energy calculation theories and applications have been significantly developed. However, such high-end building energy simulation tools require demanding time, cost and modeling efforts. In addition, as simulation tools become sophisticated and complicated, uncertainty caused by subjective judgment, modeling assumptions and stochastic building behavior is also not negligible, followed by a so-called 'performance gap' between prediction and a reality. To solve the aforementioned issues, the authors present application of Bayesian calibration, a stochastic parameter estimation technique to ISO 13790 model for an existing office buildings. In the paper, it is addressed that such simple energy model (ISO 13790) can produce accurate prediction when enhanced with Bayesian calibration and the calibrated model can be beneficially used for energy retrofit decision making. |