Approximate Bayesian Computation in Parameter Estimation of Building Energy Models
- 240 Downloads
Model calibration is a necessary step to create reliable energy models in building retrofit. Bayesian computation in model calibration has attracted more attention because it can make full use of prior knowledge on building parameters. However, the likelihood function is hard to be computed in Bayesian computation due to the complexity of building energy simulation models. Approximate Bayesian computation (ABC) is a likelihood-free method to infer unknown parameters in complicated computational models by approximating the likelihood function with simulation. The ABC method is inherently computationally intensive since a large number of simulation runs are required to find reliable inferred values. This paper proposes a method for combining the ABC technique and the machine-learning method to compute unknown parameters in parameter estimation of building energy models. The results show that this method can provide reliable estimations of unknown parameters when calibrating building energy models.
KeywordsBuilding energy simulation Approximate bayesian computation Parameter estimation Model calibration
This research was supported by the National Natural Science Foundation of China (No. 51778416) and the Key Projects of Philosophy and Social Sciences Research, Ministry of Education (China) (contract No. 16JZDH014, approval No. 16JZD014).
- 6.Bureau, C.M.: China Standard Weather Data for Analyzing Building Thermal Conditions. China Building Industry Publishing House Beijing, China (2005)Google Scholar
- 7.MOC.: Design Standard for Energy Efficiency of Public Buildings, China Architecture and Building Press (2015)Google Scholar
- 8.Kuhn, M., Johnson, K.: Applied Predictive Modeling, Springer (2013)Google Scholar
- 10.Guideline, A.: Guideline 14-2014, Measurement of Energy, Demand, and Water Savings (2014)Google Scholar
- 12.Sisson, S.A., Fan, Y., Beaumont, M.: Handbook of Approximate Bayesian Computation, Chapman and Hall/CRC (2018)Google Scholar