Meta-models for building energy loads at an arbitrary location. (September 2019)

• Record Type: Journal Article
• Authors: Yong, Seok-Gil; Kim, Jinho; Cho, Jinkyun; Koo, Junemo
• Is Part Of: Journal of building engineering. Volume 25(2019)
• Publication Date: 2019-09
• Subjects: Meta-model -- Cooling and heating load -- Building energy simulation -- Design optimization; Building -- Periodicals; 690.05
• DOI: 10.1016/j.jobe.2019.100823
• Languages: English
• ISSNs: 2352-7102

Abstract

Abstract: Meta-model-based optimization eases the burden of large time and computational costs by replacing the entire simulation process (which requires solutions of transient transfer equations) with the evaluation of simple meta-model functions (such as polynomials) while preserving the accuracy of dynamic simulations. The objective of this study is to present the development of meta-models to predict the cooling and heating load change of a building with variations in building envelope design factors under arbitrary climate conditions. The meta-models were developed using the results of building energy simulations for eight representative cities of climate conditions in the U.S. and were validated by comparing the predictions with simulation results for other cities in the US and Europe. The comparison showed that good accuracy was achieved with the developed unified single model. Highlights: Meta models for loads were developed for arbitrary climate conditions. They were validated by applying to other US and European cities. Prediction error of meta models for local weather increases with distance. Prediction error for whole zone model stays constant regardless of distance. The variations of optimal building design were analyzed with the models.