Paper Conference

Proceedings of eSim 2018: 10th Conference of IBPSA-Canada


Data-Driven Modeling for Thermal Dynamic Analysis of a Low Energy House

Zequn Wang, Yuxiang Chen

Abstract: RC-network based data-driven models can provide useful information in analyzing and evaluating the actual thermal dynamics of building systems. This paper developed two RC-network models (a simple model and a multi-zone model) to identify important thermal characteristics of a low energy house. The models are trained/validated with on-site monitored data. Both of them can capture thermal dynamics of the low energy house with satisfactory accuracies. The simplified model gives overall thermal properties of the house while the multi-zone model reveals details concerning different zones. They are used to identify the thermal resistance of the building envelope, thermal capacities of internal air and internal mass, and the effective solar heat gains. Identifiability and interrelation of the estimated model parameters are also investigated. The models can be used to evaluate and operate the constructed building and the evaluations can inform future designs.
Keywords: data-driven models, RC-network models, building thermal characteristics, thermal dynamics
Pages: 26 - 33