Abstract
Hourly dynamic energy performance study of buildings requires an in-depth understanding of dynamic thermal performance of building envelope assemblies. While two and three-dimensional building envelope thermal bridges have a great impact on whole-building energy simulations, heat storage capacity of the layers has also a significant influence. State of the art research has confirmed necessity of accurate thermal storage behavior analysis of building envelope assemblies in dynamic hourly building energy simulations. To-date, a number of studies have been conducted on the simplification of transient behaviour prediction of one, two and three-dimensional building envelope assemblies. nIn this study, the previous equivalent and simplified models for prediction of dynamic behaviour of building enclosure are reviewed, and an improved equivalent model based on frequency responses of RC-Network (FR-RCN) is presented. The model utilizes thermal RC-Network with three unknown resistances, two known resistances, and four unknown capacitances. The frequency responses of building envelope assembly are calculated either analytically (one dimensional assemblies), or numerically using COMSOL (two/three dimensional assemblies). Eureqa, a software which leverages evolutionary algorithms, is utilized in order to generate optimized unknown RC-Network resistances and capacitances considering the calculated frequency responses of the assembly. In this study, one light weight single-family home, one mass type structure high-rise building, and selected steel construction assemblies in climate zones 2 and 6 have been considered. A simple approach is also presented for the generation of equivalent FR-RCN models of variable insulation thickness assemblies. The comparison between the transient results calculated from the equivalent FR-RCN and COMSOL simulation shows good agreement. The performance of FR-RCN method is compared with other selected equivalent models, and an improvement in accuracy is confirmed.