Evaluation of simulation models for predicting the energy performance of aerogel glazing system

Abstract

Simulation model is crucial for predicting the energy performance of building components, as it is closely related to the energy-saving evaluation and product promotion of various innovative technologies. As an emerging technology, the aerogel glazing system (AGS) is distinguished from other conventional glazing systems due to its nano-porous structure and scattering characteristics. However, the accuracy of different models for predicting the energy performance of AGS has not been exhaustively quantitatively investigated. In this study, based on the detailed dynamic heat transfer and solar radiation transmission calculation model (the detailed model) and the overall heat transfer coefficient K and the shading coefficient SC model (the K-SC model), the intermediate heat transfer and solar radiation transmission calculation model (the intermediate model) of AGS is proposed. The objective of the present study is to summarize the model features and quantitatively evaluate the simulation accuracy and speed of the three models. The experimental validation was conducted based on a test cell with AGS. The heat gain simulated by the three models under different weathers, seasons, and orientations was comprehensively analyzed. The results showed that the K-SC model has the largest error, especially on sunny days. The normalized mean bias error (NMBE) and the coefficient of variance of the root mean square error [CV(RMSE)] of the transmitted solar gain calculated by the K-SC model is −55.14% and 80.99%, respectively. Compared with the detailed model, the K-SC model may underestimate the energy-saving potential of AGS by at least 25.12% in the cooling season and overestimate the energy-saving potential by at least 73.75% in the heating season. The intermediate model has better accuracy than the K-SC model by 11.43–21.69% in the cooling season and at least 23.21% in the heating season. Meanwhile, the simulation speed of the intermediate model is faster than the detailed model by two orders of magnitude.