Discussion on the Hybrid Models for Building Energy Consumption Modeling and Prediction

Abstract

Building Energy Consumption Modeling and Prediction (BECMP) gradually become more and more significant in architectural engineering construction process in dealing with energy efficiency, sustainability and environmental-friendly development goals. Based on recent research on BECMP, this paper mainly focuses on the hybrid model combining the physic-based model and the AI-driven model, to discuss improvement on simulation result accuracy, prediction on residence activity, elimination on performance gaps, and optimization on Artificial Intelligence (AI) algorithm selection in improving building performance simulation rather than the other two types of prediction models, following the recent specific application examples and evidence. It also discusses opposing development concepts of models in BECMP among scholars and the limitations of building performance prediction, including corporation of residence behavior models, low popularity of the topic and the high education requirements of users. Lastly, this paper highlights the importance of real-time data system monitor and management, suggests multidisciplinary collaboration, and appeals widespread attention on BECMP.