MOOSAS – A systematic solution for multiple objective building performance optimization in the early design stage

Abstract

There is great potential for building performance optimization (BPO) in the early design stage, but there is still a lack of methods, algorithms, and tools to support the BPO process in this stage. Through a comprehensive review, this study identified three critical issues that affect the implementation of the BPO process in the early design stage: model integration, real-time performance analysis, and interactive optimization design. This study provides a systematic solution to these three critical issues. In terms of model integration, a feature-based and graph-based 3D building space recognition algorithm is proposed to automatically convert the computer-aided design (CAD model) into a computer-aided engineering model (CAE model). In terms of real-time performance analysis, a simplified physical method, an HPC-accelerated method, and an AI-based method are explored, and a real-time energy modeling module and a real-time daylighting analysis module are developed. In terms of the interactive optimization design, a preference-based multi-objective BPO design algorithm that can consider user preferences is proposed to make full use of the decision-making ability of humans and the computing power of machines and significantly improve the optimization efficiency and result satisfaction. Based on the systematic solution, a multi-objective BPO design software, MOOSAS, is developed. MOOSAS allows real-time performance feedback, dynamic parameter analysis, and interactive optimization, supporting the BPO process in the early design stage. The innovations of this study are as follows: first, this study proposes a systematic solution to the three critical issues of the BPO process, i.e., model integration, real-time performance analysis, and interactive optimization design; second, this study develops a multi-objective BPO design software (MOOSAS) for the early design stage.