Intelligent optimization: A novel framework to automatize multi-objective optimization of building daylighting and energy performances. (November 2021)

• Record Type: Journal Article
• Authors: Dong, Yuhan; Sun, Cheng; Han, Yunsong; Liu, Qianqian
• Is Part Of: Journal of building engineering. Volume 43(2021)
• Publication Date: 2021-11
• Subjects: Intelligent optimization -- Multi-objective optimization -- Daylighting and energy performances -- Adaptive cooperative adjustment modelling -- Automatic data interaction -- Decision-making; Building -- Periodicals; 690.05
• DOI: 10.1016/j.jobe.2021.102804
• Languages: English
• ISSNs: 2352-7102

Abstract

Abstract: The existing building energy-efficient design relies highly on repeated modelling and it is difficult to obtain an alternative with desired performances as the decision-making procedure is normally too subjective. To improve the accuracy and efficiency, we aim to propose the intelligent optimization method, a user-friendly framework to automatize the whole energy-efficient design process from building information modelling to decision-making. Parametric programming and interface program design are adopted to realize the function of adaptive cooperative adjustment modelling and real time automatic data interaction between different modules with design efficiency enhanced. A decision-making method based on Self-Organizing Map clustering integrated with NSGA-II optimization algorithm is proposed to increase the accuracy of energy-efficient design. An experiment is performed to assess the effectiveness from design accuracy and efficiency as well as applicability with evaluation factors put forward. The experiment is carried out on a typical building in severe cold regions by comparing with two commonly used methods of subjective decision-making design and simulation aided design. The results show that the proposed framework is applicable by architects and performs better in design accuracy and efficiency compared with the other two methods. The performance levels of the compared three objectives are improved by 25.23%, 9.84%, 3.87% respectively and 87.5% time of a day is saved for architects to do other work. Thus, the intelligent optimization framework is a promising method to automatize the multi-objective optimization process of daylighting and energy performances by architects with both design accuracy and efficiency enhanced. The effectiveness evaluation method of different frameworks using experiment will benefit the selection and development of more applicable tools used by architects in future. Highlights: A user-friendly framework is proposed to carry out energy-efficient design. Parametric programming and interface program design are utilized to improve design efficiency. Self-Organizing Map clustering and optimization algorithm are coupled to improve design accuracy. An experiment is performed to evaluate the effectiveness of the proposed framework. The evaluation factors of design accuracy and design efficiency are proposed.