Develop a Multi-Objective Fast Optimization Framework on Improving Building Energy Efficiency And Indoor Environmental Quality For University Library Atrium

Abstract

Multi-objective optimization methods provide a valid support to performance-based design of buildings. However, building performance simulation with low efficiency is a significant barrier to the performance-driven building design. The aim of this paper is to develop a fast framework of multi-objective optimization coupled with machine learning algorithms, taking the university library atrium as an example. This paper extracted the prototypical form of university library atrium based on 44 library cases in Wuhan; Then a methodology verified with measured data for evaluating building performance was constructed, and the synergistic influence of spatial morphology parameters on the building energy efficiency(BEE) and indoor environmental quality(IEQ) was analyzed; Finally, a multi-objective fast optimization framework coupled with machine learning algorithms was used to achieve the optimal design of university library atrium. The results showed that the parameters that influence the building energy consumption, indoor thermal comfort, daylighting performance most were the Height-to-Width ratio (HWR), the skylight solar heat gain coefficient (SHGCs), and the sidewall window-to-wall ratio (SWWR), respectively, when considering the synergistic impact of atrium spatial morphological parameter. The machine learning models predicted performance 400 times faster than traditional performance simulations. And compared with the worst-performance scheme, the maximum optimization rate of building energy consumption, indoor thermal comfort, daylighting performance was 29.46%, 10.46%, and 65.56%, respectively. The multi-objective fast optimization framework could provide guidance for policy makers and architects to performance-based design in the early design stages of university library atrium and support a low-carbon, energy-efficient, healthy and comfortable campus environment.

Keywords: University Library Atrium, Spatial Morphology, Building energy efficiency, Indoor environmental quality, Machine Learning, Multi-Objective Optimization