Generative design and performance optimization of residential buildings based on parametric algorithm

Abstract

As passive green design and performance optimization are very important in the early design stage of green residences to reduce building energy consumption, energy-efficient green design and artificial intelligence technology are combined in this work for the design of residential buildings. A parametric generative algorithm is developed to automatically generate design schemes of typical Chinese urban residences based on performance-oriented design flow. By summarizing the workflow of architects, the algorithm based on the technical route is as follows: 1) spatial form features extraction of residence database; 2) automatic generation and energy simulation of new design schemes; 3) evaluation and screening of generated schemes. The generative algorithm is formulated with Rhino/Grasshopper and Python. Via a residence design case in Beijing, the design scheme with the lowest cooling and heating load among 1,595 automatically generated schemes is deemed as the optimal scheme. The total load of the optimal scheme is 15.8% lower than that of the worst scheme and 4.2% lower than that of the original scheme. Therefore, the parametric generative design of residences is able to facilitate passive green design in the early design stage and enhance energy efficiency without the increase of construction costs.