Evolutionary Algorithm Based Approach of Envelope Optimization for Large-scale Space

Building envelope optimization always occupies an important place in sustainable building design. This paper selects a large-scale space as research object, which requires more complex and flexible fenestration system. On the premise of complying with some rules, the fenestration is easily managed by adjusting the number and position of small rectangular patches on the grid-based exposed surfaces. The co-simulation of EnergyPlus and Daysim and the evolutionary algorithm named SPEA2 are combined on the Grasshopper platform, aiming to find the optimal trade-offs between minimizing total energy consumption and envelope capital cost and maximizing UDI100-2000. A three-dimensional Pareto front is obtained, and its quality including convergence, spread and uniformity are qualitatively and quantitatively evaluated. This platform is proved to be an effective tool at the initial stage of design for sustainable building.