Multi-objective optimization of traditional residential timber frames based on response surface methodology

Abstract

Traditional residential timber frames, embodying centuries of Chinese craftsmanship, continue to thrive. In today’s era of increasingly scarce global resources, optimizing these timber frames not only preserves their legacy but also contributes to structural weight reduction, aligning with the goal of lightweight construction. To determine the optimal design dimensions for these frames, this study introduces a combined optimization approach using Response Surface Methodology (RSM) and Multi-Objective Genetic Algorithm (MOGA). Using the Optimal Space-Filling (OSF) method, sample points within the design domain were collected to establish a response surface model, correlating frame dimensions with key structural metrics: maximum deformation, maximum Mises stress, and total mass. The model’s fitting accuracy was also evaluated. MOGA was employed to optimize the response surface model, yielding Pareto-optimal solutions. Additionally, a parametric sensitivity analysis was conducted on the frame based on the response surface model. Simulation experiments demonstrate that the optimal span-to-depth ratio of the frame is 0.23. Compared to the initial frame structure, the optimized frame exhibited a 10.09% reduction in mass, a 27.47% decrease in maximum deformation, and a 2.46% reduction in maximum Mises stress. The findings of this study provide valuable insights for related engineering applications and a calculation model for the fatigue damage degree of concrete was proposed.