Seismic Design Optimization of Steel Structures Using Particle Swarm Algorithm

Abstract

Earthquakes are one of the most devastating and expensive natural disasters in the world. Economical and earthquake-resistant design remains a challenge for structural engineers. This study explores the optimal design of a seismic force resisting steel frame using a population based stochastic algorithm known as Particle Swarm Optimization (PSO). PSO is able to efficiently explore a complex solution space with many design variables and constraints. PSO is also problem independent and can be built around any approach to earthquake design. As a case study, the seismic design of a three-story moment resisting frame is optimized for the linear static, linear dynamic, and nonlinear static analysis methods. An interface was created between MATLAB and OpenSees to link optimization with a well-known and freely available earthquake engineering software. This application is extended to the performance-based design of structures, in which the optimal design meets the target performance objectives of Immediate Occupancy, Life Safety, and Collapse Prevention under Frequent, Design, and Maximum-considered seismic hazard levels.