建筑结构学报2025，Vol.46Issue(6)：153-164,12.DOI:10.14006/j.jzjgxb.2024.0224
基于智能算法的自复位结构抗震韧性设计方法
Intelligent algorithm-based seismic resilience design method of self-centering structures
摘要
Abstract
Self-centering structures have more design parameters and targets than traditional structures,and the design difficulty is higher.Using traditional design methods with single target to design self-centering structures is not only incompatible with multi-design targets,but likely to increase the build cost due to the conservative design results.Hence,multiple design parameters and targets of self-centering structures are considered,and a structural intelligent seismic resilience design method based on the flexural-shear coupling model and shape function is proposed.This method utilizes a reinforcement learning model that cooperates with the multi-objective optimization to determine the optimal parameters of the flexural-shear coupling model,then determines the design parameters of each component via the parameter conversion of refined-simplified models,and accomplishes the seismic resilience design of self-centering structures.A 9-story self-centering braced steel frame is designed using the intelligent seismic resilience design method,and compared with its preliminary structure to verify the design validity.It is found that compared with the preliminary structure,the designed structure has 18％ to 43％ decrement of column section area,up to 48％ decrement of forces of the self-centering braces,37％ reduction of the floor acceleration,lower residual displacement drifts than the limit,no acceleration amplification effect,and slighter higher-mode participation.关键词
自复位结构/抗震韧性设计/智能算法/强化学习/多目标优化Key words
self-centering structure/seismic resilience design/intelligent algorithm/reinforcement learning/multi-objective optimization分类
建筑与水利引用本文复制引用
林智超,徐龙河,谢行思,张格..基于智能算法的自复位结构抗震韧性设计方法[J].建筑结构学报,2025,46(6):153-164,12.基金项目
国家自然科学基金项目(52125804,52494964). （52125804,52494964）
