基于微粒群算法的桁架频率拓扑优化

Frequency Topology Optimization for Truss Structures Based on Particle Swarm Optimization Algorithm

研究桁架结构频率拓扑优化的微粒群算法。采用混合罚函数法分开处理结构固有频率约束和其他约束条件,既保证所有约束能够严格满足,又提高了微粒群算法的收敛速度。由模态识别系数判断出虚节点自由度产生的局部振动模态,排除其对应的频率,得到结构真实的固有频率。算例计算结果表明,无论是频率极值优化问题,还是具有频率约束的结构优化问题,联合使用微粒群算法和模态识别系数都可以很方便地获得桁架最优截面和拓扑构型。

Particle swarm optimization(PSO) algorithm is applied to the frequency topology optimization for plane truss structures.The frequency constraint and other constraints are separately treated with the combined penalty function strategy in order to satisfy all constraint conditions and improve the convergence rate of PSO.The mode-discriminating coefficient is used to get rid of the inaccurate natural frequency corresponding to the local vibration mode,which is brought by the virtual node freedom.Through the computation and analyses of the numerical examples,it is found that there are well convenience and accuracy for the combination use of PSO and the mode-discriminating coefficient,no matter for solving the basic frequency extremum,or for the structural optimization with complicate frequency constraints.