基于改进遗传模拟退火算法的结构损伤识别研究

STRUCTURAL DAMAGE IDENTIFICATION BASED ON IMPROVED GENETIC SIMULATED ANNEALING ALGORITHM

针对遗传算法容易"早熟"、局部寻优能力较差等缺点,提出一种基于改进遗传模拟退火算法的结构损伤识别方法。先用节点的残余力向量来进行损伤定位,然后以节点的残余力向量构造目标函数,以单元刚度折减系数为设计变量,利用改进的遗传模拟退火算法进行损伤程度的识别。最后,对一个桁架结构模型进行数值模拟,结果表明,该算法能够准确识别结构的多处损伤,验证了该方法的有效性。

Due to shortcoming of genetic algorithm that it is often premature convergence and the poor effect of the local optimization, a structural damage identification method based on the improved genetic-stimulated annealing algorithm was proposed. Firstly, the nodal residual forces were employed to localize the damage. Then, the reducing factors of elements stiffnesses were used as the design variables to construct the optimization object function. The quantification of the structural damage were performed with improved genetic-stimulated annealing algorithm. Finally, a numerical damage identification example of a truss model was given. The results showed that the multi-damage of the truss can be identified well. The effectiveness of the proposed method was verified.