改进的布谷鸟算法在结构损伤识别中的应用

Structural damage identification based on the modified Cuckoo search algorithm

布谷鸟算法受自然界中布谷鸟寻找鸟窝产卵过程启发而提出,其参数设置简单且效率高。为了平衡布谷鸟算法的求解速度和精度,提高算法的全局搜索能力,本文应用动态步长和锦标赛选择策略对原算法进行性能改进,并将其应用于结构损伤识别。在数值模拟中,选取频率等系统参数建立用于结构损伤识别的目标函数。基于改进的算法求解目标函数。最后分析得出结构损伤情况。数值结果显示,改进后算法准确识别损伤单元,并对噪声不敏感;验证了其有效性和实用性,可进一步应用于工程。

An approach based on modified Cuckoo search algorithm is proposed for structural damage detection. The tournament selection strategy and the dynamic step are applied to modify the CS algorithm, which not only prevent the algorithm from being local optimal but also improve the computation accuracy. The natural frequency and the modal assurance criterion(MAC) are adopted to build up the objective function of the damage detection. The MCS is utilized to solve the certain objective function. The simulations show that the MCS algorithm has a significant improvement in the convergence speed and the global optimization ability compared with standard CS algorithm and GA algorithm. The MCS algorithm can judge the local damage location and degree of the structure accurately even under measurement noise, which illuminates that it has a higher damage diagnosis precision.