基于GA优化人工免疫算法的结构故障诊断

Structural fault diagnosis based on artificial immune algorithm of GA optimization

针对大型结构的故障检测与分类问题,提出了一种基于GA进化机制的人工免疫算法.该算法将样本结构模式数据作为抗原刺激抗体集合,抗体集合经过选择、交叉、变异、构建最优抗体集合这一进化过程来提高记忆细胞质量,利用训练好的记忆细胞集合实现对实测数据的故障检测与分类.在Benchmark结构模型上的仿真实验结果表明,该算法能实现有效的故障模式识别,且提高了故障分类的成功率,引入了多父体交叉操作,扩大了算法的搜索范围,且能有效利用其他抗体的优良模式,克服了单纯人工免疫算法收敛速度慢的不足.

In order to solve the problem in the fault detection and classification of large-scale structures, an artificial immune algorithm based on GA evolutionism was proposed. The sample structure mode data were taken as antigen stimulation antibody set in the algorithm, and the quality of memory cell could be improved through the evolutionary process including the selection, crossover, variation and construction of optimal antibody set. The fault detection and classification of measured data were realized with the trained memory cell set. The results of simulating experiments based on the Benchmark structure model show that the proposed algorithm can achieve the effective fault mode recognition and improve the success rate of fault classification. The multi-parent crossover operation is introduced, the search scope of algorithm is expanded, the excellent mode from other individuals can be effectively used, and the low convergence efficiency of simple artificial immune algorithm can be overcome.