基于整数编码遗传算法的传感器优化配置研究  

Optimal sensor placement based on integer-coded genetic algrithm

作　　者：覃柏英[1] 林贤坤[1,2] 张令弥[2] 郭勤涛[3] 

机构地区：[1]广西工学院,广西柳州545006 [2]南京航空航天大学振动工程研究所,南京210016 [3]南京航空航天大学机电学院,南京210016

出　　处：《振动与冲击》2011年第2期252-257,共6页Journal of Vibration and Shock

基　　金：国家自然科学基金资助项目(50575101);江苏省交通科学研究计划资助项目(06Y20);宜宾学院青年基金资助项目(QJ05-01);广西工学院博士基金资助项目(03081509)

摘　　要：动态测试中,为了将传感器配置在合理的自由度上,以便充分反映结构的动力特性,需对传感器进行优化配置。分别以模态置信度矩阵、Fisher信息阵和它们的组合为优化准则,采用整数编码遗传算法,探讨了单目标和多目标优化的传感器优化配置问题。通过与现有的模态动能法、有效独立法及基于QR分解的逐步累积法进行比较,传感器优化配置的结果表明整数编码遗传算法优于上述三种方法。In order to dispose sensors to resonable degrees of freedom,and reflect adequately dynamic characteristics of a tested structure,the sensor placements of dynamic testing must be optimized.Here,taking MAC matrix,Fisher information matrix,and their combinations as optimization criteria,respectively,the integer-coded genetic algorithm was applied in a single-objective or multi-objective optimal sensor placement problem.Compared with three existing optimal sensor placement methods,modal kinetic energy method,effective independence one,and cumulative one based on QR decomposition,the results indicated that the integer-coded genetic algorithm is better than three methods mentioned above.

关 键 词：传感器 优化配置 组合优化问题 整数编码 遗传算法 

分 类 号：O329[理学—一般力学与力学基础]