微粒群算法在传感器优化配置中的应用  

Application of Particle Swarm Optimization to Optimal Sensor Placement

作　　者：林贤坤[1] 覃柏英[2] 

机构地区：[1]广西工学院广西汽车零部件与整车技术重点实验室,广西柳州545006 [2]广西工学院理学院,广西柳州545006

出　　处：《控制工程》2013年第1期84-88,92,共6页Control Engineering of China

基　　金：国家自然科学基金(50575101);广西自然科学基金(2012GXNSFAA053208);江苏省交通科学研究计划项目(06Y20);广西教育厅科研项目(200103YB105);广西工学院博士科研基金(院科博1005)

摘　　要：为了将动态测试中的传感器配置在合理的自由度,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MAC矩阵、Fisher信息阵,及其组合为优化准则,采用微粒群算法,对传感器优化配置问题进行了研究,探讨了优化准则和优化算法对传感器优化配置结果的影响。通过与模态动能法、有效独立法及基于QR分解的逐步累积法比较,传感器优化配置的结果表明微粒群算法优于上述方法。In order to dispose sensors to reasonable freedom degrees, and reflect adequately the dynamic characteristics of tested struc- ture, the sensor placements of dynamic testing must be optimized. Taking MAC matrix, Fisher information matrix, and their combina- tions as optimization criteria respectively, the particle swarm optimization was applied to optimal sensor placement problem. The effect of optimization criteria and optimal method to sensor placement were also discussed. Compared with three existing sensor placement opti- mal method, which are modal kinetic energy, effective independence, and cumulative method based on QR decomposition, their results of the optimal sensor placements indicated that the particle swarm optimization is better than three existing method.

关 键 词：微粒群算法 整数规划 传感器 优化准则 优化配置 

分 类 号：TP27[自动化与计算机技术—检测技术与自动化装置]