基于GANN的机翼盒段试验件传感器的优化配置  

GANN Method Applied to Optimize Piezoelectric Sensor Placement for Wing Box Specimen of Plane

作　　者：谢建宏[1] 张为公[2] 梁大开[3] 

机构地区：[1]江西财经大学电子学院,南昌330013 [2]东南大学仪器科学与工程系,南京210096 [3]南京航空航天大学智能材料与结构航空科技重点实验室,南京210016

出　　处：《系统仿真学报》2007年第9期1917-1921,共5页Journal of System Simulation

基　　金：国家自然科学基金(90205031)

摘　　要：对机翼盒段试验件进行了结构分析,采用有限单元方法,建立了其结构的有限元模型,并进行了冲击压电响应数值仿真。构造了一种基于损伤检测的压电智能结构传感器优化配置的遗传神经网络(GANN)方法,采用该方法对机翼盒段试验件压电传感器进行了优化配置,得到了传感器对应于其初始布置模式下的最优配置,为该结构试验件的实际压电传感器的优化配置提供指导依据。仿真结果也表明,对于更多传感器的初始布置模式,采用该遗传神经网络方法可有效减少更多传感器的数量,从而降低成本。Based on the finite element method, the wing box specimen of a plane affixed with piezoelectric sensors was simulated, and its piezoelectric responsive signals were obtained under the impact load. A method of genetic algorithm integrated neural network （GANN） used to optimize sensor placement based on damage detection for piezoelectric smart structures was proposed. Then, the method of GANN was applied to determine the optimum piezoelectric sensor placement corresponding to its primal sensor placement for the wing box specimen. The simulation results can give a certain of guidance for the practical piezoelectric sensor placement for the wing box specimen. In addition, the simulation results show that, for the more sensors＇ primal placement, the number of sensors can be reduced effectively＂through the method of GANN, and thus leads to cost savings.

关 键 词：机翼盒段试验件 损伤检测 传感器优化配置 遗传神经网络方法 

分 类 号：TB381[一般工业技术—材料科学与工程] TP39[自动化与计算机技术—计算机应用技术]