基于并行量子遗传神经网络的自诊断智能结构传感器的优化配置*

针对压电智能复合材料层板,基于损伤检测问题,采用最小二乘小波支持向量机(LS-WSVM)网络建立损伤检测目标函数,运用量子遗传算法对目标函数进行优化,并将LS-WSVM以并行方式与量子遗传算法相结合,从而构造并行量子遗传神经网络方法,实现对智能结构损伤检测传感器的优化配置。仿真结果表明,采用该方法实现的不同数目传感器的最优布置符合工程判断,综合考虑成本与效益的因素,该方法可确定传感器对应于其初始布置模式下的最优配置数目。对于更多传感器的初始布置模式,采用该方法可有效减少传感器的数量,从而降低成本。相比于传统遗传算法,该方法中量子遗传算法具有较好的寻优能力和收敛速度。

Optimal sensor placement in self-diagnostic smart structures using parallel QGA integrated LS-WSVMs

For piezoelectric smart composite laminated plates,this paper proposed LS-WSVM as a kind of neural network to establish the performance function of damage detection,and then applied QGA to optimize the performance function.To enhance the algorithm speed,combined LS-WSVMs adopted as parallel mode with QGA,that was,constructed a method of parallel QGA integrated LS-WSVMs to optimize sensor placement based on damage detection.Simulation results show that,the optimal placements of different number of sensors are in accordance with engineering judgments,and considering the cost-effective factor,the optimal sensors’ number corresponding to its primal sensors’ number can be determined through the method.For the more sensors’ primal placement,the number of sensors can be reduced effectively through the method,and thus leads to cost savings.Compared with TGA,QGA in the method possesses the better searching ability and the faster convergence speed.