基于RBF神经网络的复合材料层合壳荷载识别  

LOAD IDENTIFICATION FOR A COMPOSITE LAMINATED SHELL USING RADIAL BASE FUNCTION NEURAL NETWORK

作　　者：董会丽[1] 郑世杰[1] 

机构地区：[1]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：《工程力学》2008年第3期64-67,共4页Engineering Mechanics

基　　金：国家自然科学基金重点项目(50135030)

摘　　要：针对有限元逆分析方法进行荷载识别的大计算量的缺陷,以及鉴于传统的BP网络的速度慢和局部极小值问题,该文提出了将有限元方法与径向基函数(Radial Base Function,简记为RBF)神经网络结合对受集中载荷作用的壳体结构进行荷载识别。通过有限元方法计算出压电元件的集聚电荷,以该电荷来构建训练样本对网络进行训练,再将没有进行训练的电荷数据送入到训练好的RBF神经网络进行预测,实现对壳体结构荷载的作用位置和大小的评估。最后给出了对壳体结构荷载识别的算例,结果表明该方法计算速度快、精度高、具有较好的应用前景。Due to the large computation expense of finite element inverse analysis and considering the demerits of the conventional BP neural network such as low convergence speed and local extremum, a new method combining finite element analysis and radial base function neural network is presented to identify concentrated load in a shell structure. The charge outputs of four piezoelectric sensors are calculated by tinite element method and used to train the neural network. Furthermore, several samples without training are input into the neural network to predict the location and the magnitude of the load. Finally, an example shows that the present method is effective, feasible and promising.

关 键 词：有限元分析 荷载识别 RBF神经网络 固体壳 压电传感器 

分 类 号：TB33[一般工业技术—材料科学与工程] O316[理学—一般力学与力学基础]