【摘要】 提出了一种基于动力有限元分析和神经网络相结合的含分层损伤层合板的诊断方法。采用作者发展的含分层损伤层合板的动力有限元分析模型和方法,计算了具有不同分层长度损伤层合板的频率和模态阻尼值,以此建立样本库。应用反向传播BP神经网络训练和形成网络。典型含层间分层损伤层合板的仿真结果表明,采用对损伤变化较为灵敏的高阶模态阻尼作为网络的输入参数进行分层损伤诊断比常用的模态频率更为合理。本文中提出的是一种用于层合板的分层损伤诊断的有效和经济的方法。更多还原

【Abstract】 Based on the dynamic finite element method in conjunction of neural networks,a strategy for detecting delamination of composite laminates was proposed.To establish the data-base for neural networks,a nonlinear dynamic finite element method,developed by the authors,was used to calculate the values of natural frequencies and mode damping of the delaminated composite plates,while a back propagation(BP) neural network was employed for training and testing the network.The typical simulating results show that the high order mode damping is better than frequencies as an input parameter for detecting delamination damage,because the former is more sensitive to the extent of delamination.The present strategy is an effective and low cost method for detecting delamination of the composite laminates.更多还原