基于径向基函数神经网络的编织复合材料结构脱层损伤监测研究

鉴于传统的BP网络的速度慢和局部极小值问题，以及针对基于实验数据训练神经网络存在样本不足的缺陷，文中提出了利用径向基函数（Radial Base Function，简记为RBF）神经网络通过有限元方法对含有脱层损伤的复合材料试件进行数值模拟，把前五阶弯曲模态频率进行修正，以修正后的前五阶弯曲模态频率再经过归一化处理构建训练样本的新思路，将实验模态分析结果经归一化处理后送入训练好的RBF神经网络进行预测，从而实现对编制复合材料梁的脱层损伤定位和损伤程度评估。最后给出了编织复合材料结构损伤大小伤识别及定位的算例，仿真结果表明RBF神经网络速度快，稳定性好，精度高，在复合材料结构损伤监测中具有光明的应用前景和重要的工程应用价值。

Damage Monitoring for Woven Composite Structure Based on Radial Base Function neural Network

Firstly,due to demerits of BP neural network,such as low convergence speed and local minimum and lack of training samples,a new method for woven composite structure,using RBF(Radial Base Function) neural network,is presented,based on computational mechanics.Secondly,two SW210 fiber glass cloth reinforced composite beams are fabricated,and their modal frequencies are measured by LMS CADA-X modal analysis and test system.Thirdly,the first five flexural modal frequencies of the FEM model with four hundred and fifty-one different conditions obtained by FEM,are modified and normalized by the method given here.Then they are used to train a RBF neural network.Finally,the first five flexural experimental modal frequencies normalized by the same method mentioned above,are input to the neural network to predict the delamination location and its damage level.The results show that the method proposed is feasible,satisfactory and promising.