BP神经网络及其在结构动力分析中的应用研究

回顾了ＢＰ神经网络在结构分析中的应用和发展过程，针对ＢＰ网络在实际应用时存在的收敛速度慢以及所需训练样本多等主要问题，提出了相应的网络快速学习算法以及训练样本正交化处理方法。同时，对网络结构设计、训练样本规范化处理、网络映射结果精度的自适应修正等具体技术问题也进行了详细讨论，并提出了相应的解决方法。文中最后给出了二个工程应用实例，结果表明：ＢＰ网络在结构动力分析中不失为一种崭新的有效方法。

Research on BP neural network and its application in structure dynamics analysis

In this paper, the application and development of neural network in structure dynamics analysis are reviewed, the fast learning algorithm and samples selection method by orthogonalization for BP neural network are presented, which can improve the convergence speed of neural network and decrease greatly quantities of training samples required in engineering practice. Some particular problems in implement of neural network such as design of network topology, normalization of training samples, adaptive modification of computing results are discussed in detail and solved satasfactorily. At last, two engineering application examples are given and results are promising.