基于神经网络的结构动力模型修改和破损诊断研究

Structural Dynamic Model Modification and Damage Diagnosis Method Based on Neural Networks

提出了一种基于神经网络的结构动力模型修改和破损诊断方法，讨论并解决了该法实施中的若干技术问题，如神经网络的拓扑结构和快速算法，结构模态特性的量化比较，结构破损特征信息的提取和处理，计算结果精度的提高等等。

A new structural dynamic modal modification and structural damage diagnostic method based on neural networks is presented in this paper, and some keys problems related to the method such as the neural networks topologic architecture and fast learning algorithms, the quantitative comparison of structural modal properties, the extraction and processing of characteristic information on faults location and extent, the improving of computing precision etc, are discussed and solved, computing results show that the dynamic modal modification method based on neural networks have not some difficult problems which must be dealt with when using other modification method, and have high modification precision. The diagnostic method based on neural networks can determinate precisely the faults extents as well as faults location.