用概率社会网络进行结构损伤位置识别

在不计测量误差情况下，神经网络能够成功地识别损伤位置及其程度，但在测量噪声影响下，神经网络的损伤识别效果则比较差，考虑到基于多变量模式分类的概率神经网络具有处理受噪声污染的测试数据的能力，本文将可能的损伤位置作为模式类，利用概率神经网络的分类能力来识别结构的损,地对两个算例，一个六层框架和一个两层框架进行数值模拟分析，并将概率神经网络与BP网络进行了比较，结果表明，概率神经网络具有更好的识别效果，是一种很有潜力的结构损伤位置识别方法。

Structural Damage Localization Using Probabilistic Neural Network

Structural damage detection based on vibration measurement has been a very active research subject during the past two decades. In recent years, there have been increasing researches focusing on the application of artificial neural networks (ANNs) in strucural damage identification. Most of them perform well with numerical examples under error free conditions, but become worse when the experimental data are polluted with measurement noise. In the paper, the probabilistic neural networks (PNNs) are applied for structural damage localization. The possible damage locations are considered as the pattern categories in PNNs. The damage location(s) is identified by means of the classification capacity of PNN. Two numerical examples are given in the paper. A performance comparison between the PNN and BP neural network for structural damage localization is carried out. The results show that with the same training and testing and testing samples, the PNN gives rise to a more accurate prediction in damage localization than the BP neural network.