悬臂板损伤数值模拟试验与WPNN识别方法

Study on Method for Structure Damage Detection Based on Neural Network

通过分析小波概率神经网络(WPNN)与数据融合技术在工程结构损伤识别中的应用原理,建立了基于小波概率神经网络和数据融合技术的模型.对悬臂板结构进行了数值模拟试验,运用损伤单元数据作为输入向量训练了WPNN与数据融合的损伤识别模型,并选取4个单元作为检验样本进行检验,检验的结果与数值试验分析吻合较好,从而表明,该方法在工程结构的损伤识别中有较好的应用价值.

The principle of structural damage detection using wavelet probabilistic neural network （WPNN） and data fusion is expounded. The model based on wavelet probabilistic neural network and data fusion is established. Numerical simulation experiment of suspended panel is made. The model based on WPNN and data fusion is trained by input vector of damage elements data, and the model is validated by 4 damage elements data that are choosed from all the damaged elements. Tested results are inosculated with the numerical simulation experiment, which indicates the effectiveness of this method in the field of structural damage detection.