基于WPNN与数据融合的损伤检测方法

Structural Damage Detection Method Based on Wavelet PNN and Data Fusion

目的为了有效利用结构健康监测系统中的多源传感器数据信息,对复杂结构的健康状况进行诊断进而提高确诊率.方法利用概率神经网络(PNN)的贝叶斯推理与诊断能力及多传感器数据融合原理,将神经网络与数据融合有机结合,使两者优势互补,提出了复杂结构损伤检测技术及其在多层框架结构中损伤检测及诊断中的应用.结果提出了基于小波概率神经网络(WPNN)与数据融合的损伤检测方法.结论基于WPNN与数据融合的损伤检测方法是可行的、有效的.

In order to make full use of multi-sensors data or information from multi-resources and to improve the diagnosis ratio for the health conditions of complex structures, a complex structural damage detection technique based on neural network (NN) and data fusion was presented by means of multi-sensors data fusion theory and probabilistic neural network (PNN). By using their advantages and overcoming disadvantages, this technique combined data fusion with NN, and it was applied to the damage detection and diagnosis of a multi-story framed structure. Therefore, a structural damage detection method based on wavelet probabilistic neural network (WPNN) and data fusion was proposed. The result shows that the method of data fusion-based WPNN is feasible and effective.