桁架结构损伤识别的数据融合方法研究

Study on Method of Truss Structure Damage Identification Data Fusion

以数据融合技术进行桁架结构的单损伤和多损伤识别；通过研究基于频率的结构损伤理论，分析了归一化的频率和损伤位置的关系，利用小波概率神经网络的算法对决策融合进行了修正，建立了基于小波概率神经网络的数据融合结构损伤识别模型，运用结构计算软件计算了一典型桁架结构的频率，并融合为小波概率神经网络算法的输入特征向量，并对桁架算例模型结构进行了损伤识别，通过桁架不同位置的损伤情况，验证了该方法的有效性，并提出了工程应用注意的问题，研究结果表明，基于小波概率神经网络算法的数据融合技术是一种较可靠的损伤识别方法，具有较好的工程应用前景。

Data fusion techniques are used to the identification of single -truss structure damage and more damage. Based on the study at the frequency of structural damage, it anglicized the relationship between the normalized frequency and damage, and the wavelet probabilistic neural network algorithm is used to modify for decision, in order to establish a model of structural damage identification based on wavelet probabilistic neural network ture vector, and figure out the damage identification in truss structure model, finally, it prove that the method is workable while propose some engineering attentions, so the results show that the algorithm based on wavelet probabilistie neural network data fusion technology is a more reliable damage identification method, which has good prospects in the engineering applications.