论文部分内容阅读
以含损伤的连续梁结构为研究对象,研究了识别结构损伤位置和损伤程度的小波神经网络方法。通过有限元分析和Lanczos法计算得到损伤结构的曲率模态参数,再利用曲率和应变的关系得到应变模态参数。根据连续小波变换理论并用Matlab小波工具箱,得到小波系数模极大值并判断出结构损伤的位置。以此为基础,将小波系数模极大值作为BP神经网络输入参数构造神经网络,通过损伤程度与小波系数模极大值之间的非线性关系,由神经网络的输出参数确定结构的损伤程度。建立了一种既能识别结构损伤位置又能确定损伤程度的小波神经网络方法。通过对一含裂缝的三跨连续梁的损伤识别计算分析,验证了该方法的有效性。
Taking the continuous beam structure with damage as the research object, the wavelet neural network method to identify the damage location and the damage degree of the structure is studied. Through the finite element analysis and the Lanczos method, the modal parameters of the curvature of the damaged structure can be calculated. Then the strain modal parameters can be obtained by the relationship between the curvature and the strain. According to the theory of continuous wavelet transform and Matlab wavelet toolbox, the modulus of wavelet coefficient is obtained and the position of structural damage is judged. Based on this, the neural network is constructed by using the modulus maxima of wavelet coefficient as the input parameter of BP neural network. By the nonlinear relationship between the degree of damage and the modulus maxima of wavelet coefficient, the damage degree of structure is determined by the output parameters of neural network . A wavelet neural network method which can identify the position of structural damage and determine the degree of damage is established. The damage identification of a three-span continuous girder with a crack is verified by the numerical results. The validity of the method is verified.