【摘要】 以框架结构为研究对象,利用小波分析和神经网络理论,结合二者的优点,运用小波分析来确定框架结构的损伤位置,运用神经网络算法来识别损伤程度,给出了基于应变模态参数识别框架结构损伤的原理,建立了一种识别结构损伤的小波神经网络方法.通过建立基于振型模态和应变模态的损伤识别方法,分别对9种不同工况下框架的裂缝位置进行识别,并对比了这2种模态下损伤位置的识别效果.然后,分别对框架的振型模态和应变模态进行连续小波变换,获得2种模态参数下的小波系数模极大值.利用神经网络去模拟小波系数模极大值与损伤程度之间的非线性关系来识别结构的损伤程度,并对比了这2种模态下损伤程度的识别效果.数值分析结果表明,小波神经网络可以有效地识别出结构的损伤位置和损伤程度,基于应变模态的损伤识别方法具有更好的准确性.更多还原

【Abstract】 A frame structure is regarded as the object of research,wavelet analysis is applied to determine the location of damage of frame structure and the neural network method to identify the damage degree.The principle of strain mode parameter identification of damaged frame structure is given and thus a wavelet neural method for identifing damage structure is proposed.Through the establishment of damage identification method of vibration mode and strain mode,the framework crack positions under 9different working conditions were identified,and the identifying effects in the two modes were compared.Then the vibration parameters and strain parameter were undergone continuous wavelet transform to obtain the two maximum mode parameters.Neural network was applied to identify the damage degree of structures through simulating the nonlinear relationship between the maximum of wavelet coefficients of the structure and the damage degree,and the damage identification effects in the two modes were compared.The numerical analysis indicates that the wavelet neural network can identify the damage location and degree of structure effectively,and the accuracy of damage identification method based on strain mode is better.更多还原