论文部分内容阅读
为了探索一种可能的结构损伤在线监测方法，使结构损伤动力识别的研究更具工程实际意义，以工程上常用的截面开口槽形梁为试件，用试验模态分析与神经网络法相结合的方法，研究了位移频响函数指标、应变频响函数指标和模态频率指标在结构损伤定量辨识中的灵敏性。结果证明了神经网络法与试验模态分析技术相结合检测结构损伤的可行性，并指出应变指标对损伤更敏感。此外，指出了对于具有复杂模态的结构无法获取模态的完善信息时，用有限个点、有限个方向的测量数据进行损伤识别是可行的。
In order to explore a possible on-line monitoring method of structural damage, the research on structural damage dynamic identification is more practical and practical. Using the slotted girder of sectionally open section commonly used in engineering as a test piece, the experimental modal analysis and neural network method are combined Method, the sensitivity of displacement frequency response function index, strain frequency response function index and modal frequency index in structural damage quantitative identification is studied. The result proves that the combination of neural network method and experimental modal analysis technique can detect structural damage and points out that strain index is more sensitive to damage. In addition, it is pointed out that it is feasible to use a limited number of points and a limited number of directions for damage identification when the complex modal structure can not obtain the modal perfect information.