应用人工神经网络技术,以反映结构损伤程度的固有频率作为神经网络输入的特征参数,建立结构故障诊断模型,对悬臂梁裂纹的深度和位置进行了诊断研究。从诊断结果说明神经网络用于工程结构的裂纹故障诊断是可行的。 Using artificial neural network technology, the natural frequency which reflects the extent of structural damage is taken as the characteristic parameter input by neural network, and a structural fault diagnosis model is established to diagnose the depth and position of cantilever crack. The diagnosis results show that the neural network is feasible for the fault diagnosis of engineering structures.