基于BP神经网络与柔度变化的结构破损诊断

A Method for Structural Damage Detection Based on Back Propagation Neural Network and Flexibility Changes

提出了一种基于 BP神经网络的结构破损诊断方法 ,该方法以结构破损前后柔度的变化作为破损诊断的网络输入 ,为了解决由于系统响应样本数据空间分布不均匀对网络收敛速度及网络诊断的影响问题 ,对网络训练样本采用广义空间格点进行了变换 ,模拟算例及应用实例均表明 ,本文方法能准确诊断结构破损位置与严重程度 ,是一种有效的结构破损诊断方法。

A method for damage detection in structures using backpropagation neural networks is proposed in this paper, which takes the flexibility difference vector of a structures as inputs of the neural network for parameter identification. The method named the Generalized-Space Lattice (GSL) transform original input and/or output data points of all training pattern onto uniformly spaced lattice points over a multi-dimensional space. Thus, the neural network can learn the training patterns efficiently as well as accurately. The proposed method is applied to on a simulated and experimental beam example. The results show that the proposed method can be used to identify the location and magnitude of damages from measured vibration data effectively.