基于曲率模态和神经网络的结构损伤诊断

Structural Damage Detection Based on Curvature Mode Shapes and Neural Network Technique

基于神经网络算法,利用曲率模态参数对框架结构损伤定位和定量识别问题进行了研究和实例分析,分别用两种网络对结构损伤进行诊断,并用灰色理论对结构损伤进行了预测。结果表明,灰色理论能成功地对结构损伤进行预测,神经网络适用于此类损伤无规律对象问题的诊断。

On the basis of the theory that curvature mode shapes can be employed to determine the locations and degrees of damage of structures, a BP neural network technique with improved input structure and a RBF neural network technique are developed to detect the structural damages, and the grey system theory is used to predict the degrees of damage of structure. The results show that while the gray system can be very successful in structural damage prediction, neural network technique is applicable to irregular structural damage detection.