神经网络在水工混凝土结构裂缝检测中的应用

Application of neural network to crack detection in hydraulic concrete structures

针对水工大体积混凝土结构的特点,探讨了混凝土结构裂缝的检测方法.基于BP神经网络建立混凝土裂缝深度与其相关参数之间的关系模型,以实测资料作为模型的训练样本,并通过实例进行了检验.结果表明,BP神经网络应用于混凝土裂缝检测分析是可行和有效的.

In accordance with the characteristics of hydraulic mass concrete structures, the method for crack detection was discussed. A model for the relationship between the crack depth of concrete and its correlative parameters was developed based on BP neural network, and the measured data were used to train and test the network model. Case study shows that the application of BP neural network to crack detection is feasible and effective.