基于BP神经网络的地下结构损伤识别研究

Damage Identification of Underground Structure Engineering Based on BP Neural Network Technique

随着地铁等地下工程建设规模的扩大与发展,对地下工程结构的损伤诊断及其安全性评价已成为亟待解决的重要问题。提出了一种基于BP神经网络的地铁地下工程结构损伤识别方法,定义结构的曲率变化率,并将其作为BP神经网络的输入标量,采用降低单元弹性模量E的方法来模拟结构的损伤位置和损伤程度,通过设定地下结构各部位的损伤程度,将结构的前4阶曲率变化率作为BP神经网络的输入。结果表明:利用BP神经网络的方法能够准确地识别出地下工程结构的损伤程度,能方便、有效地解决地下工程结构的损伤识别问题。

As the construction scale of subway engineering expands and develops,it is in urgent need of solving the important problems of damage identification on the underground structures. Based on BP neural network, a new method is proposed in this paper to detect the damage of the underground structures,and the change rate of curvature is defined as the neural network＇s input.The element＇s elasticity is reduced to simulate the structure＇s damage location and its ex- tent.By setting different elements and different damage extents of the subway underground en- gineering,the variation rates of the former four mode＇s absolute value are calculated as the BP neural network＇s input. The results show that the method can correctly figure out the magnitude of the damage,and can detect the damage of subway underground engineering con- veniently and effectively.