基于BP神经网络算法的土木结构损伤检测研究

结构损伤会导致其振动频率发生变化,基于频率测量的结构损伤识别方法简单方便,在结构健康监测领域应用广泛.本文在分析结构固有频率的基础上,把结构损伤识别问题分解为损伤状态辨别、损伤定位、损伤程度识别三个过程,对每个过程用模态参数构造标识量,并作为特征参数输入到神经网络中实现损伤识别.结合实验室一箱形钢梁的损伤检测试验进行了应用,收到满意的效果.

The Study of Structural Damage Detection Based on BP Neural Network Technique

The method based on frequency measurement to localize structure damage is widely applied to structure health monitor.In this paper,the structural damage detection is divided into three courses,which are structural damage identification,structural damage localization and structural damage severity determination course.Damage detection courses are named by modal parameters and inputted to neural network as parameters for structural health monitoring.Damages of different locations and severity on a box beam are simulated with BP neural network.