基桩缺陷的神经网络判别法

The application of neural network method in idendificating the defect of foundation piles

基桩缺陷辨识是一个高度的非线性问题,现有分析方法人为干预比较多,难以得到准确的桩身完整性结果。利用 MATLAB工具编制了基于BP神经网络模型的基桩缺陷识别程序,该方法具有较好的判别精度。应用神经网络模型结合实测资料对南京地区的一些人工挖孔桩进行完整性分析,取得了较好的应用效果。

Idendificating the defect of foundation piles is a complicated non -linear problem, there are a lot of artificial interference in the methods that we use , so we are difficult to get the accurate result about integrity of foundation piles. The thesis discusses the BP neural network model used in idendificating the integrity of foundation piles. Compiling programme to idendificate defect of foundation piles base on the model, the result of training and examination show that this model is exact on idendification of integrity. Applying the neural network model in dozens of concrete piles in Nanjing region, the method gets a good application effect.