无黏性管涌型土的BP神经网络判别法研究
Application of BP neural network to assessment of noncohesive piping-typed soils
-
摘要: 运用基于Matlab的神经网络工具箱(NNT)建立无黏性管涌型土的预测模型,并分别以试验数据为研究对象,运用BP神经网络模型对该段土样的渗透破坏形式进行预测。预测结果与规范及试验所得结果相同,说明了基于Matlab的BP神经网络运用于无黏性管涌型土预测的可行性和应用价值。Abstract: The application of neural network toolbox(NNT) based on MATLAB to the assessment of the noncohesive piping-typed soils was mainly illustrated by using the tool of NNT.The feed forward back propagation(BP) network model object for assessment of noncohesive piping-typed soils was created.As an example,the seepage destory formation in some areas of Yellow River and some test example were analyzed here.