The Application Of Fuzzy Neural Net To Assessing Rock Cut Slope Stability

Cut slope stability is an important problem during highway construction and utilizing period. The problem is increased with more and more high type roadways constructing. The problem, assessing the stability for a special project, is affected by many factors. A simple and convenience methods is necessary to solving this problem.This paper used fuzzy analysis method and artificial neural network to build a model analyzing state of cut slope. We selected rock strength, angle between surface of cut slope and major structrual plane, state of underwater, angle of cut slope, topography, effloresce, etc. as the major factors effecting cut slope stability. The cut slope stability is assessed by each factor, then the output as the input of neural network. We accomplished the final judgment through the neural network that possess learning ability. There are four layers in fuzzy logic neural net system of cut slope stability analysis. The first layer is input layer consisting of factors effecting cut slope stability state. The last layer is output layer corresponding to cut slope stability state. The first hidden layer has five neurons. The second hidden layer has twenty neurons.We have proved that the fuzzy logic neural net can assess cut slope stability through inputting model to the trained net. It is a convenience and effective method to assess cut slope stability.