Application Of Neural Network Methods Based On Gray Theory For The Breakwater Slope Stability Analyzing

This thesis summarized the building situation of Nanning dyke first, selected the statistical information about the slide disaster situation of Nanning breakwater slope which has taken place, analyzed various kinds of factors that influenced breakwater slope stability, including slope high, slope ratio, average unit weight , cohesive force , angle of internal friction , water head , dyke distance and load, applyed the Gray correlation analysis of Gray system theory to analyze Gray correlation about the stability influence factors of breakwater slope, and got the order arrange of each influence factor according to the influence degree of slope stability, and according to the analysis results , suggested the scheme that how to design and manages the Nanning dyke .Then choosing five advantage factors of Gray correlation analysis: The angle of internal friction , cohesive force , water head , dyke distance , average unit weight as the input samples, introduced improved BP algorithm of Artificial Neural Network, adopted neural network function storehouse of MATLAB to write the procedure , setted up neural network predicting model of analysis of Nanning breakwater slope stability.Finally we trained the network by using the database which came from typical Nanning breakwater slope instances , and forecasted the slope stability of Nanning breakwater slope , compared the analysis results on network appraisal with results of the traditional method one by one, the comparing results proved the method can meet the appraisal requirements of Nanning breakwater slope predicts of stability.Because the optimum seeking of the influence factor and improvement of the network algorithm have reduced the calculation time of the network greatly , this method make the appraisal of the breakwater slope stability more convenient rapidly, it is proved the gray -neural network method of appraise the breakwater slope stability have certain applying values and theoretical values, with the abundance and substantiating of the sample storehouse , the precision and suitability of this method will be improved further, and increase its applying value.