Stability Evaluation Of Highway Slope In Shaanxi Province Based On Intelligent Algorithm

With the rapid development of highway transportation, the numerous emerging events because of highway slope failures have made adverse impact on social and economic development of human being. The slope stability analysis has become one of hot issues in the research area of slope hazards. The slope engineering is such a complicated system that it has been a consensus for researchers to get solutions about slope stability problems with a sole scientific theory and method.Because highway slope geological hazards are highly nonlinearly related with a few nature or social factors, there is a certain constraint on the application of conventional slope stability analysis method. With the development of computing technology, artificial intelligence is becoming a new method in slope stability analysis, such as Support Vector Machine method. More and more complicated problems are solved with it on data acquired in different ways, due to its advantages of using limited samples, high dimension, non-linearity and local minimum effectively.Focusing on the highway slope geological hazards in Shaanxi province, this paper studied slope stability with the application of GIS spatial analysis technology to get understanding the spatial character of slope geological hazards, and with support vector machine method to set up a stability evaluating model by the parameters’ optimization.Based on the field investigation and analysis of terrain, topography, geological structure, rock and soil type as well as the other factors in Shaanxi province and combining with the data from "Daily Maintenance Management System of Shaanxi Province", the paper optimized the major factors influencing the slope hazards in Shaanxi province and established the corresponding stability assessment indicator system.Given the characteristics of the area and highway engineering situation in Shaanxi province and based on the GIS overlay method of regional influencing factors, the regional evaluation model is built for highway slope hazards with respect to the main factors of rainfall, penetration depth of surface, rock and soil type and earthquake, and the graphs of each single factor are completed. The weightings for each factors are determined by the method of quantitative analysis of correlation. Highway slope zoning map of Shaanxi province is obtained, which lay the foundation for the establishment of slope stability intelligent analysis models of different areas.On the basis of zoning map, hazard stability analysis model were put forward by using the support vector machine. The result is specially important to the small sample data in the areas of northern Shaanxi, Guanzhong and southern Shaanxi. The data of studied areas with different forms, include slope types, slope grade, the maximum slope height, degree of weathering, vegetation coverage, intensity of slope rock and soil, were put into stability analysis model while the stability state of slopes were put out.. The most widely used radial basis function is adopted as kernel function, and three different parameter optimization methods, i.e., article swarm optimization, genetic algorithm mesh searching method were selected. The best penalty factor C and kernel parameter g were obtained during the process of continuous optimization. Finally, intelligent model with a higher accuracy is determined after the optimization of support vector machine. The studies indicated that mesh searching method is less time-consuming and of worse accuracy, while particle swarm optimization and genetic algorithm is of long time consuming but high accuracy. By comparative analysis and optimizm this paper got the reliable model with 95.8 percent of precision so that the model could be utilized in the highway slope stability analysis of Shaanxi province. In a whole, a better modified support vector machine analysis model with good generalization performance and stability is obtained through the parameter optimization. The relatively high accuracy under the verification of test data proved that the stability analysis model of modified support vector machine reflected the non-linear characteristics of highway slope system. It possesses many advantages and formore application future.Based on the MATLAB platform and combining with the Microsoft Visual C++ editor, the artificial intelligent highway slope stability analysis system, which is convenient for qualitative analysis of highway slope stability, is developed by the extended programming using Libsvm tool boxes.On the time of advocating the application of big data, the highway slope stability analysis method based on intelligent algorithm, can scientifically analyze the stability of highway slope in different situations, and providesignificant reference and guiding for preventing and controlling of highway traffic disasters.