| The Three Gorges Reservoir area is one of the most serious landslide evolution areas of China,and a great deal of valuable data has been accumulated in the process of landslide treatment.The physical and mechanical parameters of rock and soil mass and the geometry of the calculated section are two complex engineering problems because they are affected by many factors which interact with each other.Besides,the acquisition of various parameters has the characteristics of high cost,long cycle and not very reliable.It is helpful to provide reliable basis and effective guidance for landslide treatment in the future by analyzing and studying a large number of landslide data that have been successfully dealt with.In particular,it is of great significance and value to correctly estimate the size and cost of the project that has not been set up by using the existing experience.On the basis of extensive collection of existing prospecting and treatment data,this paper has completed the database of geological disaster in Chongqing which belongs to the three gorges reservoir area.On the purpose of acquiring more precise calculating data and more exact budgets of landslide control by conventional ground surveys method alone before the stage of approval,the physical and mechanical parameters of rock and soil mass,the section shape of landslide and treatment measures of landslide of Three Gorges Reservoir area of Chongqing area are estimated based on data mining technology.The main research contents and the main achievements of this paper are as follows:(1)A parameter estimation system for landslide control engineering design is established based on data mining technology,and the key data such as geotechnical parameters,landslide geometric section and structural parameters of anti-sliding engineering are solved and determined.It can provides a more accurate and reliable new method for decision-making of landslide treatment project.(2)The original documents of geological disaster investigation and design of the Chongqing section of the Three Gorges Reservoir area,including 83 landslides,112 collapsed banks,62 slopes and 14 rockfall areas,have been collected and collated.With Analyzing and processing 102,000 valid data,a graphic database system including query,analysis,calculation and statistic function is set up,which can provide the foundation for the next data mining.(3)The histogram,the scatter plot and the boxplot are used to preprocess the original data,and the correlation between the factors and the calculation parameters was analyzed by Pearson correlation coefficient,and the spatial correlation of data is studied with random field theory.On this basis,the use of multiple regression,support vector machine is established including a series of commonly used data mining technologies such as severe sliding body and sliding zone soil shear strength,sliding zone of gravel content,natural moisture content and other calculation parameters of sliding zone,complete data mining model of geotechnical parameter estimation.These estimation models have been verified to meet the engineering accuracy and reliability.(4)Based on the statistical analysis of more than 310 actual landslide sections,this paper determines the type of sliding surface of the deposit layer which is sliding along the bedrock surface using the "three-stage" slide.By analyzing the factors that influence the shape of the slip surface,the influencing factors for the shape estimation of the slip surface are determined.Through the repeated study of the raw data and the use of Geostudio software,the rationality and reliability of the original computed section are studied one by one,which provides a scientific and reliable foundation for the excavation of the cross section model.According to the correlation analysis,the gradient of the leading edge,the length of the slip mass,the ground slope,the thickness of the slip mass,underground water level,and the gradient of the slide bed are selected as input variables,to establish smooth front edge Angle,length of landslide in the middle level forecast model.By contrast and inspection,the precision of the model is high,and the result is good.At the same time,the existing open GIS data is fully utilized to obtain and draw the landslide ground line and complete the estimation of the calculated cross-sections.(5)On the basis of the landslide residual thrust combined with the restrictions on the size of the landslide treatment project,from the engineering feasibility,economy and environmental protection three aspects to carry on the comparison,anti-sliding project selection suggestions are put forward.When the multi row pile retaining structure is recommended,the numerical method is used to simulate the optimum setting distance between the piles.It is recommended that the distance between the back row piles and the front row piles be 20%of the distance between the front row piles and the trailing edge of the landslide,and the total thrust will be minimal at this time.In addition,the design parameters of anti slide pile under different thrust stages are analyzed.(6)Based on the above research,with the parameters such as the landslide coordinates,the main sliding direction,the sliding body material(broken stone)content,sliding bed material lithology,sliding average thickness,groundwater depth,sliding bed slope,and the length,width of landslide that can be obtained through ground survey only,this paper can provide data for design and calculation of landslide treatment project,and provide the basis for the engineering project decision-making.When the main research is done,optional by others in the study area and outside the area of a landslide as a blind calculation of engineering examples,with the help of the results in this paper,geotechnical parameter estimation,the shape of the landslide section,the project selection arrangement and the determination of structure parameters are achieved.Compared with the actual results of blind survey and design data,the calculation of soil parameters,rock stability coefficient and residual thrust,anti slip regulation project selection and parameters of key data are in reasonable accuracy of engineering error range and on the safe side,suggesting that this paper provides estimates of the treatment method has good feasibility and practicability. |
PDF Full Text Request