Fractal Characteristics Of Landslide And Risk Assessment In The Near Dam Region Of Xulong Hydropower Station In Jinsha River

Landslide disasters in China are widely distributed, resulting in casualties and economic losses are more serious, especially in recent years with the rapid development of hydropower engineering in southwest of China,hydropower engineering trigger countless landslide, threat stability in the reservoir region and seriously endanger people’s safety and property. So it is urgent to evaluate the risk of landslide in the reservoir area before the hydropower station operated.This paper take the landslide in the near dam region of Xulong hydropower station in Jinsha River as research object, combined with remote sensing image to carry out a detailed geological survey, analyzes the landslide spatial distribution characteristics and formation mechanism of landslide; from the perspective of the fractal reveals the fractal characteristics of the landslide and calculate fractal dimension of the regional topography. Finally, based on the information model and the weights of evidence model in regional landslide of risk evaluation, the research contents and achieved results of this paper are as follows:1 Detailed geological survey of the 25 landslide spatial distribution features such as spatial position, thickness, area and volume, bedrock lithology and so on, and through the field special geological point speculated that the historical process of the formation of landslide; analysis of the formation mechanism of landslide, with rock slope structure and formation lithology are closely related.2 Focus on fractal features of landslide in Maoding estuary and calculation includes the fractal dimension of boundary shape of single landslide、spatial fractal dimension of landslide and three-dimensional fractal dimension of regional topography. The fractal dimension of boundary shape of the largest landslide is 1.0928, most twists and turns,the most obvious expansion deformation.The spatial fractal dimension of landslide is 2.0073,from the fitting curve, the spatial distribution of landslide with fractal characteristic is very high, almost all the tribes in the same line. Three-dimensional fractal dimension of largest landslide topography is 2.4561,the landslide topography is the most complex, and the development of the landslide is very high.The three-dimensional fractal dimension of regional topography of the largest watershed is 2.0583,landform is also the most complex.3 The paper apply the fractal dimension of regional topography to landslide hazard assessment, formation lithology, altitude, slope degree, aspect, fluctuation degree, river erosion, fractal dimension of regional topography, geological structure was selected as eight influence factors to establish system of risk evaluation. The paper study on the landslide distribution and variable dimension fractal characteristics of influence factors, variable dimension transformation frequency is high, fractal dimension value is greater, the influence factors in the landslide risk assessment of the contribution is greater. The maximum of the variable dimension fractal dimension of the geological structure is 1.6076, the distribution and development of landslide is the most important; the minimum of the variable dimension fractal dimension of the fluctuation degree is 0.5470, the distribution and development of landslide is the least important.4Based on the information model and the weights of evidence model in regional landslide of risk evaluation, get the most beneficial form factors of landslide is the altitude in the range of 2200~2400m, the slope to the southwest, slope degree in the range of 0~10 degree, fluctuation degree in the range of 0~30m, loose formation lithology, distance from the fault in the range of 400~600m, distance from river in the range of 0~100m, fractal dimension of regional topography in the range of 2.0275~2.0488. Two methods were used to risk zoning, information model of high risk areas accounted for 39.65% of the proportion of the region’s area, the weight of evidence model of high risk areas accounted for 33.24% of the proportion of the region’s area; using the probability analysis method to verify the accuracy, and the accuracy of the information model is P=71.67%, the accuracy of the weight of evidence model is P=66.91%, and the two methods are of high accuracy.