Research On Prediction Of Deformation And Failure Of Complicated Rock Slope Based On Graph Machine Learning

In the field of geotechnical engineering,the deformation and failure analysis of complex rock slopes is a very important research topic.The rock slope in nature is composed of rock blocks and a large number of structural surfaces.Various structural surfaces such as joints,cracks,and weak interlayers often have a huge impact on the stability of the slope.Traditional methods based on continuum,such as finite element method and finite difference method,are not suitable for actual jointed rock slopes.Methods based on discontinuous media,such as the block discrete element method,can describe this characteristic of rock mass,but when analyzing the deformation and failure of rock slopes,traditional judgment standards are often not applicable.Therefore,this paper proposes a new research method that no longer sticks to the concept of traditional computational mechanics,but looks at the problem from the perspective of graph theory.From the perspective of graph theory,the rock blocks in the rock slope are equivalent to the nodes in the graph network,and the contact between the rock blocks is equivalent to the edges in the graph network,that is,the rock block system as a whole is represented as a graph.The internet.At the same time,this paper analyzes the contact network of rock blocks by using the analysis method of graph network-graph machine learning method.The changes in the relevant characteristics of the graph network reflect the deformation and failure of the rock slope.The main research work of this paper is as follows:(1)The rock block system was graphically characterized and time series graphical data was constructed.This paper uses the knowledge of graph theory to characterize the rock block system,corresponds the elements in the rock block system to the elements in the diagram one-to-one,and proposes the concept of the rock block contact network.At the same time,this paper uses block discrete elements to perform numerical simulations,and obtains the contact relationship of the rock block system at a specified iteration interval,thereby constructing time series graph data.(2)The analysis method of rock slope deformation and failure based on community detection is proposed.This paper uses the graph-based machine learning algorithm-Louvain community detection algorithm to analyze the time series graph data of the rock block system.By comparing the calculation results of the stable and unstable simplified slopes,the deformation and the deformation of the rock slope are compared.The damage was analyzed and evaluated.(3)The analysis method of rock slope deformation and failure based on k-core decomposition algorithm is proposed.This paper uses the graph-based machine learning algorithm-k-core decomposition algorithm to analyze the time series graph data of the rock block system,and compares with other scholars' research methods to verify the deformation and failure of rock slopes.Analysis and evaluation.(4)Application case study based on graph machine learning method.This paper takes the Yanqianshan Iron Mine in Anshan City as an example,uses the MIDAS software to establish a three-dimensional geological model,and uses the research methods of other scholars to divide the equivalent joints;then the slope is excavated in layers under the condition of block Discrete element numerical simulation is used to construct the time series graph data of the rock block system;finally,the two methods proposed in this paper are used to analyze the deformation and failure of the slope,and the slope deformation under four working conditions is given.And destroy the results of the analysis.