Study On Early Warning Technology Of Stability Margin Of Rock Slope Based On Comprehensive Measurement

In view of the existing conditions of formation of geological landslide and the complexity of induced factors,the use of traditional means of measurement has poor real-time performance and low accuracy,a kind of early warning technology based on comprehensive measurement for stability margin of rock slope is proposed.In order to better analyze the nonlinear development trend of landslide,this paper uses BP neural network to establish the early warning model of geotechnical landslide.In this paper,the sensor is used to collect data which can represent the state of landslide,including rainfall,soil moisture content,surface displacement,slope angle,soil shear force,soil stress and underground three-dimensional displacement.The data measured by the sensor is used as the training sample and test sample to create BP neural network,and the comprehensive measurement of the data is realized.At the same time,the information collection and display system of landslide warning is built.The sensor signal is collected by microcomputer control module,and RF wireless radio frequency technology is used to transmit and aggregate the data.The SSH frame of JAVAEE is used to display the dynamic curves of rainfall,moisture content,surface displacement and soil stress in real time.Using WPF Framework of C# to develop online Real-time Monitoring system of Underground 3D displacement,realize Visualization of Underground reality.MYSQL database is used to store the collected data,and the history of the data can be traced back.The BP neural network model is established by using the collected samples.The input variables of the neural network are extracted by principal component analysis(PCA)to eliminate the overlapping information between the input variables and to reduce the coupling degree between the variables.Using a genetic algorithm with global search function,the weights and thresholds of the input layer,the hidden layer and the output layer are improved.The weights and thresholds are taken as variables to be improved in genetic algorithm and coded.The individuals after coding are repeatedly selected,crossed,mutated,and finally the optimal individuals are obtained and the optimal combination of weights and thresholds is obtained.Genetic algorithm is used to avoid the shortcoming of the neural network falling into the local minimum value,improve the training precision of the neural network,speed up the convergence of the neural network,and make the prediction value of the output layer closer to the ideal expected value.The performance and characteristics of GA-BP neural network and BP neural network are compared and analyzed.In this paper,the soil of rock and soil slope in a certain area of Hangzhou is taken as the experimental object,and the landslide occurrence process is simulated by using the landslide geological hazard simulation test site as the experimental platform.The sample data of landslides are collected by comprehensive measurement and the neural network model of landslide is established.The experimental results show that the prediction results obtained by using the intelligent learning algorithm of neural network are basically consistent with the actual situation,and the feasibility of the prediction of landslide stability margin by GA-BP neural network is verified.