Prediction Of Stability Of Roadway Surrounding Rock Based On Concept Lattice And Probabilistic Neural Network

The establishment of the evaluation Index system of roadway surrounding rock stability is the foundation and key of the stability evaluation of roadway surrounding rock,there are many factors influencing the stability of roadway surrounding rock,and the relationship between evaluation indexes is uncertain and concealed.At the same time,the parameter selection and adaptability of the model can also have a great influence on the prediction result,so it is one of the main research contents to choose the evaluation index rationally and improve the accuracy of the prediction model of the surrounding rock stability of the roadway.Analysis of instability types and mechanism of surrounding rock,the main factors affecting the stability of roadway surrounding rock are the material quality of surrounding rock itself,the integrality of rock mass itself,the groundwater,the stress of surrounding rock,the roadway section and so on.Based on the principle of statistics,the indexes of rock mass index RQD,uniaxial compressive strength RC,rock mass integrality coefficient,groundwater seepage quantity and joint condition are selected preliminarily.By using the concept lattice multilevel attribute reduction method to reduce the primary index,the rock mass Integrity index and joint condition are reduced,and the evaluation Index system of roadway surrounding rock stability is established.The correctness of reduction is validated by Pearson relativity theory.The 20 groups of unbalanced training samples are synthesized into 100 sets of balanced datasets using a synthetic sampling technique,and a probabilistic neural network model based on symmetric Alpha stable distribution is constructed with symmetric alpha stable distribution instead of Gaussian distribution as the base function of the model.In view of the stability evaluation of roadway surrounding rock,the genetic algorithm is used to optimize the model parameter ? and scale parameter ?,and the values of ? and ? are 0.2848,1.5963 respectively.The validity of the model is validated by the 100 training sample data being returned into the model,the prediction accuracy is greater than 96%.Based on concept lattice and probabilistic neural network,a combined model for predicting the stability of roadway surrounding rock is established,taking 10 roadways in the mining area of the pointed forest in Daye iron mine as engineering objects,the accuracy rate of surrounding rock stability of roadway is 90%,and the accuracy rate of model prediction based on principal component analysis and probabilistic neural network is 80%,and the comparative analysis shows that The prediction model of roadway surrounding rock stability based on concept lattice and probabilistic neural network has high accuracy and reliability.The study optimizes the evaluation index system of roadway surrounding rock stability,expands the application field of the concept lattice multilayer attribute reduction method,enhances the adaptability of probabilistic neural network model,and provides a new method for predicting the stability of roadway surrounding rock.