Study On Early Warning Model Of Surrounding Rock Stability Of Mining Roadway Based On SVM

Ensuring the stability of coal mine surrounding rock is a prerequisite for achieving coal mine safety.In the initial stage of actual engineering mining,it is often necessary to classify the surrounding rock stability of coal mining roadways to provide scientific references for the design of support schemes and the layout of roadways.Although the traditional fuzzy mathematics method can be classified,it has certain disadvantages.In view of the advantages of support vector machines for the processing of limited sample data this paper proposes to use support vector machines to establish a grading model for processing,and optimize through optimization algorithms to achieve accurate classification of the surrounding rock stability of roadways.Another important content for studying the stability of surrounding rock is to monitor the roof displacement data of the roadway and predict the roof displacement,so as to achieve the purpose of early warning and ensuring the safety of construction.This paper mainly studies two core contents:the surrounding rock stability classification and the roof displacement prediction and early warning method.For the classification of surrounding rock stability,this paper analyzes in detail the reasons that affect the stability of surrounding rock in coal mine roadways,and collects data on 56 typical roadways.The six parameters of surrounding rock strength,coal seam strength,burial depth,rock integrity coefficient,water inflow and coal pillar width are determined as the main influencing factors.Firstly,the roadway is classified by traditional fuzzy mathematics method,and then using the same data,a support vector machine classification model is established to classify the stability of the surrounding rock,and optimized by grid optimization algorithm,particle swarm optimization algorithm and genetic algorithm.The results show that the genetic algorithm has the best optimization results and the cross-validation accuracy rate reaches 92%.Compared with the traditional fuzzy mathematics method,the support vector machine classification model reduces manpower and material resources,and has certain scientific guiding significance.At the same time,in order to accurately predict the stability of the roadway roof during coal mining and reduce the possibility of roof accidents,this paper proposes a support vector machine-based roadway roof displacement prediction method and constructs a roadway roof displacement prediction and early warning model.Under the given conditions,the two indicators of time and distance from the working surface are selected as the model input,and the roof displacement as the model output.In order to improve the prediction accuracy of the model,the grid search method,particle swarm optimization algorithm and genetic algorithm are used for optimization,and finally the genetic algorithm is used to optimize the support vector machine parameters.Taking Hujiahe coal mine roadway as the research object,four sets of roadway roof displacement data are selected as sample data,and then the model is used for training and prediction.The analysis results show that the prediction results of the model are consistent with the actual results,which can provide good theoretical support for the danger warning of the roadway roof.