Research On Mine Dynamic Disaster Prediction Based On RS-AHP And KM-FCM

At present,with the deepening of underground coal mining,there may be coal and gas outburst or microseismic and stress in the mine.With the information era and the advent of the era of big data,the deepening of mining,the data will also be increased,increasing the difficulty of mining,mine safety disaster frequency also increases,so will be as much as possible the formation of a complete set of mine dynamic disaster prediction system,including processing and storage of data,the system also includes the algorithm efficient prediction of mine dynamic disaster.However,at present,this set of prediction technology in China is not very perfect,and each part is not very complete,which needs to be further optimized.This paper proposes,studies and optimizes three aspects of problems.The first is the speed of data operation.Too many small files lead to a large use of network resources in data operation and the execution of data operation algorithm,which will reduce the work of data extraction.The second is the data storage problem,every one of mine dynamic hazard prediction data is very has the research significance,so a lot of save data as much as possible,in order to prevent data loss,need to copy data technology improvement,but in copy will find the copy of the default storage technology will lead to uneven data load;Thirdly,in the algorithm prediction of mine dynamic disaster,it is found that the traditional prediction algorithm cannot provide good support for the prediction with the increase of sample data and dimension,so the prediction control algorithm needs to be optimized and improved.Three different methods were used for optimization and experimental verification of the three proposed problems.For the merging of small files,this paper compared the three merging methods,and finally selected the sequence file method to merge and compress the data,and put forward the defragmentation system.For carbon storage technology to optimize the paper introduces the rough set theory and analytic hierarchy process AHP,comparing with the copy of the default storage methods,found that the load balancing was solved,but this is only personal experience to assign weights to the judgment matrix,and combining the rough set theory and information entropy by objective weighting method,but this method is not combined with the actual,finally introduced the improved AHP method,this method is combined with the actual job processing important degree.For the optimization of mine dynamic disaster prediction algorithm,this paper introduces the replacement of Euclidean distance with markov distance,introduces the open search method and k-means clustering algorithm,and combines them to optimize the initialization method of the improved markov distance fuzzy clustering algorithm,and finally forms the improved km-fcm algorithm.Through the elaboration and optimization of the three problems,the relevant conclusions are finally drawn.After improving the three problems,a set of integrated algorithm prediction system for mine dynamic disaster is formed,which optimizes the data operation efficiency,improves the copy storage mechanism of data,and improves the comprehensive prediction ability of mine dynamic disaster.There are 57 papers,25 tables and 53 references in this paper.