Research On The Application Of Data Mining In The Coal Mine Gas Monitoring System

Coal-mine is a major source of energy in China, with the continuous deepening of coalmining, coal mine gas concentration increases gradually and gas accidents becomeincreasingly serious,which is threatening the country and the people’s lives and property, theeffective monitoring of the gas in the coal mine production is very urgent. Although coalmining enterprises also established a monitoring system of gas safety, it is mainly suit to localmonitoring and management, lacking of effective treatment and deep analysis of themonitoring data. In order to achieve better prediction function, this study use data mininganalysis methods to make in-depth analysis of the relevant monitoring data.The purpose of this paper is that the data mining techniques are applied to the analysis ofthe data of the coal mine gas monitoring system, that is, we use clustering, time seriesanalysis, and other new technologies to make in-depth analysis of the data which is collectedby the monitoring system, to provide an exploration and guidance for coal mine gasmonitoring and forecast.After introducing the topic background, a description of the technology involved in thesubject, and then by analyzing the characteristics of the coal mine gas monitoring dataindicators, set up a data mining processing model, and on this basis to do the followingaspects: Application of fuzzy K-means method for coal mine gas safety rating; application ofARIMA model of time series analysis method to predict the daily average concentration of thegas, which is to grasp the gas concentration distribution and trends; In order to excavationstrong association rules,we mine multi-dimensional association rules on safety monitoringdata through using the method of association rules,and proposes a method for estimating theminimum support threshold based on rule antecedent, which is used to determine theminimum support threshold in mining association rules.Empirical studies have shown that themodel performance is good, simple, the method used is reasonable and effective.