Effectiveness Analysis and Improvement Simulation of Coal Mine Safety Management under the Background of Big Data
Author:Qiao Wan Guan
Supervisor:li xin chun
With the rapid development of big data technology and the gradual improvement of security management theory,data mining technology has attracted the attention of many scholars and enterprises in the field of safety management decision-making.Faced with the increasing safety needs of employees and the high attention of public opinion,coal mining enterprises have to improve the level of safety management to adapt to the rapid development of society.However,how to accurately,efficiently and objectively evaluate the safety management efficiency of coal mine enterprises and improve its effectiveness is an urgent safety problem for coal mine enterprises.In the Thirteenth Five-Year Plan for Safety Production(No.3,2017),it was pointed out that "information technology applications such as big data of safety production should be promoted in an all-round way,and the ability of early warning and monitoring of major hazards,hidden dangers,risk control and emergency disposal should be enhanced".It shows that the application of big data technology and methods in coal mine safety management by the government and coal mine enterprises is extremely urgent.Based on the above background,this subject combines the theories and methods of safety management,system engineering and data mining,and proceeds from the big data background,studies the evolution and transformation of the current theory of coal mine safety management in China,the research paradigm of coal mine safety management and the mechanism of accident occurrence,and so on.The factors affecting coal mine safety management are deeply excavated,the efficiency evaluation of coal mine safety management and coal mine safety.Five aspects of management efficiency improvement are studied to solve the efficiency problems existing in the process of implementing safety management in coal mine enterprises,so as to formulate the development path and related strategies to improve the level of coal mine safety management.The specific research contents and conclusions of this topic are as follows:(1)Under the background of big data,the paper mainly elaborates on four aspects: safety management concept,safety management efficiency,safety management methods and safety management thinking.In terms of the concept of safety management,it is found that safety management from the four perspectives of accident cause,traditional safety,system security and big data security has a certain time sequence,but there are overlapping situations in their respective periods.In the aspect of the reform of safety management efficiency,it is pointed out that the current research on coal mine safety management efficiency lacks comprehensive,forward-looking and timeliness,and the content of safety management efficiency research under the background of big data is constructed.In terms of the change of safety management methods,this paper introduces the current research models and methods of coal mine safety management from three aspects: knowledge-driven,model-driven and datadriven,and puts forward a coal mine safety management method based on hybriddriven.Finally,this paper elaborates the change of current safety management thinking from four aspects: experience,system,risk pre-control and big data.It is found that under the background of big data,the thinking of coal mine safety management is changing from causality-correlation,static-dynamic,reverse-positive,experience-data thinking.(2)In the aspect of coal mine safety management mechanism under the background of big data,firstly,the connotation of coal mine safety management digitalization is defined,which makes the object of coal mine safety management change from the traditional physical individual to the individual-generated large data management under the background of big data,and points out that coal mine safety management under the background of big data not only has the main characteristics of big data,but also has its own shortcomings.Secondly,the data of coal mine safety management are summarized and classified according to the different structures,sources and attributes of coal mine safety management data.Then,the large and small data of coal mine safety management are compared and analyzed from the angle of data size,and the data,information and regularity transformation model of coal mine safety management under the background of large data is put forward.The results show that the data and information of coal mine safety management are reliable.The relationship between information and law is not a simple linear model,but involves four security transformation modes,in which security knowledge plays a mediating role.Finally,the theory of spatio-temporal data field is introduced to explain the mechanism of coal mine accidents from the data point of view.The results show that when there are errors in association rules generated by spatio-temporal effect of coal mine safety data,it may lead to coal mine safety accidents.When the new coal mine safety data,information,knowledge and rules generated by the spatial effect of coal mine safety data do not match the original coal mine data space,it will lead to the asymmetry of safety information or unstable safety rules,and cause accidents.When the collision and intersection of time and space data of coal mine safety results in time and space intersection effect,which causes the instability of man-machine-loop-pipe interaction in coal mine production system,and leads to the unsafe behavior and hidden dangers of coal mine safety can not be controlled timely and effectively.(3)In the aspect of data mining of coal mine safety management,this paper firstly puts forward three-dimensional structure diagrams of big data mining of coal mine safety management from three aspects: time dimension of safety production,data mining method dimension and safety domain dimension.At the same time,it gives six main functions and algorithms of big data and ten fields of big data application in coal mine safety management.Secondly,the basic flow of big data in coal mine safety management efficiency research is elaborated from five aspects: problem definition,data preprocessing,modeling,model evaluation and model application and optimization.Then the decision tree classification function is used to classify and count the number of unsafe behaviors of miners.The results show that training,attendance,experience and age are all factors affecting the frequency of unsafe human behavior.Training factors have the greatest impact on unsafe behavior.Finally,Apriori association rule algorithm is used to mine the relationship between hidden dangers in coal mine safety.Apriori association rule mining algorithm is used to correlate the factors such as department,time,location,month and risk level of hidden danger management in coal mine.Strong association rules among multiple dimensions are found to avoid or reduce the occurrence of coal mine accidents.At the same time,the use of these strong linkages can improve the efficiency of our investigators.(4)The efficiency evaluation of coal mine safety management.In this chapter,the data-driven method and model-driven method are combined to evaluate the efficiency of coal mine safety management.Firstly,the advantages and disadvantages of modeldriven and data-driven evaluation methods for coal mine safety management efficiency are introduced.Then,the DEA-BP neural network hybrid driving coal mine safety management efficiency evaluation model is proposed.By using DEA-BBC model to carry out static analysis of coal mine safety management efficiency,and using DEAMalmquist index to carry out dynamic analysis.Then,DEA is used to get the efficiency evaluation results and code them as the expected output of the second evaluation prediction.Finally,BP neural network is used to further evaluate and predict the safety management efficiency of different coal mine enterprises.The results show that the safety management efficiency of coal mine shows an upward trend,which shows that the current safety management efficiency of coal mine enterprises is not stable.At the same time,the number of enterprises with effective safety management efficiency also shows an upward trend of shocks.In the process of dynamic analysis,coal enterprises are blindness in improving safety efficiency.They fail to notice that safety efficiency is the result of technical efficiency and technological progress index,which leads to the decrease of technical efficiency and the increase of technological progress efficiency or the decrease of technological progress efficiency.Finally,for coal mine enterprises with insufficient output,we should strengthen the management of accident and hidden danger data,find out the strong association rules between coal mine accidents and hidden dangers,and use these rules to improve the detection rate of hidden dangers and reduce the number of casualties.For coal mine enterprises with redundant input,the emphasis should be placed on personnel structure and safety input efficiency,and the investment should be reduced appropriately to increase scale efficiency.(5)In the aspect of simulation and optimization of coal mine safety management efficiency level under the background of large data,the boundary of coal mine safety management efficiency system is defined as four levels: employee safety management subsystem,hidden danger management subsystem,accident management subsystem and safety input management subsystem by using system dynamics.The causality diagram and flow chart are constructed by introducing large data influence coefficient as adjustment variable.Finally,taking Wanglou Coal Mine as an example,the following conclusions are drawn: the increase of large data impact coefficient is consistent with the change trend of coal mine safety management efficiency level,and the change speed of different factors influencing system safety management efficiency level is different.Among them,the influence of big data on the safety management of employees is more obvious.The increase of the influence coefficient of big data on the safety management of miners will inevitably lead to the improvement of the efficiency of coal mine safety management.But after reaching a certain coefficient,the improvement of the efficiency of safety management is slower and slower,which indicates that the influence of big data on the efficiency of coal mine safety management is limited.It is more and more difficult to improve the whole management level,and it is difficult to increase the efficiency simply by changing the single big data influence coefficient.After changing several large data impact coefficients,it is found that the efficiency of coal mine safety management has greatly increased,among which the most obvious effect is to change all the impact coefficients at the same time.The second is to change the large data impact coefficients of three subsystems,and the last is to change the impact coefficients of two subsystems.By adjusting the proportion of different big data influence coefficient and SD simulation analysis,the change rate of safety management efficiency level under different schemes is compared,which can provide reference for coal mine enterprises to analyze and improve large data.There are 63 figures,35 tables and 257 references in this paper.