Research On Risk Early Warning In Mine Safety Production

The industry of coal mine is a high risk industry in China, the frequent occurrenceof coal mine accidents caused a great loss to the country and people, which seriouslyaffect the construction of harmonious society, it also had a bad influence in internationalcommunity. In the management of coal mine safety production, changing the passiveand the empirical mode of safety management to the proactive and the comprehensiverisk early-warning model is the key to solve the frequent occurrence of coal mineaccidents effectively in China. Meta-synthesis method from qualitative to quantitative isregarded as risk early warning guidance in mine safety production, on the basis ofcollecting comprehensive risk information, judge risk state of system in advance bywarning knowledge rules and warning models, then achieve the goal of alarm in timeabout current unacceptable risks and real-time warning about future risks. Preventionand control measures can be taken in advance based on the results of early warning, sothe coal mine accidents can be eliminated, reduced and controlled. Implementing riskearly-warning in mine safety production has great significance in implementing thepolicy of the safety production, which is Safety First, Prevention Primary,Comprehensive Treatment, it also has great significance in improving the currentserious situation of safe production in coal mine.In this paper, complexity theory is regarded as guiding ideology, meta-synthesismethod from qualitative to quantitative is regarded as supervising method, and itintegrates many theories and methods, such as, complex science theory, risk earlywarning theory, data mining technology, combined with the empirical analysis, theknowledge of risk early warning in coal mine safety production was studied.First, the complex evolution mechanism of coal mine disasters was studied, thenproposed a new method of early warning on risk state. By analyzing the complexcharacter of mining production system we can find that mining production system is notonly a dissipative structure system, but also a complex giant system, meta-synthesismethod from qualitative to quantitative is the most effective way to solve this problem,so this method is regarded as supervising method of risk early warning; Based on thetheory of self-organized criticality, analyzed the power rate about accidents of peopledied in coal mine and people died in coal mine by gas, the results showed that there is acharacter of self-organized criticality in Chinese coal mine accidents; According to the complex character of coal mine, put forward a idea of early warning on risk state, therisk state of system is divided into three states, they are far away from critical state,critical state and supercritical state, the risk state of system about coal mine safetyproduction is constantly changing during transient and steady state in far away fromcritical state, but those changes can be predicted and controlled; The critical state is aspecial form about the evolution of risk states, the evolution of risk state in critical statemay be mutation, this evolution is unpredictable, so the risk in critical state is the key toprevent and control. The supercritical state is the further positive development of criticalstate, at this state, coal mine accident must occur and must be alarmed at once. Theprevention and control in early warning is to ensure that the risk state must keep intransient and steady state during far away from the critical state all times.Second, the theoretical system of the risk early warning in coal mine was studiedand preliminarily established. According to the principle of ‘preventive treatment ofdisease’, analyzed the risk early warning method of ‘syndrome differentiation andtreatment’ in coal mine; The concept of dominant risk source and recessive risk sourcewere proposed by risk early-warning features; Put forward the risk early warningprinciples, which are Human-oriented, Prevention First, Timely and in Time;Established the procedure of risk early warning, they are marking the alert meaning,seeking the alert source, analyzing the alert symptom, forecasting the alarming degreeand putting prevention and control, and established the method of risk early warning.Third, the establishment of risk early-warning index system in coal mine wasstudied. Put forward a method to get subjective weight by combining analytic hierarchyprocess (AHP) with set-pair analysis, studied the method to get the objective weightbased on rough set theory, and put forward a method to get objective weight bycombining rough set theory with information entropy theory, then used examples toanalyze and validate; The indexes system of recessive risk were set up by four riskfactors,they are people, machine, environment and management, then the membershipfunction of system indexes were established by fuzzy mathematics method.Fourth, the establishment of knowledge base about risk early warning in coal minewas studied on the base of data mining technology. Proposed the representation of riskearly warning in coal mine, they are production representation and frame representation.Production representation is suitable for dominant risk representation, and framerepresentation is suitable for recessive risk representation; Point to productionrepresentation, the decision tree such as classification and regression tree(CART)& C5.0were used to empirical study about risk early warning. Point to framerepresentation, case-based reasoning (CBR) was used to empirical study about risk earlywarning.Finally, the risk early warning model based on set-pair analysis theory was studied.Point to the uncertain feature of risk, analyzed the description of risk state byconnection degree based on the description of macro and micro evolution, and analyzedthe uncertain interval of connection degree by catastrophe model; proposed thirdwarning degree for dominant risk and fifth warning degree for recessive risk base onset-pair analysis model, then verified by examples.