Study On Safety Trend Forecasting In Mine System Based On Environment Factors

In order to achieve the purpose of safety production in mining management,mining safety problem usually has been studied through the analysis of its internalaffecting factors from the causes of mining accidents or the research on mining systemsafety trend forecasting based on mining data. But the mining system is severelyrestricted by environmental factors, so it is necessary to study the mining system safetytrend which is based on environmental factors.After analyzing the environmental factors of unsafe man behavior and insecurityphysical state by Accident-Causing Theory, environmental factors attributing to unsafeman behavior and insecurity physical state have been studied and the environmentalfactors system of mining safety system has been built. The degrees of eachenvironmental factor have been identified by using Analytic Hierarchy Process.According to the weighted Markov chain theory, the mining system safety trendinfluenced on environmental factors has been forecasted by using the weightedMarkov chain prediction model and analyzed based on the number of mining safetyincidents. The relationship between the fitted values sequence and the number ofmining safety incidents has been analyzed based on seasonal index trend model. Themining transfer states space have been divided by its boundary lines. When the numberof mining safety incidents is larger than the fitted values, the upper limit boundary linehas been determined by the maximum value of their residuals and the upper boundaryline has been determined by the mean value of their residuals. When the number ofmining safety incidents is less than the fitted values, the lower limit boundary line hasbeen determined by the minimum value of their residuals and the lower boundary linehas been determined by the mean value of their residuals. The weight for the weightedMarkov chain model based on the affecting of environmental factors has been adjustedby the matrix product of the weight which reflects the relationship of each statustransition probability matrix affected by lag time and the effect degree ofenvironmental factors. The results show that environmental factors in addition tosystematic intrinsic properties include natural conditions, social environment, laws andregulations, enterprise management operating environment, and each weight is0.0616, 0.0619,0.2265,0.4411and0.2089. The seasonal index trend model has a greatguidance to mining system safety trend and its fitting degree is0.8149. The weightedMarkov chain prediction model is effective to the forecast of mining system safetytrend.