Research On The Fractal Characteristics Of Temporal-spatial Distribution Of Coalmine Accidents In China

To effectively control relevant influencing factors for reducing coalmineaccidents, research on those accidents as to grasp the law of accidents happening andthe causing mechanism have been carried out. The characteristics of randomness,nonlinearity and dynamism as the accidents processes are influenced by someuncertain factors determine that researching on accidents would be quite sophisticated.The fractal theory is a newly developed subject in recent year, and this theory hasbecome a powerful tool for describing nonlinear problems. Analysis and discussion ofthe fractal characteristics of temporal-spatial distribution and relativity as wellregularity on temporal distribution of coalmine accidents from2001to2010in Chinahave been made by using relevant methods of fractal theory in this thesis. The mainresearch contents and results are as follows:(1) Concrete analysis of China’s coalmine accidents fractal characteristics ofdifferent time domains, regions and accident types from2001to2010have been madeby using the geographic information processing software. The results show that thefractal characteristics of China’s coalmine accidents of spatial distribution areobjectively existent; the fractal dimensions of each season are respectively1.0071,1.0054,0.9805and0.9658, and the complexity of accidents’ distribution decreasefrom spring to winter successively; the fractal dimensions of Shanxi, Gansu, Hunanand Guizhou are respectively1.1894,0.9025,1.2670and1.2693, showing that miningtechnology conditions are in accordance with the fractal dimensions of coalmineaccidents; the fractal dimensions of roof, gas, transportation and flood damage arerespectively1.0527,1.0297,0.8256and0.8189, and the complexity of aboveaccidents decrease successively.The fractal characteristics of coalmine accidents of spatial distribution are ofobjective existence through finding the power law relation between scale r and gridnumber N(r). Not big differences of those fractal dimensions of different time domains, regions, accident types are found, and they are close to those of the wholefractal dimension value. It shows that each spatial distribution continues the wholespatial fractal characteristics, they share much of similarity.(2) Analysis of the singularity strength of each partition and the rangeability ofeach element in the multi-fractal set. As to the multi-spectrum f(a)-a, the range ofparameter a: Δa=7.8. This shows that the singularity strength of partitional coalmineaccidents’ spatial distribution is of a large value and the spatial distribution ofcoalmine accidents is irregular, and the relevant Δ(a)=0.3082>0shows that theprobability of accidents’ increasing is small comparing to their reduction in eachpartition.(3) The power law relation between the scale of death toll and the frequentnessof accident number in each region’s coalmine accidents has been manifested, so it canbe concluded that coalmine accidents are with self-organized criticality, backing upthe existence of fractal characteristics of coalmine accidents.(4) The time series correlation dimension and effective freedom degree ofcoalmine accidents are calculated by applying phase space reconstruction methodbased on G-P algorithm, and the values are5.2922and10respectively. It reveals thatthe phase space’s fractal dimensions of coalmine accidents in China trend to be afinite value. Besides, the embedded dimension corresponding to saturation correlationdimension reflects the upper bound of the variable number when modeling dynamicalsystem of coalmine accidents, and providing scientific basis for reasonably selectingvariable number.(5) The Hurst index H of time series of coalmine accidents is obtained byadopting the method of rescaled range analysis (R/S Analysis), and H=0.9005>0.5. Itproves that the time series of China’s coalmine accidents possesses positivecorrelation long-range dependence with a relatively significant sustainability.