煤矿安全事故分布分形特征和时间序列研究  

Study on the fractal character of safty accident distribution and time series in coal mine

作　　者：肖素丽[1] 谢焱石[1] 李锦鹏[1] 

机构地区：[1]南华大学安全工程,衡阳421001

出　　处：《煤矿现代化》2010年第6期71-73,76,共4页Coal Mine Modernization

摘　　要：本文根据分形理论对我国煤矿安全事故进行分析,研究了煤矿安全事故分布的分形特征以及煤矿顶板事故时间序列的分形特征。通过盒子数计算了我国煤矿2001～2005年之间安全事故的分形维数,数据拟合发现,煤矿安全事故的分布具有双分形特征,即安全事故的分布具有两个无标度区间,且两个无标度区间分界点是110km,说明分界点的两个无标度区间受到不同安全因素的控制。应用R/S分析法对2003～2005年我国煤矿顶板事故时间序列进行分析,发现煤矿顶板事故时间序列具有分形特征,且赫斯特指数H>0.5,这说明顶板事故的发生具有正相关性,通过对顶板事故的统计可以看出,从2003～2005年,我国煤矿顶板事故开始逐年减少,因H>0.5,可以预测未来顶板事故的发生具有减少趋势。In this paper, accident of our country coal mine was analysed based on fractal theory. This paper researched the fractal characters of distribution of coal mine accident and peak accident of coal mine time series. Fractal dimension of accident from 2001 to 2005 was calculated through box amount. Computation result show that the distribution of accident of coal mine are bi-fractal. There are two no-length region and their dividing point is 110km in the distribution of accident. It is showed that the different factor control the dividing point of two no-length region . Through analyzing Peak accident of coal mine time series from 2003 to 2005 by R/S, we found that the peak accident of coal mine time series have fractal charaters, and the Hurst exponent H0.5, the result showed accident have correctitude correlation stating result of peak accident of coal mine showed that the peak accident of coal mine incre reduced annually from 2003 to 2005, through those data we can forecast reduced current of peak accident of coal mine due to H0.5.

关 键 词：分形理论 分形维数 时间序列 赫斯特指数 

分 类 号：TD7[矿业工程—矿井通风与安全]