矿井瓦斯涌出量时间序列的分形特性分析

应用分形论中的R/S分析法对矿井瓦斯涌出时间序列进行分析。计算了时间序列的赫斯特指数、概率分布分形维数和轨迹分维数。讨论了它们在矿井瓦斯涌出预测中的指导意义。计算结果表明正常情况下矿井瓦斯涌出量的变化趋势具有很好的持久相关性，可以用分形论对其作出趋势预测；而发生火灾时，持久相关性不明显。但两种状态下的概率密度分布都满足分形分布；时间序列轨迹都具有明显的分形特性。

Fractal feature analysis of time series of gas emission in coal mine

In this paper, with the concept of rescaled-range(R/S) analysis of fractal theory ,we analyzed the fractal feature on the time series of gas emission in coal mine, and computed the Hurst exponent H,fractal dimension of probability distribution and the curve on the time series and discussed the theoretical significance in prediction of gas emission. According to the result of the analysis above, we can predict the variable tendency of gas emission when system is in normal state. But it is difficult because the system is more complex in fire state than in normal state. But probability distribution satisfies fractal distribution and the curve is a fractal in two states.