煤系岩体参数概率分布及Bayes优化

Probability distribution fitting and Bayes optimization of rock mass parameters of coal measure strata

正确确定煤系岩体力学参数的概率分布是松散含水层下采煤顶板稳定可靠性分析的关键。以淮南矿区部分煤矿8煤顶板砂岩强度参数为例，以 K － S 检验法对数据进行概率分布拟合，获取参数分布类型及统量。以此大样本概型为先验函数，，具体工作面顶板岩体小样本参数的概型为似然函数，基于 Bayes 方法对其优化，得到验后分布的概型参数。计算结果表明，8煤顶板砂岩强度参数全部接受正态分布和对数正态分布，优化后的方差有所下降，可以提高覆岩采动稳定可靠性分析结果，从而达到优化目的。

The correct determination of probabilistic distribution mechanical parameters of rock mass is the key to analyzing the roof stability in coal mining under the confined aquifer. Take the coal roof＇s sandstone strength parameters of No. 8 coal seam in Huainan mining area for an example, the probability distributions of data distribution are fitted by the K - S test to obtain the distributing style of parameters and statistics. Take the large sample projective scheme as prior function, the small sample projective scheme of concrete roof undulation rock mass as likelihood function and optimize it using Bayes method and then obtain the paramet.ers of the post distribution. The calculated results indicate that the coal roof＇s sandstone strength coefficient of lower coal group are all distributed normally and follows lognormal distribution, the analysis results＇ reliability of roof stability can be improved as the lower optimized variance so as to achieve the goal in the optimization