中文摘要：

针对煤矿矿井突水对煤矿安全生产的不利影响,基于水化学分析结果,建立多元混合模型模型,分析计算矿井突水水样,并与传统的BP神经网络原理和模糊综合评判法相对比,分析结果与矿山实际情况相符合。结果表明：多元混合模型不仅能准确分析出矿井突水水样主要来源,而且计算简单准确,受水化学分析样本量和离子种类数量限制较小,可以作为一种矿井突水水源判别工具在工程实践中应用。

英文摘要：

In order to figure out the source of bursting water in an accurate manner, the multivariate mixed models theory and hydro- chemical analysis were used to build the multivariate mixed models for aquifer and bursting water source. By combining the built equations, numerical analysis was taken for the bursting water source. The gained result was compared with those gained by the BP neural network method and the fuzzy comprehensive evaluation method. The comparison shows that these results are basically consistent with each other, proving that the multivariate mixed models can well be counted upon as a reliable tool. The tool is able to accurately figure out the source of coal mine water invasion and the calculation process is simple and accurate; it is less restricted by water chemical analysis of sample size and ionic species number, thus making a great contribution to the improvement of the safety production of coal mine.