106 / 2021-11-30 22:46:57

Identification model of mine water bursting source based on multi-attribute data mining theory and its application

water bursting in mine; water source identification; unascertained measurement identification; intelligent recognization

（1）资源开发与利用 > 11. 矿井水害、污染防控与矿井水资源化利用

Due to coal resource starvation for Carboniferous-Permian Period in eastern-central China, the focus for coal mining has been gradually shifted to the Jurassic Coalfields in Ordos Basin, midwest China. The roof of coal seam in Ordos Basin is weakly-cemented, in which the pores and fractures are widely developed, and with strong water abundance for aquifers in overlying strata, water inrush from roof often occurs. So how to accurately identify the water bursting source has always been an urgent problem in the field of preventing and controlling mine water disasters. Groundwater dynamic monitoring and hydrochemical characteristics suggest that when water inrush occurs, indicators such as water level, water inflow, water temperature and the ratios between characteristic elements usually show precursory abnormal changes. Taking the roof water inrush occurred in Guojiahe coal mine of Shanxi Province when the No. 3 coal seam (Jurassic Period) mining as the research background, water level, water inflow, water temperature, the ratios between characteristic macroelement, the ratios between characteristic trace elements and the ratios between characteristic stable isotopes are selected as identification indexes, datamining in the variation characteristics of precursory abnormal factors when water inrush occurs, and then the identification model based on rough set theory and unascertained measurement function, intelligent recognition model based on AWOA-ELM were constructed, besides, the contribution rates of each potential water source mixed in mine water bursting source were quantified, which provides a new method for accurate identification of mine water bursting source.