The Model Research On The Recognition Of Water Inrush With Using The Water Chemical Composition In Hebi Mining Area

Water damage is always one of the major threaten in the safe instruction of a colliery,so it is a premise of controlling the type of water inrush exactly and quickly in the work of the protection of mine water.Based on the source of inrush water in Hebi mining area,we opened the water quality survey and collected 79 groundwater samples in each water-filled aquifer and the goaf water samples, and finally we summarized the water chemical characteristics of each water-filled water.Then,with using the method of Principal Component Analysis,we determined the “TDS principal component”“Na-Cl principal component”,“Mg-HCO3- principal component”,which can reflect the water features in a number of indicators,on this foundation,we set up the model of Grey correlation,Fuzzy Math,BP neural network,Bayes discriminant,Fisher discriminant to recognize the water inrush which are based on the water quality,and 79 water quality testing datas collected in 2013.Recognition results show that Fisher discriminant model and BP neural network had highest rate which were 84.6%;Grey correlation model and Bayes discriminant model are the second;Fuzzy mathematical model's recognition rate is low.After compared with each other,we choosed the Fisher discriminant model as the mathematical models to identify the sources of water inrush,and its recognition rate can reach more than 80 percent,it is simple to operate and easy to learn,so it has been already adopted by Hebi mining area. This conclusion has been adopted by we undertake research projects of Based on automatic recognition of mine water inrush water quality water software research and development of Hebi mining area.