Application of dynamic weight in coal mine water inrush source identification

Water inrush accidents in coal mining seriously restrict production activities and accurately identifying the source of inrush water can effectively ensure mining safety. Based on groundwater, chemical composition of main aquifer in Pingdingshan coal mine area established a dynamic weight identification model of mixed inrush water source based on variable weight synthesis theory. The results indicate: through Piper trilinear diagram analysis, the pore groundwater sample Q11 of Quaternary system and the groundwater sample P6 of Permian sandstone system with abnormal content of Na + + K + were removed; and determine the state variable weight parameter α of the main aquifer (Quaternary α 1 = − 0.2997, Permian α 2 = − 0.3766, Carboniferous α 3 = − 0.3085, Cambrian α 4 = − 0.2469); construct the variable weight matrix of the water sample, combined with the principle of membership degree, the dynamic weight discriminative model of water source is established. The discrimination results show that the accuracy of the model reaches 100%. Significantly higher than the Piper–EWM water source identification model based on constant weight (76.9% accuracy). This model considers the configuration level of various variables in the water sample, weakens the influence of content extreme value on weight, the weight of the same index will change dynamically between different water samples, get rid of the constraint of fixed weight, reflect the weight of each index in the discrimination process more scientifically, greatly improve the identification accuracy, and provide a strong guarantee for the safety of coal mining activities.