Research On Prediction Of Floor Water Inrush Based On GWO-Elman Neural Network

China is the world's largest coal producer and consumer.However,due to complex hydrological and geological conditions and large mining depths,mine accidents are endless.In particular,water inrush accidents in coal floor will usually cause serious casualties and property losses once they occur.Therefore,the prediction of coal mine floor water inrush has become the focus of the coal industry research.Hongqi Coal Mine is a typical North China coal field.The average thickness of the 3 coal seams as the main coal seam in the mine field is 5.48 m,accounting for 64.3%of the total thickness of the recoverable and partially recoverable coal seams.The coal seam is seriously threatened by floor water inrush,and floor water inrush occurred many times during mine construction and production.Therefore,predicting the water inrush from the bottom of its 3 coal seams is not only of great significance and high value for the safety production of Hongqi Coal Mine,but also has certain guiding significance for the safety production of North China coalfields that are also threatened by water inrush from the bottom.The effective thickness of the water-resisting layer is one of the influencing factors of water inrush.In order to obtain its value,firstly,on the basis of analyzing the relevant geological data of the coal mine,using the formula formula method,FLAC3D to establish the numerical model method and the on-site measurement method to predict the depth of the floor damage And get the value of the floor depth of the working face of the mining area.Then use the obtained value as the value for calculating the effective thickness of the water-repellent layer,and finally calculate the value of the effective thickness of the water-repellent layer.On the basis of determining the main influencing factors of coal floor water inrush,the relevant data are input as input samples to the gray wolf optimization algorithm to obtain the optimal weight and threshold for Elman neural network optimization,and then use the obtained data to establish a GWO-Elman neural network-based Prediction model of floor water inrush.After that,the test sample is input into the model for verification,and the result proves that the accuracy of the model is high.Finally,the model was used to predict the floor water inrush from the unmined face.The prediction results can guide the actual safety production of the mine.