Application study of a water conducting fracture zone height prediction model based on bp neural network optimized by genetic algorithme

It is important to accurately predict the height of the water conducting fracture zone (WCFZ) for the production safety of coal mines and ecological environmental protection. Through the field measurement data of WCFZ at working face in Shendong and neighboring mines. The correlation degree between the working face length, mining height, mining depth, coal seam dip angle, mining speed, hard rock proportionality coefficient and the height of WCFZ was analyzed by using gray relationship analysis theory. The BP neural network model optimized by genetic algorithm was trained, and the predicted WCFZ values of empirical formula, GA-BP and BP neural network model were compared with the actual measured values of WCFZ in the mine area. The comparison analysis showed that the GA-BP neural network model had the smallest error of 7.78% in the prediction value. Therefore, the GA-BP neural network model has sufficient prediction accuracy and generalizability and can be applied to study the WCFZ in the Shendong mine area.