Prediction on Surrounding Rock Deformation of Roadway based on a Neural Network Model
The measured data are taken as training samples, and the surrounding rock deformation of ventilation roadway of the 21175 working face in Huopu coal mine is predicted by building a wavelet neural network time series model whose network structure is 4-18-1. The result shows that there is high coincidence degree between the prediction deformation and the field monitored data. The maximum relative error is less than 1%, and the average relative error is less than 0.5%. Therefore, the prediction values provide the theoretical basis of support design and safe production for roadways with the same conditions.