Abstract:

The first, a BP net is built and trained with six input variables of burial depth of coal seam coal seam thickness gas content the distance of coal working face and adjacent coal seam daily advance distance average daily output and the absolute Gas emission as output variables of the neural network model. Then the Monte Carlo method is tried to predict the development trends and the input variables change with time behavior simulation by random sampling of 6 groups input variables, the simulation results are used as the input layer node, then the output value is the forecast of next production cycle Gas emission. The results showed that this prediction method has a high precision and this prediction method can be a good guide for the production of mine Gas control work.

Key words: BP neural network, Monte Carlo method, The Prediction of Gas emission