Forecast the Dangerous Level of Gas Outburst Based on GASA Neural Networks in Coal Mine

To decrease casualties and actualize unmanned roadway, the neural nets based on hybrid learning scheme of genetic algorithm and simulated annealing (GASA) with output and hidden weight optimization is put forward to forecast the coal and gas outburst in roadway face. Although the neural net based on OHWO does better than others in the convergence, it has the instinctive shortcomings of local minima and oscillation problem. GASA combines the ability of evolution of GA and probability searching of SA. The proposed neural nets based on GASA with global search ability can simultaneously optimize weight and threshold of artificial neural network. The improved NN can accurately capture the complicated relationships among feature values of coal-gas outbursts and dangerous circumstances. Results show that it is a highly feasible and creative technique to forecast coal-gas outburst in roadway face.