Advanced pattern recognition based on neural network applied in coal structure

The coal and gas outburst disasters could be forecasted in time by recognizing the coal structure types. In order to classify the coal structure, an advanced pattern recognition that combined ultrasonic reflection with BP neural network was put forward. The patterns were sorted and recognised based on a reasonable consideration of ultrasonic speed, ultrasonic attenuation coefficient, characteristics of ultrasonic transmitting, ruggedness coefficient and other parameters relating to types of coal structure. Results demonstrate that the advanced coal structure pattern classification can distinguish coal structure types effectively. It is significant for the advanced coal structure pattern classification to forecast disaster of coal and gas outburst.