Application of Data Fusion Theory in Coal Gas Fire Prediction System

In the coal mining production, the fire catastrophe is very dangerous to mining worker life and the whole coal well when the gas explosion is happening. So it is very important to predict fire happening and emit the alarm. In this paper, a novel coal gas fire prediction system is proposed based on multi-sensor data fusion theory. The four different kinds of sensor parameters of the temperature, CO strength, smog thickness and gas concentration are used to predict the fire happening probability. Fuzzy logic characteristic fusion algorithm and the BP neural network fusion algorithm are introduced to reason the fire happening probability. The system is designed with the ARM7 to calculate the fire probability and send out the warning for real time control. The theory analysis and simulation results show that the system has some merits such as celerity capability, reliable and flexible. It gives a new way to raise the correction report rate of the fire detection and extend the application of the existing gas monitoring and control system.