Multiple Type Discriminating Mine Fire Sensors

It was determined that a selection of different types of fire sensors could be used to discriminate mine fires from nuisance emissions produced by diesel equipment. A neural network (NN) was developed for application to coal, wood, and conveyor belt fires in the presence of diesel emissions and evaluated with the successful prediction of twenty-two out of twenty-three mine fires based upon a fire probability determination. The optimum sensor selection for the NN was comprised of a carbon monoxide (CO) sensor, two different types of metal oxide semiconductor (MOS) sensors, and an optical path smoke sensor.