Research On Prewarning Of Mine Fire Disaster In Coal Mine Goaf Based On Information Fusion

Coal is an important resource of our country, but also a high-risk industry, coal mine fireaccident-prone, have caused significant damage to the national economy and security. The vastmajority of coal fire occurred in the mined-out goaf one can not look directly at or reachhidden locations, coal mining goaf fire warning authenticity, reliability and timeliness directimpact on the production order and economic efficiency of enterprises. Research coal mininggoaf fire warning has important practical significance.In this paper, based on analysis of the safety factor of the coal mining goaf, therelationship of the various safety factors and the fire broke out a more comprehensive study,especially in the coal mining goaf the relationship between temperature and the fire broke outin-depth study. Mainly include coal mining goaf based on safety information fusion firewarning information collection, pre-processing, information fusion and decision-making ofthree parts. Information pretreatment de-noising method based on wavelet transform of thenoisy signal feature extraction, low-pass filtering, signal reconstruction, signal de-noising.Proposed the fusion prediction based on neural network and DS evidence theory is acombination of two fire method, the method extracts the average rate of change of values andthe cumulative value of all safety factors characteristic value, the use of improved BP neuralnetwork partial information fusion, basic probability distribution function to construct anindependent theory of evidence, judgment and DS evidence theory based on the weightdistribution on multi-feature fusion, a coal mining goaf fire warning, to solve a single safetyfactors described goaf environment more one-sided defects. Written in Matlab platformsimulation software, simulation results show that: the proposed multi-feature two informationfusion in decision-making can better predict coal mining goaf fire warning, a more accurateearly warning capability and fast prediction speed. In this paper, based on information fusion coal mining goaf fire warning after asimulation test results show that the real-time early warning gobs of spontaneous combustionstate, so that disaster prevention, mitigation, in reality, has high reliability and practicality.