Study On Early Warning Method For Coal Spontaneous Combustion Based On Multi-information Fusion

Coal spontaneous combustion is very severe in our country, which greatly threatenssafety for coal mines. Early monitoring, identification and warning of coal spontaneouscombustion are important for prevention and control of coal combustion fire. In this paper,according to the results of coal spontaneous combustion experiment, the correlation betweenindex gas and coal temperature is analyzed, the quantitative identification index of degree ofcoal spontaneous combustion is determined, the early warning models for coal spontaneouscombustion are established base on multi-information fusion and pattern recognition method,the monitoring method of diversified characteristic information of coal spontaneouscombustion and early warning method for coal spontaneous combustion base onmulti-information fusion are studied, a multi-information fusion warning system for coalspontaneous combustion is developed.In the large experiment device with15tons coal loaded, the process of coal spontaneouscombustion from room temperature to ignition point temperature is simulated, the correlationbetween index gas and coal temperature is analyzed, the quantitative identification index ofdegree of coal spontaneous combustion is determined, index system of monitoring andwarning for coal spontaneous combustion fire is established; Based on the experimental dataand the observed results in coal mine, mathematical coal combustion early warning modelsare established by means of the fuzzy clustering and pattern recognition, support vectormachine and neural network methods.A combined coal spontaneous combustion monitoring method for goaf of the bundle tubemonitoring and temperature monitoring base on wireless ad-hoc networks is studied. Basedon heat transfer model of coal spontaneous combustion and the gas flow models, the positionof monitors for collecting key characteristic parameters of coal spontaneous combustion is optimized; judgement method for coal spontaneous combustion degree and the position ofsource of self-ignition by means of fusion diversified coal combustion characteristicinformation, temperature and gas, from different monitor system and different site monitoringis proposed; coal spontaneous combustion fire warning system based on multi-informationfusion integrating mine safety monitoring, bundle tube monitoring and temperaturemonitoring with wireless ad-hoc networks is developed, the comprehensive monitoring,identification and early warning for coal spontaneous combustion fire is realized; consideringspecific conditions of the trial mine, a construction program on early warning system for coalspontaneous combustion based on multi-information fusion is made, the construction ofintegration communication network and installation of the system hardware and software arecompleted; applied engineering industrial experiment shows that the acquisiting andcommunicating of all kinds of monitoring data are highly efficient and reliable, operation ofthe system is stable, the key functions and technical specifications of the system meet therequirements of early recognition and warning for coal spontaneous combustion.The research of early warning for coal spontaneous combustion based onmulti-information fusion will greatly promote development of monitoring and preventiontechnology of coal fire.