Research On Forecasting Method For Spontaneous Combustion Of Coal Seam Existing CO Gas Based On D-S Theory Fusion

Our country has been severely suffered by coal spontaneous combustion disasters nowadays. Coal spontaneous combustion not only causes the resource wastes and brings the mine enterprises huge economic losses directly, but also seriously threatens life safeties of enterprise staffs, destroy mine equipment and facilities, and endanger the surrounding natural environment. Therefore, timely and accurate forecast of coal seam spontaneous combustion source is particularly important to the safety production of enterprises. CO gas is a key index for coal spontaneous combustion forecasting. However, if CO gas exists in coal seam itself, CO gas emitting outward in the process of mining will interfere with forecasting work of mine coal spontaneous combustion.Based on the problem mentioned above, this paper takes Linnancang Mine with Kailuan Group as the research object. As for the argument whether CO gas exists in the mine coal seam, by field and laboratory experiments verification that CO gas occurs in the coal seam, the affecting factors of CO gas occurrence causes are analyzed. On this basis, this paper aims to put forward CO as the core indicators, supplemented by coal temperature and C₃H₈,C₂H₄ indicating gases. Coal spontaneous combustion pre-warning index system based on Basic Probability Assignment ?BPA? will be established by extracting the information characteristics of monitoring data. By using multi-source information fusion of D-S ?Dempster/Shafer? evidence theory, the spontaneous combustion fore of coal seam with CO occurrence will be realized.1. In-situ drilling and gas-taking experiment and coal sample vacuum desorption experiments will be adopted. Cointegration theory will be used to prove native CO gas occurrence in the coal seam. And its influence on CO gas occurrence cause of Linnancang Mine will be analyzed from the aspects of mine geology structure, coal macromolecular structure, adsorptive property of coal petrography, permeability of coal petrography and pore structure characteristics of coal petrography, etc,.2. By using intelligent coal heating oxidation tank, and the effective simulation of mine spontaneous combustion process, the corresponding relationship between gas index parameters and coal temperature are summarized and analyzed in this paper. The gases including CO, C₂H₄, C₃H₈ are selected to be signs for coal combustion forecasting. Based on the change characteristics of CO gas concentration, the early stage of coal spontaneous combustion are divided into five phases, including the occurrence gas desorption phase, low oxidation crisis phase, moderate oxidation crisis phase, high oxidation crisis phase and extremely high oxidation crisis phase. By adopting Fuzzy Membership degree and IS interval set, coal spontaneous combustion forecasting system is built on the basis of Basic Probability Assignment ?BPA?.3. By using multi-source information fusion D-S evidence theory, spontaneous combustion D-S evidence theory fusion forecasting simulation verification is carried out in terms of coal mining working face and goaf respectively. This paper proposes an evidence conflict measurement method based on the relative difference to deal with the evidence conflict on multi-source information fusion. The simulation result shows that the forecating methods for coal spontaneous combustion presented in this paper ensure the effective retaining of valuable information through feature recognition. Multi-source fusion increases the accuracy of the forecating result, and at the same time provide effective decision-making basis in order to adapt to the complicated condition of underground mine.