Research and application of the underground fire detection technology based on multi-dimensional data fusion

Abstract

Underground mining has been plagued by gob fires triggered via self-igniting coal, the emission of toxic fumes, particulate matter, and possibly induced explosions from the gob fires make mining extremely dangerous. The spontaneous combustion of underground coal seriously affects the mining of adjacent coal seams and the overall safety production. The key to control the shallow coal seam fire is to accurately locate its the location and scope by surface detection method. Based on the non-dimensional normalization processing of different types of data and the weight analysis of analytic hierarchy process (AHP), a multi index data fusion method of temperature, gas and radon concentration was proposed by using the practical radon measurement, borehole temperature measurement and gas measurement methods, and the index of fire zone delineation (IFZD) based on multi-dimensional data fusion was obtained. According to the content total method based on fractal theory, the anomalousthreshold of IFZD was determined by piecewise linear fitting in double logarithmic coordinates. Eventually, the multi-dimensional data fusion underground fire detection method was established completely, and the method was applied to the fire detection of shallow coal seam in typical integrated mine. The results showed that the weight coefficients of the detection indexes (CO, CO2, SO2, temperature and radon) in the fire area were 0.12, 0.03, 0.05, 0.12, 0.68, and the lower bound of the anomaly of IFZD was 0.29. According to the comprehensive isoline map, abnormal area plan and stereogram drawn through IFZD and its outlier threshold, there were eight areas with the total area of 27061 m2, which developed from the abandoned well, main well location to the southwest and northeast. The isolines of abandoned air shaft and main shaft were densest and the value was the largest, indicating that coal oxidation in this area is severest, so it is necessary to strengthen the ground fracture investigation and sealing. This method solved the problems such as poor accuracy of single index circle fire zones, and difficulty in direct fusion of multi-index data with different dimension and range, and difficulty in determining the abnormal lower boundary of the fusion index after dimensionless normalization, which provided a new technical method support for the detection of underground shallow coal seam fire area.