Multi Sensor Data Hierarchical Fusion Application In Coal Mine Gas Warning System

China as a country with extensive resources, coal has been occupying the irreplaceable role in the national production and life. According to the statistics of coal in the energy structure in China accounted for74.29%(according to the Chinese Statistical Yearbook)。The frequent occurrence of coal mine accidents has become a great hidden danger of people influence the stability and well-being, although in recent years the country’s strong investment and research staffs efforts, safety accident overall decline obviously, but there are still rely on a single sensor warning, low accuracy rate of early warning, prone to false alarm phenomenon, not only affects the production efficiency and may lead to coal mine workers drop one’s guard.So this paper is dedicated to construct a comprehensive multi sensor information acquisition with data hierarchical fusion thought fusion in security early warning system model. Taking into account the mine disaster uncertainty and nonlinear problems of attention on intrinsic relatedness accident factors, extracting multi-sensor information. Through the analysis of the general framework of fusion of multi-sensor information information, early warning system structure of multi-sensor based information system, and the data layer, feature layer and decision layer fusion model.This paper presents a hierarchical multi-sensor data fusion technology in order to improve the precision of multi-sensor detection system of early warning, reducing the rate of state unknown or state error during the multi-sensor monitor process. In the data level, using chromatography analysis method to determine the membership degree and the corresponding weights, using fuzzy evaluation method in the feature level and using D-S evidence theory in the decision level data fusion. The experimental mine collected gas, wind and temperature data for experiment. The experimental results show that compare to initial data the membership of security, slight, dangerous states are increased by8.3%,6%,29.2%. The effectiveness of this technology was tested and verified.