Research On Gas Monitoring Method Based On Cloud Model And D - S Evidence Theory

In recent years, mine accidents are frequent occurrence in China. Among them, the coalmine gas explosion accident is the highest frequency of occurrence and the most serious damageof the disaster, the reason of gas explosion has many factors, the imperfect monitoring system inthe pit and the lack of scientific identification and elimination system of gas explosion accident isone of the main causes of the accident. So through the study of coal mine gas monitoring toeffectively avoid the high risk accidents and increases accident precaution and safety productionefficiency has a very important practical significance.This paper tries to realize the state evaluation and decision of the coal mine gas monitoringand improve the accuracy rate of prediction by the method of combines the multi sensorinformation fusion technology, cloud model and D-S evidence theory. In coal mine gasmonitoring system, we can set up corresponding risk degree of adjectives according to the safetyspecification for all kinds of coal mine gas monitoring state, and set up cloud model according tothe danger rank descriptor corresponding to the upper and lower limits of the detectionparameters, extraction of dangerous degree of signal from the feature level fusion results and thesensor information to tectonic the basic probability assignmen（tBPA）of evidence synthesis, andrealize the decision level fusion by using the improved D-S evidence theory method.Aiming at the limitation of traditional D-S evidence combination method, we puts forwardthe improvement method, comparative study the algorithm and characteristics of several evidencesynthesis rules, in-depth study the normalization and astringency of the continuous sub-stepmethod of D-S evidence combination, we puts forward several theorems and corollaries of D-Sevidence fusion, and gives the corresponding proof. For processing the high conflict evidence, weput forward a new synthesis method of the mixed using of D-S and the weighted average. Finally,this paper tries to use time series analysis method to forecast the future security situation of thecoal mine.The simulation results show that: The gas monitoring method based on cloud model and D-Sevidence theory which proposed in this paper, the decision results satisfy the normalizationcondition and has good convergence, which can effectively improve the accuracy ofdecision-making.