Research Of Technology For Mine Fire Prediction By Using Smell Identification

Base on the testing about the function and the affecting factors of smell sensors, studying the technique and the method of mine fire prediction by using smell detecting and artificial nerve network analysis in the paper. Make use of it, we distinguished fire smell of the different material from coal, wood, rubble and oil and identified the proportioning of the mixture constitute by them successfully. It introduced the category, function and principle of the smell sensors in the paper. It also described application and development of the method about smell identify in every profession include internal and external. It analyzed the mark gas and it's index of coal spontaneous combustion prediction by studying the characteristic of gases product during coal samples of different degree coal metamorphism program heating. These mark gases include CO, alkenes, the ratio of alkene to alkyl, alkyne, and the ratio of alkyls. Author pointed out their application and it's advantage and disadvantage in the paper. In the paper, studied the influence regulation of humidity, temperature etc. to the smell sensors by using the smell detect equipment. The results are that the humidity influence is linear and the temperature influence is non-linear. Studied the sensitive characteristic of the smell sensors to fire gases, analyzed the absorb curve types and the absorb return curve types of smell sensors to the odor materials. Make sure the experiment condition of smell detecting from combustible by a series of contrast experiment. The experiment condition includes particle size, heating curve and the way of moisture pretreatment. We studied the relation between the smell characteristic curve and degree coal metamorphism by detecting the smell form this coal during heating and oxidation. We analyzed the characteristic curve of fire gases product in comparison with the smell curve and conclude that the smell sensors can catch the small changes of smell from the coal in the early oxidation period. The temperature that predict coal spontaneous combustion can be reduced 20 to 30℃by smell analysis than that by gases analysis. Base on the artificial nerve network analysis, we studied the identified method that we can distinguish the smell from others sash as coal smell, mood smell, rubble smell and oil smell. We can also confirm the ratio of the admixture by this method. We received some correct results. But we can get authentic result only after the temperature at 100℃.