Study On The Characteristics Of Gas Occurrence In Middle And High Rank Coal And The Technology Of Risk Prediction

Abnormal gas emission is one of the main causes of gas overrun accidents,generally speaking,It is also regarded as pne of the omens of gas outburst and other gas disasters.There are many factors that can lead to abnormal gas emission,coal gas occurrence characteristics,stress,strain,pressure and other external interference are the main factors of abnormal gas emission.Because of its complex mechanism of action,some scholars have done a lot of research,the mechanism of action is still unclear.In this paper,the characteristics of pore structure of coal are studied as the breakthrough point,it focuses on the gas occurrence characteristics and the influence of gas occurrence in gas outburst risk.By tested in the laboratory,the pore and fissure development laws of coal samples of different coal rank are systematically analyzed,the fractal characteristics of their structures are analyzed,the effects of internal factors such as industrial composition,metamorphic degree,pore size distribution and fractal dimension of coal and external factors such as pressure and temperature on gas adsorption and desorption characteristics of coal are explored,the gray analysis model is established by using the influencing factors of coal gas occurrence,a deep learning intelligent predictive system for gas outburst risk is designed.The whole study provides necessary theoretical support for the prediction of gas outburst risk,the main research results are as follows:(1)The distribution laws of different metamorphic degrees,pore size distribution,specific surface area,pore volume,pore structure type,pore connectivity and pore fractal dimension of coal are systematically revealed.The study found that,the specific surface area of micropores accounts for the main part of the total specific surface area,the volume of macropores and fractures occupies the main body of the total pore volume.With the increase of coal rank,the average aperture decreases gradually,the pore volume decreases gradually,the connectivity of pores gradually becomes worse,the specific surface area increased first and then decreased.Coal with different surface characteristics has different pore structure types,the pore structure of the dim middle rank coal is mainly composed of cylindrical pores open at both ends,the pore is blocked by impurities,medium rank coal with multiple bright layers,its pore structure is mainly composed by small holes and small holes in the shape of ink bottles with thick necks,the pore structure of high-grade coal with multi-layer bright layer is mainly formed by small holes and micropores in the shape of ink bottle with thin neck,the pore structure of bright high rank coal is mainly consist of cylindrical holes and wedge holes with small pore diameter.The fence method is innovatively used to evaluate the interval of pore fractal dimension,it is found that the pore roughness of medium and high rank coal is higher than that in medium and small rank coal,the micropore roughness of high rank coal is the highest.(2)Exploration of gas occurrence form in coal pore structure,through quantitative analysis of micropore structure of coal samples with different metamorphic degrees,the limit adsorption equilibrium state of gas in coal is studied by using micropore filling theory,single molecular layer adsorption theory and multi molecular layer adsorption theory,a quantitative method was established to characterize the gas adsorption capacity of coal samples,The occurrence characteristics of gas molecules in different scale pore structure of coal reservoir are obtained.Research shows,the main occurrence mode of gas molecules in coal pore structure is adsorption state,micropore filling is the main mode of adsorption,the adsorption mode of multi molecular layer is complementary,in addition,there is a certain amount of monolayer adsorption.(3)The influence of internal factors,such as industrial composition,metamorphic degree,pore characteristics and fractal dimension of coal,and external factors such as pressure,temperature has been studied as gas occurrence characteristics.The results show that,the changing direction of fixed carbon content,metamorphic degree,total specific surface area,micropore specific surface area,micropore fractal dimension and pressure of coal samples are positively correlated with the changing direction of gas limit accumulation,the changing direction of ash content,volatile content,moisture content,temperature,particle size of coal samples are inversely correlated with the changing direction of gas limit accumulation.The grey correlational method suitable for small sample system and poor informational system is used to model and calculate the influence effect of each factor,according to the correlational coefficient between factors and the limit value of adsorption,the order is from large to small,proportion of specific surface area of micropores,particle size,temperature,fractal characteristics of micropores,moisture content,fixed carbon content,pressure,ash content,total specific surface area,coal rank and volatile content.According to the theoretical research and using the design method of artificial intelligence,a deep learning intelligent predictable system of coal seam gas outburst risk is constructed.The neural processing unit makes the intelligent BP artificial neural network model have the remarkable advantages of self-learning,self-training and errorcorrection.The momentum factor is introduced and the batch processing method is used to optimize the algorithm of the model.The key influencing factors are obtained by the grey correlational analysis to optimize the index of the model.After the optimization,the gas occurrence characteristic data and gas outburst risk cause data of 30 groups of coal seams are respectively input into the predictable system,the deep learning intelligent predictable system can carry out intelligent learning and training according to the sample data.The operational calculation are as follows,the outburst risk predictive system using coal gas occurrence characteristic data as input layer index and the outburst risk predictive system using gas outburst risk causal data as input layer index,the two systems have similar predictive accuracy and similar running time,it is verified that the deep learning intelligent predictive system of outburst risk,which takes the coal gas occurrence characteristic data as the systematic index,is a predictive technology that can be successfully applied in the engineering practice.Figure[59]table[19]reference[126]...