Research On Prediction Method By Seismic Attributes For High-risk Coal Of Coal And Gas Outburst

Coal and gas outburst is one of the main dynamic disasters that seriously affectthe safety of coalmine production in China, and it has many influence factors. Ingeneral, the most outbursts occurred in high-risk coal of coal and gas outburst. On theother hand, the technology of seismic attributes is one of the main means in structuraland lithologic exploration. In this dissertation, I discussed the detecting possibility ofhigh-risk coal of coal and gas outburst after the analyzing on the seismic attributes ofcoal bed reflection. By a case study, I validated the feasibility of this method.By the establishment of a wedge model of coal seam and tectonic coal model, Ianalyzed the relationship between seismic attributes and the thickness of coal seam ortectonic coal. On the condition of thin bed, this relationship is relatively simple. Andthis can be used to predict the thickness of coal bed or tectonic coal by seismicattributes and attributes crossplot. Nevertheless, the numbers of seismic attributes isnumerous, and it is necessary to optimize seismic attributes and select the attributesthat have closest relationship with the thickness of coal or tectonic coal. Thus, on thebasis of above forward modeling data, I established an optimization model of seismicattributes based on GA-BP neural network. In addition, by the actual optimization, Iachieved a good outcome. Moreover, choosing the optimized seismic attributes asinput, I built a thickness prediction model of coal bed and tectonic coal based on FNNand SVR. By the actual prediction of model data and real data, I found that theprediction accuracy of above model is high.On the other hand, not only the thickness of tectonic coal but also the thicknessof coal bed can affect seismic attributes. Thus, I had given a correction method thatcan adjust the influence of coal bed thickness on seismic attributes.Through the actualadjustment of seismic attributes, I found the corresponding relationship betweenseismic attributes and the thickness of tectonic coal was better than before thisadjustment. Then, I optimized the adjusted seismic attributes and the outcomes werechoosing as the input of predicting model. By using of the predicting model of FNNand SVR to predict the thickness of tectonic coal of15#coal respectively, I obtainedtheir prediction and averaged them as the final prediction of tectonic coal’s thickness.After that, by using of a comprehensive index method, many factors that affect coaland gas outburst, such as tectonic coal thickness, the fluctuation of bed’s dip angle, thesmall structure and buried depth, were used to quantitatively forecast the high-riskcoal of coal and gas outburst of15#coal in LunaII mining area. By comparision with at hand geological data, the forecasting results had good consistency.