Analysis and Application of Key Factors Influencing the Development of Coal, Rock and Gas Horizontal Wells in the Eastern Ordos Basin

The deep coal bed methane in the Ordos Basin are very abundant, it has enormous development potential and good development prospects. However, the geological characteristics of the Deep Coal Bed Methane vary rapidly in different regions, during the exploration and evaluation phase, renovation experiments were conducted under different process conditions, and the gas well productivity showed significant differences. In order to reveal the main controlling factors affecting the production capacity of the Deep Coal Bed Methane, and fully utilize the latest dynamic and static data of the Deep Coal Bed Methane in the pilot test area, according to the integrated analysis of geological engineering, The coal seam geological characteristics are described in detail, and the production indicators of the producing wells are evaluated in detail. Different geological and engineering factors areitatively evaluated by Pearson correlation analysis, systematic clustering, machine learning and other methods.The results indicate that: ① Under similar regional geological characteristics such as coal lithology and thermal evolution maturity, among the geological factors, thickness and gas content have a significant impact on coalbed methane production. Engineering factors such as the length of coal encountered, total fluid volume, sand addition volume, sand addition intensity are well correlated with production; ② Engineering factors are significantly more correlated with the daily production in the first year than geological factors, and the of geological and engineering composite factors is significantly better than that of single factors. Increasing the well control volume and scale is conducive to high production of gas wells; ③According to the coal seam thickness of 6–10 m and average sand addition intensity of 5.5 t/m in the layer of horizontal well development, the horizontal section length required for a single horizontal well to achieve an EUR of 50 million cubic meters is 1000–1500 m; ④A new method for predicting the production capacity of a single well in deep coalbed methane based on geological and technological parameters created by comprehensively utilizing deep neural network, support vector machine and random forest models, and the coincidence rate of blind well verification using 22 gas wells reached 91%.The research results have important guiding significance for the optimization of the Deep Coal Bed Methane target areas and the formulation of development plans, and also have reference value for the optimization of development technology strategies in other Deep Coal Bed Methane blocks in China.