Fracturing Effect Evaluation Of Lateral Drilling Horizontal Well In Sulige Gas Field

The contradiction between supply and demand of natural gas resources is becoming increasingly serious in our country,speeding up the development of tight gas reservoirs is of a strategically important significance to ensure the national energy security and sustainable development of natural gas.With the rise of artificial intelligence,machine learning method has been widely used in the field of oil and gas field development,and has achieved a lot of research results.In this study,for gas Wells of different geological types,based on the main controlling factors of the side-drilling fracturing effect,the machine learning method was used to establish the post-pressure productivity prediction model,and the response surface method(RSM)was introduced to optimize the side-drilling fracturing process parameters of low production and low efficiency gas Wells,so as to propose the corresponding stimulation measures.Firstly,39 laterally drilled fractured gas Wells in Sulige gas Field were taken as the research object.The linear function normalization(Min-Max)method was used to standardize and reduce the dimensionality of the original data set composed of 11 factors of geology and fracturing technology.From the geological point of view,four parameters such as the number of sand bodies drilled in a single well,effective sand body thickness,permeability and gas saturation were selected as evaluation indexes.The entropy weight-ideal point method was used to establish an evaluation model,and the variation rate of proximity degree was introduced to divide gas Wells into class I(11),Class II(24)and Class III(4).At the same time,considering the lower limit of economic index of gas well productivity after side drilling fracturing,9 low production and low efficiency gas Wells were selected.Secondly,based on the geological classification results of gas Wells,the improved grey correlation theory is used to determine the main controlling factors that affect the fracturing effect of different types of gas Wells.Among them,the main control process factors of class I,II and III gas Wells are respectively the number of equivalent pressure sections,fracturing fluid volume,horizontal section length,fracturing fluid volume,horizontal section length,equivalent pressure sections,horizontal section length,fracturing fluid volume and sand addition volume.Then,taking different types of gas Wells(except low production and low efficiency gas Wells)as the research object,based on the main control process factors of sidetrack fracturing effect,GM(1,N),support vector machine of grid search method,support vector machine of particle swarm optimization algorithm and other mathematical methods were used to establish the productivity prediction model of sidetrack fracturing.Through comparative analysis,it is found that support vector machine(SVM)based on particle swarm optimization algorithm has the highest calculation accuracy,and the determination coefficient R~2 is 0.9012.Finally,eight low-productivity and low-efficiency gas Wells in classⅠandⅡgas Wells were taken as the research object.The support vector machine productivity prediction model of sidetracking fracturing based on particle swarm optimization algorithm was introduced to analyze the influence of the interaction between the main control technology factors of sidetracking fracturing effect on the productivity after sidetracking fracturing,and the sidetracking fracturing process parameters of low-productivity and low-efficiency gas Wells were optimized.Corresponding stimulation measures are proposed to provide technical support for further efficient development of the gas field.