Evaluating the Dominant Factors Affecting Shale Oil Productivity Based on Machine Learning——A Case Study of West Area 233,Qingcheng Oilfield
The Ordos Basin is rich in shale oil and gas resources and has great development potential.Based on concept of geological engineering integration,this paper uses machine learning technology to construct a multiple linear regression model of the cumulative oil production of shale oil horizontal wells and their geological and engineering parameters for a total of 55 horizontal wells on 15 platforms in the West area 233 of Qingcheng Oilfield.The joint influence effects between various factors are analyzed in depth.Research has found that the cumulative oil production of a single well is influenced by a combination of geological and engineering factors,and the degree and direction of their impact on production vary dynamically with the development process.After about 15 months of production,the production status of the oil well tends to stabilize.At this time,the constructed multiple linear regression model can predict the single well production capacity well and meet the accuracy requirements.Among the geological factors,the drilling encounter rate has the most significant impact on oil production,followed by reservoir length,oil saturation,and formation coefficient;The engineering factors are most affected by the sand addition intensity,followed by the number of fracturing sections,fracturing cluster density,displacement,and inflow intensity.This study clarifies the changes in the impact of geological engineering factors on production under different production stages,analyzes the mechanism of single factor impact,and optimizes the main control factors,providing important theoretical basis and practical guidance for optimizing fracturing process parameter configuration and accurately predicting single well production capacity.
Qingcheng shale oilmachine learninggeological engineering parametersmain control factor analysis
NETL
NSTL
万方数据
