Research on the Prediction of Natural Gas Single-Well Operation Cost Based on Machine Learning

DOI: https://doi.org/10.62517/jike.202504301

Author(s)

Zhuohan Guo1, Ying Tian1,*, Yuhan Ma1, Yinan Chen2, Lin Liang2

Affiliation(s)

1Exploration and Development Research Institute, PetroChina Southwest Oil and Gas Field Company, Chengdu, Sichuan, China 2Key Laboratory of Energy Security and Low-carbon Development, Southwest Petroleum University, Chengdu, Sichuan, China *Corresponding Author

Abstract

In order to overcome the current problem of large amount of data required in the measurement of single well development and operation cost, the calculation process is complicated, and it is difficult to make fast and accurate prediction. By searching for many factors affecting the operating costs of natural gas wells, six machine learning algorithms, including linear regression, decision tree, AdaBoost, XGBoost, Support Vector Machine (SVM) and Artificial Neural Network (ANN), are used to construct a model to verify the usability of machine learning algorithms in the prediction of operating costs of single wells for natural gas development. Combined with the actual sample data, it is found that most of the machine learning models run with a high degree of fit to the real data, among which the elastic network regression model in the linear regression model has a prediction accuracy of 98%, which demonstrates the model's superiority in the prediction of natural gas single-well operating costs.

Keywords

Machine Learning; Single-Well Operating Cost; Cost Components; Prediction Method; Elastic Network Regression

References

[1] Ting-ting Wang, Cheng Huang, Qi-chen Zhang. The peak of China's oil production forecasting and strategies research. China Mining Magazine, 2015, 24(12):38-44. [2] Fang-zhou Zhang, Hu-yong Yan, Li-quan Yang, et al. Application of Improved Grey Neural Network Model to Oil Yields. Computer Technology and Development, 2013, 23(06):241-244+248. [3] Dan Jian, Xi-yu Ren, Juan Yue, et al. Grey correlation analysis of oil and gas operating cost components. Economic & Trade, 2017, (14):116. [4] Yan-ping Wang. Oil and Gas Operating Cost Prognostication Research of Da Gang Oil Field Company. China University of Petroleum, 2007. [5] Shang-bin Dong. An empirical study of oil operating cost forecasting. Oil-Gas Field Surface Engineering, 2009, 28(10):21-23. [6] Feng Li, Xiao-hui Zhang, De-bin Qu, et al. Prediction of operating costs in water-driven oilfields based on principal component regression modelling. Journal of Oil and Gas Technology，2012，34(9)：136-139. [7] Ping Li, Qiong Mao, Xin-ying Wang, et al. Establishment and Application of the Operation Cost Prediction Model Based on Multivariate Linear Regression Analysis. Petroleum Planning & Engineering, 2018, 29(03):33-37+52. [8] Wu Chen, Tao-hong Wu, Chen Chen, et al. Prediction of Oilfield Operation Cost Through Partial Least Squares Regression-A Case Study on DX Oilfield. Journal of Southwest Petroleum University (Social Sciences Edition), 2019, 21(01):8-13. [9] Yan Hu, An-qi Gao, Yu-lu Tian. Discuss Determination Method and Application of Oil and Gas Operating Cost in Petroleum Enterprises. Modern Industrial Economy and Informationization, 2024, 14(07): 187-189+194. [10] Xi-xi Wu, Chun-hua Hou, Wu Chen, et al. Study on Combined Forecast Method for Operating Cost of High Water-cut Water Flooded Oilfield. Technology &Economics in Petrochemicals, 2014, 30(06):5-9. [11] Qing Zhou, Cheng-zuo Lin, Rong Li. Support Vector Machine Based Classification Prediction Method for Operational Costs. Petroleum and New Energy, 2016, 27(03):16-18. [12] Jin-rong Feng. Research on Data Analysis Method of Single Well Production and Operation Cost Based on Neural Network. Xi’an Shiyou University, 2020.