Construction of Statistical Models of Oil Production Based on Robust Machine Learning Algorithms

Vladimir S. Timofeev,Andrey V. Faddeenkov, Anastasiia Yu. Timofeeva, Arslan V. Nasybullin, Ildar I. Mannanov

The relevance of the study is due to the fact that the effective operation of oil fields nowadays requires a management system based on continuous monitoring and analysis of large volumes of data on various technological parameters. The article aimed at designing of models, which describe the change in oil production volume of wells. The technology of panel data analysis, which takes into account the presence of models structural heterogeneities on the use of robust parameter estimation algorithms, was selected as a main approach. The paper presents oil production models comparative analysis results from an autoregressive model of panel data together with a model of structural changes with the use of various algorithms for estimating unknown parameters. The analysis results reflected in the article are of practical use for specialists in oil field management.

Volume 12 | Issue 6

Pages: 2702-2707

DOI: 10.5373/JARDCS/V12I6/S20201230