Application of machine learning models in predicting initial gas production rate from tight gas reservoirs

Ugwumba Chrisangelo Amaechi, Princwill Maduabuchi Ikpeka, Johnson Ugwu, Ma Xianlin

Research output: Contribution to journalArticleResearchpeer-review

11 Downloads (Pure)

Abstract

Tight-gas field development is capital intensive and the amount of data being generated in the process is immense as most platforms are now being digitized. Machine learning tools can be used to analyse these data in order to build patterns between several dependent and independent variables. In this study, two machine learning predictive models (ANN and GLM) were used to determine the expected recovery rate of planned new wells. The study approach is based on the analysis of reservoir rock/fluid properties and selected Well parameters to build decision-based models that can predict initial gas production rate from tight gas formations. Production data were retrieved from 224 wells and used in developing the model. The results obtained from these models were then compared to the actual recorded initial gas production rate from the Wells. Results from the analysis carried out revealed a Mean Square Error (MSE) of 1.57 on GLM model whereas the ANN model gave an MSE of 1.24. Key Performance Index for the ANN model revealed that the reservoir thickness had the highest (36.5%) contribution to the initial gas production rate followed by the flowback rate (29%). The reservoir/fluid properties contribution to the initial gas production rate was 53% while the hydraulic fracture parameters contribution to the initial gas production rate was 47%.
Original languageEnglish
JournalThe Mining Geology Petroleum Engineering Bulletin
Publication statusAccepted/In press - 18 Mar 2019

Fingerprint

Learning systems
Gases
Mean square error
Fluids
Tight gas
Rocks
Hydraulics
Recovery

Cite this

@article{cb3707d36ef346c4873f2cc8ceb2e28c,
title = "Application of machine learning models in predicting initial gas production rate from tight gas reservoirs",
abstract = "Tight-gas field development is capital intensive and the amount of data being generated in the process is immense as most platforms are now being digitized. Machine learning tools can be used to analyse these data in order to build patterns between several dependent and independent variables. In this study, two machine learning predictive models (ANN and GLM) were used to determine the expected recovery rate of planned new wells. The study approach is based on the analysis of reservoir rock/fluid properties and selected Well parameters to build decision-based models that can predict initial gas production rate from tight gas formations. Production data were retrieved from 224 wells and used in developing the model. The results obtained from these models were then compared to the actual recorded initial gas production rate from the Wells. Results from the analysis carried out revealed a Mean Square Error (MSE) of 1.57 on GLM model whereas the ANN model gave an MSE of 1.24. Key Performance Index for the ANN model revealed that the reservoir thickness had the highest (36.5{\%}) contribution to the initial gas production rate followed by the flowback rate (29{\%}). The reservoir/fluid properties contribution to the initial gas production rate was 53{\%} while the hydraulic fracture parameters contribution to the initial gas production rate was 47{\%}.",
author = "Amaechi, {Ugwumba Chrisangelo} and Ikpeka, {Princwill Maduabuchi} and Johnson Ugwu and Ma Xianlin",
year = "2019",
month = "3",
day = "18",
language = "English",
journal = "The Mining Geology Petroleum Engineering Bulletin",
issn = "1849-0409",
publisher = "Sveučilište u Zagrebu, Rudarsko-geološko-naftni fakultet",

}

Application of machine learning models in predicting initial gas production rate from tight gas reservoirs. / Amaechi, Ugwumba Chrisangelo; Ikpeka, Princwill Maduabuchi; Ugwu, Johnson; Xianlin, Ma.

In: The Mining Geology Petroleum Engineering Bulletin, 18.03.2019.

Research output: Contribution to journalArticleResearchpeer-review

TY - JOUR

T1 - Application of machine learning models in predicting initial gas production rate from tight gas reservoirs

AU - Amaechi, Ugwumba Chrisangelo

AU - Ikpeka, Princwill Maduabuchi

AU - Ugwu, Johnson

AU - Xianlin, Ma

PY - 2019/3/18

Y1 - 2019/3/18

N2 - Tight-gas field development is capital intensive and the amount of data being generated in the process is immense as most platforms are now being digitized. Machine learning tools can be used to analyse these data in order to build patterns between several dependent and independent variables. In this study, two machine learning predictive models (ANN and GLM) were used to determine the expected recovery rate of planned new wells. The study approach is based on the analysis of reservoir rock/fluid properties and selected Well parameters to build decision-based models that can predict initial gas production rate from tight gas formations. Production data were retrieved from 224 wells and used in developing the model. The results obtained from these models were then compared to the actual recorded initial gas production rate from the Wells. Results from the analysis carried out revealed a Mean Square Error (MSE) of 1.57 on GLM model whereas the ANN model gave an MSE of 1.24. Key Performance Index for the ANN model revealed that the reservoir thickness had the highest (36.5%) contribution to the initial gas production rate followed by the flowback rate (29%). The reservoir/fluid properties contribution to the initial gas production rate was 53% while the hydraulic fracture parameters contribution to the initial gas production rate was 47%.

AB - Tight-gas field development is capital intensive and the amount of data being generated in the process is immense as most platforms are now being digitized. Machine learning tools can be used to analyse these data in order to build patterns between several dependent and independent variables. In this study, two machine learning predictive models (ANN and GLM) were used to determine the expected recovery rate of planned new wells. The study approach is based on the analysis of reservoir rock/fluid properties and selected Well parameters to build decision-based models that can predict initial gas production rate from tight gas formations. Production data were retrieved from 224 wells and used in developing the model. The results obtained from these models were then compared to the actual recorded initial gas production rate from the Wells. Results from the analysis carried out revealed a Mean Square Error (MSE) of 1.57 on GLM model whereas the ANN model gave an MSE of 1.24. Key Performance Index for the ANN model revealed that the reservoir thickness had the highest (36.5%) contribution to the initial gas production rate followed by the flowback rate (29%). The reservoir/fluid properties contribution to the initial gas production rate was 53% while the hydraulic fracture parameters contribution to the initial gas production rate was 47%.

M3 - Article

JO - The Mining Geology Petroleum Engineering Bulletin

JF - The Mining Geology Petroleum Engineering Bulletin

SN - 1849-0409

ER -