An Artificially Intelligent Technique to Generate Synthetic Geomechanical Well Logs for the Bakken Formation
Artificially intelligent and predictive modelling of geomechanical properties is performed by creating supervised machine learning data models utilizing artificial neural networks (ANN) and will predict geomechanical properties from basic and commonly used conventional well logs such as gamma ray, a...
Main Authors: | , , |
---|---|
Format: | Article |
Language: | English |
Published: |
MDPI AG
2018-03-01
|
Series: | Energies |
Subjects: | |
Online Access: | http://www.mdpi.com/1996-1073/11/3/680 |
_version_ | 1818035745649065984 |
---|---|
author | George Parapuram Mehdi Mokhtari Jalel Ben Hmida |
author_facet | George Parapuram Mehdi Mokhtari Jalel Ben Hmida |
author_sort | George Parapuram |
collection | DOAJ |
description | Artificially intelligent and predictive modelling of geomechanical properties is performed by creating supervised machine learning data models utilizing artificial neural networks (ANN) and will predict geomechanical properties from basic and commonly used conventional well logs such as gamma ray, and bulk density. The predictive models were created by following the approach on a large volume of data acquired from 112 wells containing the Bakken Formation in North Dakota. The studied wells cover a large surface area of the formation containing the five main producing counties in North Dakota: Burke, Mountrail, McKenzie, Dunn, and Williams. Thus, with a large surface area being analyzed in this research, there is confidence with a high degree of certainty that an extensive representation of the Bakken Formation is modelled, by training neural networks to work on varying properties from the different counties containing the Bakken Formation in North Dakota. Shear wave velocity of 112 wells is also analyzed by regression methods and neural networks, and a new correlation is proposed for the Bakken Formation. The final goal of the research is to achieve supervised artificial neural network models that predict geomechanical properties of future wells with an accuracy of at least 90% for the Upper and Middle Bakken Formation. Thus, obtaining these logs by generating it from statistical and artificially intelligent methods shows a potential for significant improvements in performance, efficiency, and profitability for oil and gas operators. |
first_indexed | 2024-12-10T06:59:56Z |
format | Article |
id | doaj.art-4522d6902e2540559a8d87836aee1870 |
institution | Directory Open Access Journal |
issn | 1996-1073 |
language | English |
last_indexed | 2024-12-10T06:59:56Z |
publishDate | 2018-03-01 |
publisher | MDPI AG |
record_format | Article |
series | Energies |
spelling | doaj.art-4522d6902e2540559a8d87836aee18702022-12-22T01:58:21ZengMDPI AGEnergies1996-10732018-03-0111368010.3390/en11030680en11030680An Artificially Intelligent Technique to Generate Synthetic Geomechanical Well Logs for the Bakken FormationGeorge Parapuram0Mehdi Mokhtari1Jalel Ben Hmida2Department of Petroleum Engineering, University of Louisiana at Lafayette, Lafayette, LA 70504, USADepartment of Petroleum Engineering, University of Louisiana at Lafayette, Lafayette, LA 70504, USADepartment of Mechanical Engineering, University of Louisiana at Lafayette, Lafayette, LA 70504, USAArtificially intelligent and predictive modelling of geomechanical properties is performed by creating supervised machine learning data models utilizing artificial neural networks (ANN) and will predict geomechanical properties from basic and commonly used conventional well logs such as gamma ray, and bulk density. The predictive models were created by following the approach on a large volume of data acquired from 112 wells containing the Bakken Formation in North Dakota. The studied wells cover a large surface area of the formation containing the five main producing counties in North Dakota: Burke, Mountrail, McKenzie, Dunn, and Williams. Thus, with a large surface area being analyzed in this research, there is confidence with a high degree of certainty that an extensive representation of the Bakken Formation is modelled, by training neural networks to work on varying properties from the different counties containing the Bakken Formation in North Dakota. Shear wave velocity of 112 wells is also analyzed by regression methods and neural networks, and a new correlation is proposed for the Bakken Formation. The final goal of the research is to achieve supervised artificial neural network models that predict geomechanical properties of future wells with an accuracy of at least 90% for the Upper and Middle Bakken Formation. Thus, obtaining these logs by generating it from statistical and artificially intelligent methods shows a potential for significant improvements in performance, efficiency, and profitability for oil and gas operators.http://www.mdpi.com/1996-1073/11/3/680Bakken Formationunconventionalgeomechanicsshear waveartificial intelligencepredictive modelingsupervised machine learningartificial neural networks (ANN)regressionhypothesis testingcost-savings |
spellingShingle | George Parapuram Mehdi Mokhtari Jalel Ben Hmida An Artificially Intelligent Technique to Generate Synthetic Geomechanical Well Logs for the Bakken Formation Energies Bakken Formation unconventional geomechanics shear wave artificial intelligence predictive modeling supervised machine learning artificial neural networks (ANN) regression hypothesis testing cost-savings |
title | An Artificially Intelligent Technique to Generate Synthetic Geomechanical Well Logs for the Bakken Formation |
title_full | An Artificially Intelligent Technique to Generate Synthetic Geomechanical Well Logs for the Bakken Formation |
title_fullStr | An Artificially Intelligent Technique to Generate Synthetic Geomechanical Well Logs for the Bakken Formation |
title_full_unstemmed | An Artificially Intelligent Technique to Generate Synthetic Geomechanical Well Logs for the Bakken Formation |
title_short | An Artificially Intelligent Technique to Generate Synthetic Geomechanical Well Logs for the Bakken Formation |
title_sort | artificially intelligent technique to generate synthetic geomechanical well logs for the bakken formation |
topic | Bakken Formation unconventional geomechanics shear wave artificial intelligence predictive modeling supervised machine learning artificial neural networks (ANN) regression hypothesis testing cost-savings |
url | http://www.mdpi.com/1996-1073/11/3/680 |
work_keys_str_mv | AT georgeparapuram anartificiallyintelligenttechniquetogeneratesyntheticgeomechanicalwelllogsforthebakkenformation AT mehdimokhtari anartificiallyintelligenttechniquetogeneratesyntheticgeomechanicalwelllogsforthebakkenformation AT jalelbenhmida anartificiallyintelligenttechniquetogeneratesyntheticgeomechanicalwelllogsforthebakkenformation AT georgeparapuram artificiallyintelligenttechniquetogeneratesyntheticgeomechanicalwelllogsforthebakkenformation AT mehdimokhtari artificiallyintelligenttechniquetogeneratesyntheticgeomechanicalwelllogsforthebakkenformation AT jalelbenhmida artificiallyintelligenttechniquetogeneratesyntheticgeomechanicalwelllogsforthebakkenformation |