Comparison of Kalman Filter and Steepest Descent Method for Assisted History Matching
Every new generation of petroleum industry always search for new reliable method to approximate reservoir parameters in high resolution for long periods and large number of grid blocks for reservoir management purposes. History matching is a system that reduces the difference between the model perfo...
Saved in:
Main Author: | |
---|---|
Format: | Final Year Project |
Language: | English |
Published: |
Universiti Teknologi PETRONAS
2014
|
Subjects: | |
Online Access: | http://utpedia.utp.edu.my/14329/1/FARANADIA%20AB%20AZIZ%20%2813800%29%20Final%20Report.pdf http://utpedia.utp.edu.my/14329/ |
Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
id |
my-utp-utpedia.14329 |
---|---|
record_format |
eprints |
spelling |
my-utp-utpedia.143292017-01-25T09:37:00Z http://utpedia.utp.edu.my/14329/ Comparison of Kalman Filter and Steepest Descent Method for Assisted History Matching Ab Aziz, Faranadia T Technology (General) Every new generation of petroleum industry always search for new reliable method to approximate reservoir parameters in high resolution for long periods and large number of grid blocks for reservoir management purposes. History matching is a system that reduces the difference between the model performance and its historical behavior. There are many discoveries of optimization methods that can be applied for history matching purposes, but not all methods are suitable and reliable enough. This report studies the process of history matching, demonstrates some of the steps required and then review and compares the application of one method over another method for assisted history matching purpose. Assisted history matching is a technique that integrates forward model, formulation of an objective function and optimization. Among all of the optimization methods of history matching, Kalman filter and steepest descent methods are chosen. Steepest descent is one of the gradient based methods while Kalman filter is a non-gradient based method. Both of these methods are investigated and the results are compared. Two different sets of reservoir parameters of synthetic model are used to obtain both historical and simulated model and production data. Forward model is constructed and an objective function also is obtained in order to observe the discrepancy between the calculated and historical data. Hence, this report compares Kalman filter and steepest descent methods in order to investigate the more reliable methods that can provide better result in terms of their accuracy, CPU time and reliability for assisted history matching Universiti Teknologi PETRONAS 2014-05 Final Year Project NonPeerReviewed application/pdf en http://utpedia.utp.edu.my/14329/1/FARANADIA%20AB%20AZIZ%20%2813800%29%20Final%20Report.pdf Ab Aziz, Faranadia (2014) Comparison of Kalman Filter and Steepest Descent Method for Assisted History Matching. Universiti Teknologi PETRONAS. (Unpublished) |
institution |
Universiti Teknologi Petronas |
building |
UTP Resource Centre |
collection |
Institutional Repository |
continent |
Asia |
country |
Malaysia |
content_provider |
Universiti Teknologi Petronas |
content_source |
UTP Electronic and Digitized Intellectual Asset |
url_provider |
http://utpedia.utp.edu.my/ |
language |
English |
topic |
T Technology (General) |
spellingShingle |
T Technology (General) Ab Aziz, Faranadia Comparison of Kalman Filter and Steepest Descent Method for Assisted History Matching |
description |
Every new generation of petroleum industry always search for new reliable method to approximate reservoir parameters in high resolution for long periods and large number of grid blocks for reservoir management purposes. History matching is a system that reduces the difference between the model performance and its historical behavior. There are many discoveries of optimization methods that can be applied for history matching purposes, but not all methods are suitable and reliable enough.
This report studies the process of history matching, demonstrates some of the steps required and then review and compares the application of one method over another method for assisted history matching purpose. Assisted history matching is a technique that integrates forward model, formulation of an objective function and optimization. Among all of the optimization methods of history matching, Kalman filter and steepest descent methods are chosen. Steepest descent is one of the gradient based methods while Kalman filter is a non-gradient based method. Both of these methods are investigated and the results are compared.
Two different sets of reservoir parameters of synthetic model are used to obtain both historical and simulated model and production data. Forward model is constructed and an objective function also is obtained in order to observe the discrepancy between the calculated and historical data. Hence, this report compares Kalman filter and steepest descent methods in order to investigate the more reliable methods that can provide better result in terms of their accuracy, CPU time and reliability for assisted history matching |
format |
Final Year Project |
author |
Ab Aziz, Faranadia |
author_facet |
Ab Aziz, Faranadia |
author_sort |
Ab Aziz, Faranadia |
title |
Comparison of Kalman Filter and Steepest Descent Method
for Assisted History Matching |
title_short |
Comparison of Kalman Filter and Steepest Descent Method
for Assisted History Matching |
title_full |
Comparison of Kalman Filter and Steepest Descent Method
for Assisted History Matching |
title_fullStr |
Comparison of Kalman Filter and Steepest Descent Method
for Assisted History Matching |
title_full_unstemmed |
Comparison of Kalman Filter and Steepest Descent Method
for Assisted History Matching |
title_sort |
comparison of kalman filter and steepest descent method
for assisted history matching |
publisher |
Universiti Teknologi PETRONAS |
publishDate |
2014 |
url |
http://utpedia.utp.edu.my/14329/1/FARANADIA%20AB%20AZIZ%20%2813800%29%20Final%20Report.pdf http://utpedia.utp.edu.my/14329/ |
_version_ |
1739831989960179712 |
score |
13.18916 |