Full text loading...
-
Geologically constrained history matching with PCA
- Source: First Break, Volume 30, Issue 11, Nov 2012,
-
- 01 Nov 2012
- Previous Article
- Table of Contents
- Next Article
Abstract
Principal component analysis has been used in a variety of disciplines such as pattern recognition, machine learning, and image processing. In this article, principal component analysis is used for history matching, i.e., adjusting the uncertain parameters of a reservoir model in order to match the simulated results with the observed behaviour. This approach incorporates geological realism into the history matching process by limiting the search to models that satisfy the prior geostatistical constraints. The Brugge model is presented as a test case against which the algorithm is validated. The results show an improvement over previously published data, yielding higher net present value and lower predictive error.
EAGE Publications BV