About This Item
- Full TextFull Text(subscription required)
- Pay-Per-View PurchasePay-Per-View
Purchase Options Explain
Share This Item
The AAPG/Datapages Combined Publications Database
Environmental Geosciences (DEG)
Abstract
DOI: 10.1306/eg.01241817008
Predicting gas
migration
through existing oil and gas wells
![Previous Hit](/data/images/arrow_left.gif)
![Next Hit](/data/images/arrow_right.gif)
James A. Montague,1 George F. Pinder,2 and Theresa L. Watson3
1Department of Civil and Environmental Engineering, University of Vermont, Burlington, Vermont; [email protected]
2Department of Civil and Environmental Engineering, University of Vermont, Burlington, Vermont; [email protected]
3T.L. Watson & Associates, Inc., Calgary, Alberta, Canada; [email protected]
ABSTRACT
The ability to accurately predict the probability of fluid migration
from depth through existing wells based on known well properties, such as age and depth, would be enormously helpful in understanding how
migration
pathways develop and the identification of potential
migration
without extensive field tests. The presence of fluid pathways is an important environmental issue because such pathways allow gas, either naturally occurring methane or sequestered CO2, to move into the atmosphere. In this paper, we explore the ability of various predictive models to forecast gas
migration
at existing wells in Alberta, Canada, based upon the characteristics of existing deep wells. Alberta was selected as a
case
study because of the availability of data in an area that has required wells to be tested for pathway development after rig release since 1995. Wells that do not demonstrate pathway development require no further testing until the well is abandoned. We show that accurately predicting fluid
migration
requires detailed information on well construction, production, and fluid properties, and even then, the models considered in this study misclassify a large number of wells. This suggests other factors may contribute to pathway formation. Of the models investigated, random forests provide the best results on this data set, correctly identifying 78% of the wells used.
Pay-Per-View Purchase Options
The article is available through a document delivery service. Explain these Purchase Options.
Watermarked PDF Document: $14 | |
Open PDF Document: $24 |