Havana — a fault modeling tool
Artikel i vetenskaplig tidskrift, 2002

Improved knowledge on faults and hydrocarbon seal put pressure on geologists and reservoir engineers doing reservoir modeling. All geo-knowledge must be built into the reservoir models to assure that it is taken into account in the decision processes. The need for advanced modeling tools is increasing. This paper describes the development of a fault modeling tool, the methodology behind it and examples of fault modeling studies. The general focus is on the uncertainty related to faults. The tool can be used for sensitivity analysis of fault effects, including studies of the flow effects of all faults scales, adding faults to simulation grids, and studies of the geometric uncertainty of the faults. The work started out as a development of a tool for stochastic modeling of sub-seismic scale faults. The faults can be added to a flow simulation grid as both displacement and seal. The current tool has been designed to operate together with the Eclipse flow simulator and the IRAP RMS program package. IRAP RMS is the main tool for visualizing output and Eclipse is used to examine the effect of the faults on hydrocarbon recovery. The techniques for modeling of fault seal, outputting results in a format that Eclipse can directly utilize, and the possibility for displacing simulation grids has proved useful also to seismic scale faults. This has led to further development, more detailed fault models and improvements of the general fault modeling capabilities. Examples of fault modeling, including three field examples, Statfjord, Heidrun and Sleipner, are presented to illustrate ways of including fault modeling as part of the reservoir modeling workflow.


Hollund Knut

Petter Mostad

Göteborgs universitet

Chalmers, Institutionen för matematisk statistik

Björn Fredrik Nielsen

Lars Holden

Jon Gjerde

Maria Grazia Contursi

Andrew J McCann

Chris Townsend

Einar Sverdrup

Norwegian Petroleum Society Special Publications

Vol. 11 157-171





Sannolikhetsteori och statistik