Skip to main content
European Commission logo print header
Zawartość zarchiwizowana w dniu 2022-12-27

AHM DEMONSTRATION PROJECT

Cel

The aim of the project is to cut the cost in manpower and computing resources used in history matching a numerical model of a reservoir by 40% or more. A secondary aim is to provide a system which will assist the less experienced engineers in running history matches.

There are two main components to the system : The Adaptive History Matching (AHM) component which elicits the required information from the user, selects the next run of the simulator, and measures the quality of the run. The expert systems component which will optimise the solution space over which the AHM component searches for the best solution.
The AHM component uses a totally new mathematical approach, which we call Bayesian Sequential Design, and combines Bayes theory with the relatively new experimental design technique known as Sequential Design. This approach enables AHM to use all of the engineers knowledge and beliefts about a reservoir even where he is not certain something is true, and the results of previous runs of the simulator in choosing the next run. The expert system is built using an expert system shell, developed by ILOG of Paris. From a pilot survey of engineers in Europe and North America, we have built a preliminary set of inference rules for the expert system.
It is divided in 4 phases :
1) Pilot Study
2) Enhance/Validate
3) Final Testing
4) Commercialisation.

Zaproszenie do składania wniosków

Data not available

Koordynator

AEA Technology
Wkład UE
Brak danych
Adres
Winfrith Technology Centre
DT2 8DH Dorchester
Zjednoczone Królestwo

Zobacz na mapie

Koszt całkowity
Brak danych