COBEM99

S4 Ciências e Engenharia do Petróleo

Title:

COMBINING AUTOMATIC HISTORY MATCHING AND GEOSTATISTICAL TECHNIQUES TO IMPROVE OIL PRODUCTION FORECAST

Summary :

THE DEVELOPMENT OF AUTOMATIC HISTORY MATCHING TECHNIQUES IS RECEIVING MUCH ATTENTION RECENTLY. THE METHODS ARE BECOMING INCREASINGLY FASTER MAKING PRACTICAL APPLICATIONS POSSIBLE. IN GENERAL, THESE METHODS PROVIDE A SINGLE SOLUTION THAT MATCHES THE OBSERVED DATA, BUT AS THE SOLUTION OF INVERSE PROBLEMS IS NOT UNIQUE, THERE IS NO GUARANTEE THAT PRODUCTION FORECASTS MADE WITH THE ADJUSTED MODEL WILL PRODUCE GOOD RESULTS. THUS, IT WOULD BE DESIRABLE TO KNOW THE CONFIDENCE INTERVAL OF THE ADJUSTED MODEL. THE METHODOLOGY PRESENTED IN THIS PAPER COMBINES A HISTORY MATCHING TECHNIQUE WITH GEOSTATISTICAL MODELING TO PROVIDE EQUIPROBABLE RESERVOIR IMAGES THAT TAKE INTO CONSIDERATION PRODUCTION DATA. THE METHOD USES THE RESERVOIR IMAGE GENERATED BY THE OPTIMIZATION ALGORITHM TO GENERATE SEVERAL GEOSTATISTICAL IMAGES OF THE RESERVOIR TO PROVIDE A MEASURE OF UNCERTAINTY. ALTHOUGH THE COMPUTATIONAL COST OF THE METHODOLOGY IS HIGH, THE INFORMATION ABOUT THE UNCERTAINTY OF THE PRODUCTION FORECAST THAT IT GENERATES COMPENSATES THE ADDITIONAL TIME NECESSARY TO OBTAIN THE RESULTS. THE PAPER SHOWS A COMPLETE EXAMPLE WHERE THE ADJUSTED PARAMETER IS THE PERMEABILITY FIELD OF THE RESERVOIR. THIS METHODOLOGY COULD BE APPLIED WITH ANY OPTIMIZATION METHOD, AND IT PROVIDES AN EXCELLENT WAY TO ACCESS PRODUCTION UNCERTAINTY. KEYWORDS: AUTOMATIC HISTORY MATCHING, OPTIMIZATION, RESERVOIR CHARACTERIZATION.

Author :

Portella, Ricardo Cunha M.
Prais, Fábio 0

Paper View :