An Automatic History Matching Example
Albert C. Reynolds, F. Zhang, Jan Arild Skjervheim · 2003
D-45 AN AUTOMATIC HISTORY MATCHING EXAMPLE Summary 1 F. ZHANG J. A. SKJERVHEIM and A. C. REYNOLDS Department of Petroleum Engineering University of Tulsa 600 South College Avenue Tulsa OK 74104-3189 USA Bayesian statistics provides a convenient formal framework for generating realizations of reservoir variables consistent with a prior geostatistical model and conditional to production data. (Here reservoir variables or model parameters refer to reservoir simulator gridblock log-permeabilities.) When the prior probability density function (pdf) for reservoir variables is multivariate Gaussian an approximate sampling of the a posteriori pdf can be conveniently obtained by applying the randomized maximum likelihood method.