Advanced Assisted History Matching of a Large Mature Oil Field Based on a Huge Number of Grid-Block Level Parameters and Saturation Functions

Abstract

Constructing reservoir models that are consistent with geophysical and geological static data is well understood. A persisting challenge is to condition such models to the available production dynamic data through the process of history-matching. A new algorithm is utilized to identify influencing grid-block properties based on analytical sensitivity calculations. The derived sensitivities allow efficient modification of grid-block properties and saturation functions to improve the history-match. This innovative approach preserves the geological model features, because changes are done at the grid-block level and are only as small as needed to achieve a good match. Usually, the number of parameters is so immense that engineers have to either restrict their attention to a small subset of the parameters (and likely missing crucial ones), or unnecessarily pay extremely high simulation costs. In this paper, we employ a new assisted history-matching technique that computes the parameter sensitivities analytically and for each grid-block. Here, the derivatives of the mismatch with respect to each parameter are rigorously computed based on the black-oil simulator's fluid flow equations. Hence, a single simulation run followed by a derivatives calculation session is sufficient to detect how each parameter affects the mismatch, and consequently, to decide how (or whether) to change each parameter to improve the match. This technique was successfully applied to history-match a mature oil field in the Middle East. Two sets of parameters are modified: The permeability in 3 directions per grid block, and relative permeability curves for about 50 saturation regions. The goal was to match water-cut for individual wells. With this analytical technique, excellent improvement in the match was achieved after only a dozen simulation runs and within a couple of days. Because the modifications are at the grid block level and minimal (only as and where needed), the technique preserves the original geological features of the model to a great extent. Eliminating the need of manual local modifications (e.g. box multipliers) is an important advantage of the method. The relative permeability curves have been tweaked successfully for numerous saturation regions using Corey model parameters. The ability to adjust many curves successfully using just a few simulation runs represents a significant advancement in the field of assisted history-matching.