Overview: Reservoir Simulation and Visualization (June 2007)

Abstract

This year I was in undated with quality papers relating to the topic of reservoir simulation. A quick look at the titles of the papers presented in this area over the past year showed a clear trend. "History matching" is far and away the number one topic of interest in the reservoir-simulation community. In fact, 25% of the papers available for review this year contained the term "history match" or "history matching" in the title. In light of this trend, all the papers presented for review and recommended reading this month relate to algorithms, modeling guidelines, and case studies involving history matching. Gone are the days when reservoir-simulation history-match studies were allowed to take several months or longer to complete. Fewer simulation engineers are trying to manage more reservoir-simulation projects. The natural outcome is a need for automation. There is the danger of automatic processes generating simulation results that are not feasible from an engineering perspective. As several of the authors point out, there is a caveat to automated history matching—engineering judgment is still required to guide the process. However, in addition to matching history, the automation procedure also must quantify the measure of the uncertainty of the reservoir. With this new-found capability to generate numerous models with similar quality of history match, the real test of a history-matched model is the predictability it provides. In addition, a secondary trend emerged from the stack of SPE papers on my desk. Not only is there a need to produce history-matched results faster, there is a need to produce these results for more-complex reservoirs. Increased energy is being spent on enhanced characterization of naturally fractured reservoirs and those with more-complex fluid systems. The need to automate the history-matching process of these complex reservoirs tends to be even greater because of the inherent computer times required for each simulation run. In searching for a third trend among the papers on my desk, the subject of visualization jumped out at me. Actually, the subject jumped out at me by its absence, because there were no papers that specifically addressed the subject. Yet, in a way, all the papers on automated history matching addressed it. In each of the reviewed papers, there are charts, figures, and diagrams that attempt to graphically uncover the key results from a set of hundreds or thousands of reservoir-simulation runs. As we invent new processes to automatically create more data, we need to continue to develop innovative ways to "see" the results. Reservoir Simulation and Visualization additional reading available at the SPE eLibrary: www.spe.org SPE 99578 "Efficient History Matching in Naturally Fractured Reservoirs" by T. Gang, SPE, University of Tulsa, et al. SPE 100489 "Soft Computing Algorithms Accelerate and Improve the History-Matching Process: Elk Hills, California—29R Reservoir" by T. Firincioglu, Nitec LLC, et al. SPE 102676 "Compositional and Grid-Orientation Effects in History Matching" by D. Eydinov, University of Bergen, et al. SPE 106044 "Event-Targeting Model Calibration Used for History Matching Large Simulation Cases" by Stig Selberg, SPE, Scandpower Petroleum Technology, et al.