Extended Brugge benchmark case for history matching and water flooding optimization

The Brugge benchmark case designed for the SPE Applied Technology Workshop (ATW) held in Brugge in June 2008 has proven to be valuable for testing and comparing methods of history matching, production optimization and closed-loop optimization by its extensive use in literature. Key features that contributed to its usefulness for history matching included the relatively large size of the problem, the large amounts of data and the complexity of the fine scale truth model, which gave rise to realistic upscaling issues. For optimization, the usefulness of the Brugge field case derives from the complexity and uncertainty in the reservoir model, the number of control settings to be optimized, and the number of constraints on the control variables.As a closed-loop optimization exercise, the primary weakness of the Brugge benchmark was the low frequency of the feedback loop. In terms of optimization, the Brugge case will remain a good test case for different aspects of optimization by using one of the realizations from the initial ensemble. In terms of history matching, matching data in the first 10 years (without individual completions) has become relatively easy with improvements in history matching technology. History matching with observations from individual well completions (after the first 10 years), however, remains challenging.In the original setup, the well constraints and data after the first 10 years are different for different participants due to the start of optimization at the end of Year 10. We will release an updated dataset to include well constraints and production data from individual well completions for years 10 to 30, as well as changes in saturation and pressure, which can be interpreted as coming from time-lapse seismic data. With this additional data, the Brugge benchmark case will remain a challenging problem to test techniques for history matching. © 2012 Elsevier Ltd.