A pattern flood management method based on a streamline simulator was developed to support waterflood designs. The methodology was applied on a structurally complex oil field in the North German basin. Studies are being conducted to understand the potential for extending the current waterflood in this oil field. The objective of this study was to investigate if a conventional simulation-based waterflood design could be enhanced using streamline simulation.

An alternative to using streamline simulation could be the post-processing of streamlines based on outputs of a full-field finite difference (FD) simulation model. However, there are limitations to this approach, including robustness and time considerations, especially when multiple runs with field-scale reservoir models are required. The streamline simulator contains a pattern flood management algorithm designed for optimizing the performance of waterfloods using multiple value criteria. The algorithm continuously balances patterns during forecasting runs converging to optimal injection and production rates while honoring well and field production constraints. A unique set of pattern performance diagnostics are ancillary products, for example pattern efficiencies and leakage fractions.

The full-field FD dynamic model of the aforementioned oil field was adapted for the streamline simulator. Both simulation models delivered similar results at the field and well levels and matched historical observed data satisfactorily. The best pattern flood model converged on a rate schedule that led to a 4% increase in oil production, a 17% decrease in water production, and a 5% reduction in the water injection volumes over the best performance achieved using a conventional voidage replacement strategy in the FD model. These findings were validated by executing the full-field model on a FD simulator with the recommendations from the pattern flood simulation run. The streamline simulation runs executed about seven times faster. To investigate the well count optimization potential, rigorous analyses were performed on the pattern information produced by the enhanced runs. A 12.5% reduction in well count, in terms of injectors and producers, could be achieved, and the pattern flood management algorithm converged on a rate schedule that still led to an increase of 2.3% in oil production, a 22% decrease in water production, and a 10% reduction in injection volumes.

The streamline-based simulation study proved useful in improving the existing waterflood design. Speedup in runtime allowed ample investigations and analysis within a given time period. Detailed analysis of allocated rate schedules and pattern information across numerous forecast runs gave deeper insight on the problem. The study highlighted that any well pattern has associated with it an optimal rate scheduling strategy. Hence, the two components are important aspects of any successful waterflood design. The recommended rate schedules are model based and hence subject to uncertainty, requiring updates as additional information becomes available over time.