Reservoir optimization for the synthetic brugge field

Increasing availability of data and tools for history matching and optimisation brings the long-term goal of closed loop reservoir management closer. However, still many issues are not solved. An example is the interaction between the history match of the reservoir model and the optimisation. How important is a good history match for the optimisation? We have tested this for a synthetic field using 5 realisations generated from well logs, maps and geological facies classifications. The synthetic field has 20 producers and 10 injectors, which all have three inflow control valves that can be controlled. First, the fit of the realisations to 10 years of production data is determined. Next, the water flooding strategy for each realisation is optimised using a lifecycle optimisation tool based on adjoint gradients. Finally, the resulting water flooding strategy is applied to the synthetic field and the realised net present value (NPV) is calculated. The results for this small number of realisations showed a clear relation between the quality of the fit to the production data and the realised NPV. The next step is to history match the realisations using production data and time lapse seismic and repeat the optimisation