Partition optimization approach for polymer flooding optimal injection problem

Considering the segmental structure of polymer flooding, a multi-stage optimization approach for polymer flooding optimal injection problem is proposed. By using the control vector parameterization (CVP) method, the control vector of the optimal problem is reformulated as a piecewise-constant form in which each constant piece represents an injection slug. Meanwhile the polymer flooding oil reservoir model is partitioned in time domain, and each model part corresponds to an injection slug. The maximum of the net present value (NPV) is used as optimum index. The polymer injection concentrations and partition lengths are selected as decision variables. The injection concentration box constraints are also considered. The proposed multi-stage optimization approach can optimize the switch time between two adjacent injection slugs. Since only injection concentrations and partition lengths act as decision variables, the computation can be decreased significantly. The simulation test of the polymer flooding one-dimensional laboratory experiment model demonstrates the effectiveness of this approach.