The joint approach for well placement and control optimization uses an outer-loop algorithm for well placement optimization coupled with an efficient inner-loop optimization of well controls. The joint approach is implemented in an optimization framework where well placement is solved using a derivative-free pattern-search method in a highly distributed environment. The embedded control optimization uses a gradient-based solver with derivatives computed by an adjoint procedure, and is performed at each well placement iteration.

For optimization purposes, the MLO reservoir model is run using Stanford’s General Purpose Research Simulator (ADGPRS). A collaborative environment between Industry Partner Total E&P and the IO Center Research Team was established to closely treat and discuss project stages such as problem and constraint definition, model validation, effective implementation of optimization framework, and adaptation of solutions to industry standard. Results will be submitted for publication in collaboration with Industry Partner.