|
|
|
Abstract:
|
Flooding optimization in reservoir engineering maximizes the yield of a reservoir-model. The single shooting method is most encountered in literature. However, this method lacks handling of state-constraints. The output is therefore limited to a theoretical solution, where the aim is to realize an optimization strategy in real-life. This creates the need for more advanced methods. This survey will investigate two methods with a better constraint handling. The multiple shooting method handles constraints to some extent, where the simultaneous method can satisfy all constraints. The ability of handling constraints is related to a higher level of discretization, leading to significantly larger NLP problems. Although constraints can be satisfied using the simultaneous method, this will result in a very large NLP. It will be studied up to what extend the simultaneous method is preferable to single shooting.
|