Optimizing streamline Fluxes
Streamline methods are popular for simulating of many industrial reservoir flows, particularly those encountered in oil and gas production. A crucial step in streamline simulation is the allocation of fluxes to pre-computed streamlines which connect injection wells and producer wells.
This is in general an over-determined system. Constraints, some hard and some soft, that are put on the flux allocation include:
- the total production and the total injection for each well and each well face are matched as closely as possible
- all assigned fluxes need to be positive
- fluxes assigned to nearby streamlines are similar
In practice, the flux assignments are typically done in an ad hoc manner.
In this project more rigorous and effective ways of determining flux assignments are to be investigated.
The project will likely consist of the following phases:
- Construct a (limited) test suite with small problems (a couple of projectors and injectors per case in
- Formulate the problem as a constrained linear algebra problem.
- Design a few possible solution strategies and evaluate these using the test suite
- Apply the most promising approach(es) to a couple of realistic problems. Additional questions will arise
such as how to treat multiple streamlines connections.
More information: see a preliminary report here.
This project is linked to ongoing research at Stanford University.
Gunilla Kreiss will be responsible for the project.
The examples will be given by Prof. Margot Gerritsen, Stanford University.
Matlab will be used to solve the problem.