In the face of declining oil and gas reserves, increasing the productivity of oil fields has become a technological challenge of the utmost importance. Accordingly, well positioning and trajectory are critical factors and industry is seeking to acquire advanced methods to optimize well placement during field development.
This development phase relies on the simulation of fluid flows in the reservoir to determine:
- well position,
- and type.
Optimization consists in maximizing an objective NPV (Net Present Value) function incorporating field production gains and drilling costs, based on calculations derived from simulation outputs for each well (water, oil and gas production). The mixed nature of the optimization variables (continuous, integer, binary), the non-linearities of the NPV function and the associated calculation cost add considerably to the complexity of this problem.
IFPEN thus developed a two-step methodology consisting in(1, 2):
- determining the position and type of vertical wells only, thereby reducing the number of variables. This makes it possible to tackle a mixed, non-linear problem (with integer and continuous variables) using a global direct search method, based on numerical simulations;
- optimizing the trajectories of the branches coming off vertical wells, using a simplified analytical model.
This sequential optimization approach enables resolution via adaptations of traditional MINLP (Mixed Integer Non Linear Programming) methods. It can also be considered for the purposes of determining dates for bringing a well into production or closing down a well, or the implementation of adapted strategies once a well’s profitability starts to decline.
(1) C. Lizon, 2016, Mixed nonlinear optimization for integer and real variables: application to well location problem in reservoir engineering, thèse de doctorat, École polytechnique.
(2) C. Lizon, C. D’Ambrosio, M. Le Ravalec, L. Liberti, and D. Sinoquet, A Mixed-Integer Nonlinear Optimization Approach for Well Placement and Geometry, Proc. 14th European Conference on the Mathematics of Oil Recovery (ECMOR), 2014.
>> DOI: 10.3997/2214-4609.20141889
Scientific contact: Delphine Sinoquet