Water flooding is a frequently used technique to increase oil recovery after primary depletion. The presence of high permeability zones can have a large influence on the recovery, because they can cause early water breakthrough and trapping of by-passed oil. Smart well technology gives us the opportunity to counteract these effects by imposing an appropriate pressure or flow rate profile along the injection and production wells. In the current study we focus on water flooding with fully penetrating, smart, horizontal wells in 2 dimensional, horizontal reservoirs with simple, large-scale heterogeneities. The water flood is improved by changing the well profiles according to some simple algorithms that move flow paths away from the high permeability zone in order to delay water breakthrough. For all cases where early water breakthrough plays a role, it was possible to improve the water flooding process with these simple algorithms. For all these cases acceleration of production was possible. The increase in ultimate recovery obtained is very much dependent on total production time allowed. The shorter this time, the better the improvement in recovery will generally be. The increase in recovery obtained by applying the optimization algorithm varied between 0% - 20%. The delay in breakthrough time achieved by our optimization routine varied from 7–168 %. Our algorithms result in flow profiles that do not change in time. Both results from our own study and from literature show that time-varying flow profiles can at least further accelerate production. The general principle behind optimization in our cases was to reduce the difference in time of flight from injector to producer as much as possible.
Optimization of vertical well water flooding processes with fixed well locations, has been studied by Asheim1, and Sudaryanto and Yortsos2,3. Asheim studied optimization of water flooding for multiple vertical injection wells, and a single vertical production well (artificial water drive). He also studied a scenario with two vertical production wells and a natural aquifer (natural water drive). The optimization objective he used was maximum Net Present Value (NPV). As base case he used rate allocation based on the permeability-thickness product. Sudaryanto and Yortsos2,3 studied optimization of water flooding at water-breakthrough for two vertical injection wells, and a single vertical production well for incompressible fluids (total injection equals total production at all times). He used "bang bang" optimal control theory in which wells operate only on their extremes (fully open or closed). Optimization of the displacement process was done by optimization of the switch times for opening or closing of the injection wells. As a base case he used a scenario in which the injection rates remain constant throughout the displacement process and are chosen such as to get simultaneous breakthrough at the producer. One major difference between the approaches followed by Asheim and Sudaryanto is that in Asheims optimization the injection and/or production profiles change gradually over time, whereas in Sudaryanto's "bang bang" approach the injection profiles change abruptly. What both methods have in common is that they use combinations of two sources and one sink, or two sinks and one source. This differs from the present study.