Abstract Following the delineation and development of a field, the recovery of oil by cyclic steam stimulation and other thermal methods is controlled by actions taken at the surface. A number of operational variables effect the optimal management of steam stimulation programs. Generally these variables relate to the generation and allocation of steam for injection at production wells. Because the number of variables can be large, the management of steam stimulation programs is seen to require that solutions be found to constrained multivariate optimization problems.
In this paper, a novel and integrated approach is presented to the optimal management of cyclic stimulation programs. The approach combines the use of technologies to simulate the flow of steam in networks and wellbores with an industrially tested variant of the Successive Quadratic Programming algorithm (SQP) for process optimization. Examples are presented that illustrate the ability of the integrated approach to optimize the allocation of steam in distribution systems for which (1) the objective is to minimize the loss of thermal energy in surface piping and wellbore tubing, and (2) the objective is to increase the value of oil production that follows stimulation. Extensions of the approach to more realistic allocation problems are found (i) to require the inclusion of steam generation costs in the formulation of optimization problems, and (ii) to require the incorporation of mechanistic reservoir models for the stimulation of oil production by steam injection.
Cyclic steam stimulation, or ‘huff-and-puff’, is currently the most widely used technology for increasing the recovery of heavy oils by a thermal mechanism. By definition the process differs from steam-flooding in that the stimulation of wells is limited so as not to affect the production at adjacent wells. Typically, cyclic stimulation involves the rapid, but temporary, injection of steam into a select group of wells within a field. This portion of the process is termed the stimulation phase: its duration is on the order of weeks. Following the stimulation phase, there is a ‘soak’ phase during which the injection stops and steam is allowed to transfer heat to the reservoir. This stage lasts for a period of several days. Lastly, following the soak phase, there is a production phase that is driven by the pressure gradients naturally present in the reservoir. During the production stage, the recovery of oil in stimulated wells increases over that in non-stimulated wells because of a number of mechanisms. These include:the reduction of oil viscosity by heating,
the steam-driven drainage of oil by gravity,
the natural drive effects of expanding steam, and in some instances
the displacement of oil by surface subsidence.
Because the enhancement of oil production at wells is temporary, the stages of the stimulation process are repeated until the process becomes uneconomical.
Following the delineation of a field, the economic success of steam stimulation processes are controlled almost entirely by engineering and production decisions that are implemented at the surface. During the development of a field, decisions are made on a number of variables that can affect the performance of subsequent stimulation cycles. Examples include:the location and placement of wells,
the layout and levels of insulation for pipes, tubing and annuli,
the capacities of steam generators, and
the layout of the production network including items of process equipment such as separators.
Following the development of a field, the success of stimulation programs is affected by operational decisions that are made on a daily basis. These include:
P. 155^