Effective Method for Field Development Optimization and Its Application in Gulf of Mexico Oil Field

Abstract

An effective method is proposed for the solution of field development evaluation and optimization problems with categorical decision variables and subsurface uncertainties. Multiple reservoir models are applied for the representation of the subsurface uncertainties. A selected percentile of hydrocarbon recovery or an economic indicator (e.g. recovery factor, total recovery, or Net Present Value (NPV)) is maximized. The software package was developed for optimizing selections of: depletion scheme (e.g., primary depletion, gas injection, or water injection),facility size,well patterns and spacing,number of production or injection wells,well locations and trajectories from a discrete set of potential options,late-life sidetracks for existing and new wells,completion intervals for stacked reservoir units,order of reservoir zone development,well drilling schedules, etc. We consider problems with a limited number of potential values for each categorical decision variable. These combinatorial optimization problems are difficult to solve because the number of potential combinations is very large (factorial) and many hours of computer time are required for objective function evaluation running multiple reservoir models. We therefore apply a hybrid optimization method combining stochastic and sequential procedures. The sequential procedure is the novel part outlined in this paper. A reference depletion plan based on engineering experience is applied as an initial guess. Stochastic Genetic Algorithm (GA) or Particle Swarm Optimization (PSO) is executed until the objective function fails to improve in several iterations. Then, it is followed by the sequential procedure. The maximized objective function is the field oil recovery or economic indicator. The stochastic and sequential steps are repeated if the objective function is increased in the sequential procedure. The following operations are executed in each iteration of the sequential procedure: First, the previously evaluated development case with largest objective function value is identified as a starting point. Second, all potential values of the decision variables are evaluated sequentially changing one variable from the starting point. This field development optimization procedure was successfully applied in a large offshore oil field in the Gulf of Mexico (GOM). The field contains three major reservoir zones with high permeability and porosity. The field is structurally complex and it comprises of many fault blocks. We evaluated and optimized the waterflood performance and development drilling program selecting a) the numbers of new producers and injectors drilled in three reservoir zones; b) the locations and zonal completions of 41 new wells and sidetracks; and c) the well drilling schedule. NPV was maximized in the optimization procedure. The application of the hybrid method with the sequential procedure in the GOM oil field has the following benefits: The optimized field development case resulted in a 7% larger NPV and 3% larger oil recovery than those determined solely by GA.The individual well sensitivities on location and zonal completion were examined as part of the sequential procedure.The incremental value of each new well was determined.

Optimizing Waterflooding Process for Recovery Using Nonlinear Model Predictive Control: Case Yariguí - Cantagallo

“…These approaches use optimization methods to maximize the oil recovery for reservoirs under waterflooding stage. It can be found applications implemented in the Gulf of México [1][2][3] solving nonlinear programming problems, or methodologies similar to the work proposed by Jansen et al [4] which include a regulatory layer and history matching by using Ensemble Kalman Filter. The two-level closed Loop Reservoir Control strategy is detailed in works developed by Van Essen et al [5], A. Razapour [6] and E. Gildin [7].…”