Integrating Mathematical Optimization and Decision Making in Intelligent Fields

Abstract

In this paper a decision-making approach that can be applied to problems that are relevant to the oil and gas industry is presented. This methodology is supported by state-of-the-art mathematical optimization algorithms, and is based on the formal integration of the decisions in question with well-studied optimization procedures. The integration of the methodology with the application adds to its robustness. Two different types of problems are formulated and solved. The first kind is based on deciding which wells have to be shut in during a given production interval whilst simultaneously optimizing the controls for each selected well. The second category involves deciding for a group of wells which ones have to be injectors or producers, and at the same time searching for optimal well locations. In all the results obtained we can systematically see that the set of decisions proposed by the integrated approach mean substantial improvement in field production. For example, in the first class of problems studied, the production oil target is satisfied, and up to 50 percent of produced water is saved with respect to the reference case. The huge amount of information available, for example, in Intelligent/Smart Fields or Closed-Loop Reservoir Management can be utilized for rigorously making solid decisions. In this work we put an emphasis on integration of real-life decisions with a realistic simulation-based mathematical optimization framework. This framework can be also useful for establishing a common language for decision makers and researchers within a given organization, and as a consequence endowing the decision-making process with agility and robustness. It should be stressed that ultimately it is human interpretation and intuition that drives the making of crucial decisions. Automated tools should be understood as an additional (and hopefully valuable) source of information for making these important decisions.

Opening New Opportunities With Fast Reservoir-Performance Evaluation Under Uncertainty: Brugge Field Case Study

“…The optimization problems in the two cases presented can be approached by use of derivative-free methods (Conn et al 2009;Echeverría Ciaurri et al 2011;Kramer et al 2011)-which can be computationally expensive, but they often perform efficiently when the number of optimization variables is very small-or gradient-based techniques combined with numerical gradients, because (exact) derivative information may be difficult to obtain in reservoir-flow simulators when field-plateau rates are considered as control variables. We can expect that the computational cost associated with the solving (within acceptable practical accuracy) of an optimization problem that presents only one variable will be in most situations relatively small (e.g., no more than 10 objective-function evaluations).…”

“…General constraints could be dealt with by means of the introduction of a constraint-violation measure (in addition to the production-potential measure) combined with techniques inspired by the filter method for constrained optimization (Fletcher and Leyffer 2002;Echeverría Ciaurri et al 2011). Field-rate constraints (e.g., maximum field-fluid-production rate) may be transformed into well constraints through heuristics based on productivity/injectivity measures.…”

Role of Artificial Lift in Wells With a Downhole, Flow Control Completion

“…The problem of finding the optimum, real-time, ICV control setting has now become a complex, multidimensional, non-linear problem. (Echeverria-Ciaurri et al, 2012) have provided a good overview of the available control optimization methods. They also suggested that the selection of a suitable optimization method should include upscaling, parameterization, conditioning and identification of a suitable control algorithm.…”