Enhancing Optimization of Reservoir Simulation Processes in Next-Generation Reservoir Simulators

Temizel, Cenk (Aera Energy) | Saputelli, Luigi (Frontender Corp.) | Nabizadeh, Mehdi (International Petro Asmari Company) | Balaji, Karthik (University of Southern California) | Suhag, Anuj (University of Southern California) | Ranjith, Rahul (University of Southern California) | Wijaya, Zein (HESS)

OnePetro 

Abstract

In field development and management, optimization has turned out to be an integral component for decision-making. Optimization involves computationally intensive complex formulations but simplifies making decisions. For reaching the optimal solution to a defined objective function, optimization software can be combined with a numerical reservoir simulator. Hence, robust and faster results are imperative to optimization problems.

To maximize cumulative recovery and net present value (NPV), the reservoir simulator works on maximizing these predefined objective functions that can be multi-objective leading to Pareto sets with "trade-offs" between objectives. In optimization algorithms with predefined objective functions, there is a need for these objective functions to be flexible by using conditional statements through procedures, since generally they do not provide the flexibility required by the physical reservoir fluid flow phenomenon to "maneuver" throughout optimization iterations.

In this study, a commercial reservoir simulator is coupled with an optimization software. As the need was discuss earlier, conditional statements are implemented in the simulator as procedures. Operating the software/simulator combination under pseudo-dynamic objective functions is achieved through these procedures. Highest recovery for the time period mentioned in the conditional statement for the simulation is achieved by trying sets of combinations of parameters, which also makes the optimization process faster and more robust. Throught the use of these conditional statements, the procedures are able to implement piecewise objective functions as codes for a given time frame.

The objective function to be maximized by the optimization process in this study cumulative production. The optimized recoveries with pseudo-dynamic objective functions provide an enhanced recovery, as compared to that of an optimization case with predefined constant objective function in the optimization software throughout the iterations of the optimization and simulation process.