A new hybrid production optimization algorithm for the combined CO2-cyclic solvent injection (CO2-CSI) and water/gas flooding in the post-CHOPS reservoirs

Abstract

In this paper, a new hybrid production optimization algorithm, namely, the steepest ascent‒particle swarm optimization (SA‒PSO), is developed to optimize the combined CO2-cyclic solvent injection (CO2-CSI) and waterflooding (WF) or gasflooding (GF) in the post cold heavy oil production with sand (CHOPS) reservoirs. The field-scale CO2-CSI, CO2-CSI + WF, and CO2-CSI + GF performances are predicted by upscaling the laboratory tests and using the numerical simulations with the considerations of the wormholes propagation and foamy-oil flow. The net present value (NPV) is chosen as an objective function and maximized by adjusting the well controls, such as the bottom-hole pressure (BHP) and water/gas injection rate. The efficacies of the following three optimization methods are evaluated and compared: the steepest ascent (SA) method, particle swarm optimization (PSO), and SA–PSO. It is found that the SA‒PSO gives a higher global optimum NPV than the SA algorithm and requires fewer numerical simulation runs than the PSO. It combines the fast convergence of the SA algorithm and the global search of the PSO. It is also found that the CO2-CSI + GF has the highest NPV in the post-CHOPS reservoir among the three enhanced oil recovery (EOR) processes. For the CO2-CSI + GF, the gas injector should inject CO2 at the highest injection rate during the early cycles to displace the foamy oil but should be shut in during the late cycles to control the gas channeling. The optimum NPVs of the CO2-CSI, CO2-CSI + WF, and CO2-CSI + GF are achieved when the BHPs of the producers are maintained at the lowest pressure. It is uneconomical to apply the WF immediately after the CO2-CSI due to the early water breakthrough.