Computational Fluid Dynamics for Gas Lift Optimization in Highly Deviated Wells

In the implementation of gas lift, understanding flow behavior in highly-deviated well is critical in avoiding production loss due to liquid fallback and blockage, even in highly-productive reservoir. In this work, we utilize Computational Fluid Dynamics (CFD) to optimize gas lift design under various flow behavior in highly-deviated well. The analysis is directly implemented into Arjuna offshore field case. Arjuna offshore field has gas-lifted wells, producing from a high-permeability reservoir. However, several wells suffer from huge production loss due to the effect of well's deviation. In deviated well, there exists frequent liquid fallback causes blockage, therefore, reducing the production. Motivated by this issue, we use CFD framework to perform gas lift optimization. We firstly adopt the geometry of gas-lifted wells as the computational domains for our simulation. An image-based meshing technique is deployed to capture the well's trajectory and internal geometry. We secondly utilize compressible Navier-Stokes equation and Finite Volume Method to evaluate the flow behavior. We capture the location of liquid fallback and liquid accumulation at elbows to estimate production loss. We consider the variation of viscosity, density, gas lift valve placement, injected gas rate, and reservoir pressure. We finally perform gradient-based optimization utilizing production loss as the objective function to obtain optimum design. The final result is then used to optimize the current design. The simulation results show that productivity index, pipe diameter, and deviation heavily influence the amount of production loss. At big pipe diameter and high deviation, the gravitational force governs the fluid flow. Therefore, slugs are developed and accumulated at elbows. This accumulation blocks gas flow and reduces production. Changing the gas injection rate affects the lifted density. High injection rate triggers segregation between the liquid and gas, while low injection rate does not reduce the liquid density. Shifting the gas lift valve placement influence the mixing between the liquid and gas. It also determines the cost of gas injection. Hence, we need to optimize both parameters at once. Six of thirty wells in Arjuna field experience severe liquid fallback, therefore, the production significantly decreases. The simulation shows up to 40% coverage of the pipe internal diameter, which blocks the gas flow. We perform the optimization by shifting the gas lift valve placement and adjusting the gas injection rate. By implementing the study result into the field case, we manage to improve the production by 20%. We provide an effective way to connect high-resolution simulation to the field design and revise the current concept in designing gas lift well completion. The simulation allows engineers to provide more insight on flow assurance in highly deviated wells.