Surrogate-Based Optimization for the Design of Rotary Gas Separator in ESP Systems

Abstract

Summary The rotary gas separator (RGS) is used to separate gas from the two-phase flow and reduce input gas into the electrical submersible pumps (ESPs). When the pump handles flow with high gas void fraction (GVF), it causes deterioration of the head and performance of the pump system. Thus, performance prediction of the RGS is very important. In this study, optimization of separator blades in the RGS is carried out using computational fluid dynamics (CFD), and surrogate methods are used to increase gas separation efficiency. Three-dimensional (3D) Navier-Stokes equations are solved by using the Euler-Euler approach for the inhomogeneous two-phase flow (water/air), and modeling interphase forces in ANSYS® CFX 15.0 software (ANSYS, Inc., Canonsburg, Pennsylvania, USA). The kriging (KRG) model is used to approximate the objective function as a function of design variables including the inlet angle, straight blade length, and tip-clearance gap. Sensitivity analysis showed that the tip-clearance gap has more effect on efficiency. The results for optimized values indicated that the separation efficiency increased, and efficiency is progressively transferred from the high-efficiency region to the low-efficiency region.