Optimal designs of an ESP to handle upto 10% GVF

The present work aims at improving the performance of an electrical submersible pump (ESP) in multi-phase flow (MPHF) scenario. A single-stage ESP was designed and the MPHF was simulated considering maximum 10% gas volume fraction (GVF) by a CFD solver. A non-homogeneous MPHF model along with K-epsilon turbulence model was used to solve air-water two-phase flows. The reference geometry of the pump was altered by varying pump impeller inlet and exit angles to maximise the efficiency and the head, which were the objectives of the present problem. The CFD solver evaluated several designs and weighted-average surrogates were constructed. Populations for a multi-objective genetic algorithm were generated from the surrogates and Pareto optimal fronts were obtained. It was found that the gas bubbles get separated on pressure surface of the impeller. The conflicting objectives produce different designs which can help the pump designers to find better designs for the different GVFs.