Analysis of an electrical submersible pump (ESP) slippage and head degradation when handling Newtonian and non-Newtonian fluids

Electrical submersible pumps (ESPs) are widely used for artificial lift in mature heavy-oil fields (water cut > 90%), where produced fluids may transition from Newtonian to shear-thinning behavior as oil–water emulsions evolve. This work investigates how fluid rheology affects the hydraulic performance and internal flow mechanisms of a four-stage ESP (specific speed, N s = 646 ) using a previously validated transient computational fluid dynamics (CFD) framework for a Franklin Electric 4400 (4-in.) pump, benchmarked against dedicated test-loop experiments. Two Newtonian fluids (water and sunflower oil) and two non-Newtonian carboxymethyl cellulose (CMC) solutions (0.5 and 1.0 wt%) were considered; the latter were described using a Cross-viscosity model to capture shear-thinning effects. The results show that performance deterioration with non-Newtonian fluids is markedly less severe than with a viscous Newtonian oil. For the CMC solutions, the high shear imposed by the impellers reduces the effective viscosity across the stages, yielding head and power levels that approach those of water in parts of the operating range. Particular emphasis is placed on impeller energy transfer and slippage. Slip-factor correlations (Stodola, Wiesner, and Li) were assessed, and the Li correlation best matched the experimental head, with a mean squared error (MSE) of 9.12%. Stage-wise CFD measurements of the impeller exit relative flow angle ( β 2 ) remained essentially constant for Newtonian fluids. In contrast, the shear-thinning cases showed an approximately 1° increase in β 2 per stage, consistent with the stage-wise reduction in effective viscosity. Finally, the loss analysis shows that friction losses increase with flow rate. In contrast, incidence/shock losses dominate near shut-off and become significant again under high-flow off-design conditions, with larger values observed for the more viscous cases.