The Effect of Hub Leakage Flow on Two High Speed Axial Flow Compressor Rotors

Abstract

The effect of hub leakage flow on the performance of two high speed transonic rotors is investigated through numerical simulations and experiments. The leakage flow emanates from a small gap between the stationary and rotating parts of the hub flow path upstream of the rotor. Results of both the experiments and CFD simulations show that the introduction of a small leakage flow (0–25% of the main passage flow) can reduce the total pressure rise produced by the rotor across the entire span and generate a significant deficit in the total pressure profile near the hub. Numerical simulations done with a sinusoidal distribution of the leakage flow across the rotor pitch show that this deficit is present even when there is zero net leakage. Particle tracer studies of CFD simulations show that this deficit is due to the flow blockage produced by the radial migration of the low momentum leakage fluid. The performance degradation trends predicted by the simulations are qualitatively confirmed in the experimental investigation.

SWIFT Code Assessment for Two Similar Transonic Compressors

“…This terminology implies that there is something wrong with the flow in that region and has prompted much discussion in the literature. Hah et al 15 suggested that the deficit is due to a corner stall on the rotor suction surface, while Shabbir, et al 16 showed evidence that it is due to air leaking into the flow path from the gap between the stationary and rotating parts of the hub at the rotor leading edge. Chima, et al 9 suggested that the low total pressure is an intrinsic feature of the rotor that tends to be smeared out by the central-difference schemes used by many CFD codes.…”

“…7 computed total pressure profiles agree fairly closely with the data except in the region of low total pressure below 40 percent span. Other researchers have suggested that this region is due to a corner stall 15 or to hub leakage 16 . Previous work with SWIFT showed that both the central-difference and H-CUSP upwind differencing schemes smeared out the total pressure distribution in this region 9 .…”

Comparison of the AUSM+ and H-CUSP Schemes for Turbomachinery Applications

“…29 Alternatively, Shabbir, et al proposed that flow leakage between the centerbody and rotor disk could generate enough blockage to produce the deficits. 30 In this paper we suggest that the C-D schemes used in most codes smear details of the and distributions, while the upwind schemes used previously by Hah and Weber, and now in this work provide increased accuracy that gives better agreement with the experimental data.…”

“…The numerical speed of sound is given by: (30) where f is a scaling factor, is an average interface Mach number, and is the local relative Mach number limited between a cutoff Mach number and 1.…”

Multi-Fidelity Simulation of a Turbofan Engine With Results Zoomed Into Mini-Maps for a Zero-D Cycle Simulation

“…al. [22]. Wall functions are employed to model the turbulent shear stress adjacent to the wall without the need to resolve the entire boundary layer.…”