A Numerical Investigation Into Secondary Flows at the Inlet to a Low-Pressure Compressor Transition Duct

Abstract

To reduce fuel-burn and CO2 emissions from aero gas turbines there is a drive towards very-high bypass ratio and smaller ultra-high-pressure ratio core engine technologies. However, this makes the design of the ducts connecting various compressor spools more challenging as the larger required radius change increases their aerodynamic loading. This is exacerbated for the duct which feeds the engine core as it must accept the relatively low-quality flow produced by the fan root. This is characterised by a hub-low pressure profile and large secondary flow structures which will inevitably increase loss and the likelihood of flow separation. This paper presents a numerical design study aimed at modifying an existing fully annular, low-speed test facility to produce larger, more representative rotor loss cores. The CFD domain comprises a single stage axial compressor and transition duct representative of the low-pressure spool in a high bypass ratio turbofan. The rotating and stationary frames were coupled using a sliding mesh and a uRANS approach with Reynolds Stress closure for the turbulence modelling. The methodology was first validated using existing experimental data before examining a number of parameters such as inlet boundary layer thickness, inlet swirl, tip gap, blade solidity, thickness, lean. It was found that a combination of an increased inlet hub boundary layer and a reduction in solidity sufficiently increased the hub loading and encouraged the development of significantly large loss cores. Preliminary CFD results also showed that the increased rotor loss cores promoted better mixing through the outlet guide which subsequently reduced hub boundary layer thickness and secondary flows. Despite incurring a larger total pressure loss this suggested that the transition duct moved further from stall and could therefore potentially be made shorter.

Experimental Investigation of Secondary Flows and Length Reduction for a Low-Pressure Compressor Transition Duct

“…accelerating). Relative to other duct studies, the geometrical parameters suggest a modest level of aerodynamic loading but as highlighted by Walker et al [1] and Tsakmakidou et al [12] the true aerodynamic loading is more a function of the hub-low pressure profile and secondary flows delivered by the fan root flow. The ESS is designed to be load bearing so, in addition to removing ~30° of swirl from the rotor efflux, the 48 ESS have a low aspect ratio (~0.7), a physically long chord and large thickness (t/c ~ 0.12).…”

“…However, the experimental configuration used by Walker et al [1] did not produce representative rotor secondary flows and loss cores. Hence Tsakmakidou et al [12] built on the work by numerically examining the parameters which affect the generation of secondary flows through the rotor. The main aim of Tsakmakidou et al [12] was to design suitable modifications to the test rig compressor used by Walker et al [1] such that the rotor secondary flows were larger and more representative.…”

“…Hence Tsakmakidou et al [12] built on the work by numerically examining the parameters which affect the generation of secondary flows through the rotor. The main aim of Tsakmakidou et al [12] was to design suitable modifications to the test rig compressor used by Walker et al [1] such that the rotor secondary flows were larger and more representative. They found that this could be achieved by a combination of reduced solidity, increased diffusion factor (reduced blade count) and increased hub inlet boundary layer thickness.…”

“…The datum geometry shown in Fig. 9 is the same configuration employed by Walker et al [1] and as a starting point for the numerical investigation of Tsakmakidou et al [12]. The ESS and duct are representative of the LP-IP duct and are taken from the Integrated Core Technologies (iCORE) Project, a collaborative research program lead by Rolls-Royce plc., and funded by the Aerospace Technology Institute.…”

“…Although the rotor produces the required hub-low profile it does not generate the secondary flows of the correct size (see Walker et al [1]). Hence, Tsakmakidou et al [12] numerically developed a new rotor to increase the secondary flows. This was achieved by increasing the hub inlet boundary layer thickness and increasing the diffusion through the rotor passage.…”