In previously published experimental work completed at the Gas Turbine and Transmissions Research Centre (G2TRC), oil fed to an aeroengine location bearing via underrace feed was seen to shed from the cage, forming a film on static surfaces near the bearing and subsequently shedding into the bearing chamber. A high-fidelity computational model of the two-phase flow in an aeroengine bearing chamber must adequately reproduce such behaviour but there are significant challenges in modelling both the oil breakup after shedding and the subsequent film formation. It is very computationally costly to resolve an oil film interface using the Volume of Fluid (VOF) approach at regions of thin film and it is unacceptably inaccurate to resolve thick film using an explicit thin film modelling technique such as the Eulerian Thin Film Model (ETFM). A proposed solution is to couple together VOF, ETFM and discrete phase modelling (DPM). Previously published G2TRC work shows how VOF and ETFM can be successfully coupled. This paper investigates the coupling of ETFM and DPM. The evaluation of film momentum transport and air-particle momentum transfer/Lagrangian particle tracking are studied using a low Reynolds number turbulence model.

Validation is required to ensure that these models work together as intended. To this end a preliminary CFD study was carried out on a published case investigated experimentally and computationally in which a jet is injected into a duct via a nozzle, breaking up into droplets before forming a wall film. The droplets are produced by primary atomization due to liquid instabilities at the injection point. Secondary breakup occurs due to surface instabilities prompted by the high-velocity cross-flow. Small droplets are transported downstream whereas larger droplets deflect minimally hitting the wall and forming a thin film. In the work presented here quantitative film thickness data from experiments and prior simulations are compared to current data.

The success of the simulation is found to depend on shear-transportation, turbulent dispersion of the particles, particle grouping, mass transportation as well as accurate prediction of interfacial shear-stresses. With suitable modelling parameters it was possible to predict film thickness to within 28.9% of those seen experimentally. The present ETFM-DPM modelling showed improvements over previously published models in prediction of shear-stresses and film transportation as the ω-equation could be integrated through the viscous sublayer. The developed approach is now mature enough to be applied to the bearing chamber geometry investigated experimentally at G2TRC and this is proposed for future work.

This content is only available via PDF.
You do not currently have access to this content.