14th European Conference on Turbomachinery Fluid dynamics & Thermodynamics

Paper ID:


Main Topic:

Axial compressors



Christophe Montsarrat - PhD Student - LMFA Ecole Centrale de Lyon, ONERA and Safran Aircraft Engines
Jérôme Boudet - Ecole Centrale de Lyon
Julien Marty - ONERA
Eric Lippinois - Safran Aircraft Engines


The tip-clearance size of compressor rotors has a big effect on both performance and off-design conditions. Increasing the value of the tip gap can lead to a reduction in the efficiency and the surge margin of about 1 %. Mechanical constraints impose a minimal value to the tip clearance in order for the compressor to operate while a further reduction would generally be beneficial aerodynamically. However, the sensitivity to the gap size can show more complex tendencies. For example, some studies have shown a non-linearity in the performance decrease with the gap size, with an optimal value different from zero on some machines for some flow conditions. These flow conditions are not precisely understood and the influence of the tip gap size on the flow structures developing in the clearance region is not completely resolved. Parametric studies from RANS simulations and experiments have made real progress possible in the comprehension of the sensitivity of tip-leakage flow (TLF) to the clearance size, but the detailed resolution of the flow structures is only accessible through higher numerical resolutions. In this study, the flow around the rear rotor of a 3-stage high-pressure axial compressor has been simulated with a hybrid RANS/LES approach, for three gap sizes, at peak efficiency. Zonal detached-eddy simulation (ZDES) enables LES resolution out of the boundary layers and following the separation points, with a limited number of elements in the blade passage. In the present case, upstream of the third rotor, the casing boundary layer (modelled with RANS) is thick and immerses the three gaps investigated. However, from the detachment of the tip-leakage vortex (TLV), the ZDES model switches from RANS to LES thus treating the convection of the TLV in the passage with LES. Generally, the ZDES approach preserves the structures farther in the passage than RANS, with lower diffusion and less losses. Regarding the sensitivity to the tip-gap size, the results confirm the classical tendency from the literature with increased losses with increasing gaps. The TLV is more intense and generates more losses with the larger gap. The study also investigates the point of roll-up, the TLV trajectory and the distribution of axial vorticity in the vortex core and its evolution along the passage related to the ratio between the boundary-layer integral quantities and the tip-gap size.


Download it! Paper is available for download