Low Reynolds numbers influence on cascade losses in low pressure turbine

Аuthors

Khairulin V. T.^*, Bagin S. V.^**, Shvyrev A. A.^***, Tikhonov A. S.^****, Gaskarov A. V., Samokhvalov N. Y.^*****

Company «Aviadvigatel», 93, Komsomolsky avenue, Perm, GSP, 614990, Russia

*e-mail: khairulin@avid.ru
**e-mail: bagin@avid.ru
***e-mail: shvyrev@avid.ru
****e-mail: tikhonov@avid.ru
*****e-mail: samohvalov@avid.ru

Abstract

To analyze the effect of low Reynolds Numbers on losses in a low pressure turbine, the experimental study of gas-dynamic characteristics of two-dimensional cascades for three turbines was performed at low Reynolds Numbers at CIAM’s U300S test bench.
The cascades have similar design and consist of 6 airfoils. Cascade 1 has profiles with loads applied to the mid-portion, Cascade 2 has profiles with loads applied forward and Cascade 3 has profiles with loads applied rearward. All cascades have identical relative pitch and chord and identical flow entry angles b_1к and effective flow exit angles b_2эф. Reynolds Number was changed as by adjusting total inlet pressure and varying corrected speed at the cascade exit l_2is. Turbulizing screens were used in the experimental study.
The study shows the following:

• with the lower turbulence level atthe cascade inlet (Tu = 2%) the level oflosses inCascades 1and 3are nearly the same, while losses inCascade 2are in1.5 time lower;
• with the mean turbulence level at the cascade inlet (Tu = 5%) Cascades 1and 2 have equal level of losses (The difference is less than 1%)
• with the high turbulence level at the cascade inlet (Tu = 12%) Cascade 2 has distinct advantage over Cascade 1 under operation conditions l_2is.

To verify the experimental data a numerical simulation of Cascade 1 and Cascade 2 was performed at the inlet turbulence of 5%. The following conclusions can be made therefrom:

• numerical simulation based on RANS transient models makes it possible to define approximate place of flow separation but does not allow evaluating the losses increase when the Reynolds Number is becoming lower.
• numerical simulation based on RANS full turbulence models (k-e, k-w, SST, etc.) for simulating separation-induced flows at low Reynolds Numbers makes no sense because such models do not reflect a physically well-understood flow phenomenon, and the high level of losses in these models is achieved through high eddy viscosity of nonphysical nature.

Keywords:

low Reynolds, testing, turbulence model, low pressure turbine blade

References

1. Binder A., Schroeder Th., Hourmouziadis J. Turbulence Measurements in a Multistage Low-Pressure Turbine, ASME Journal of Turbomachinery, 1989, vol. 111, pp. 153-161.
2. Popovic I., Zhu J., Dai W., Sjolander S.A., Praisner T., Grover E. Aerodinamics of family of three higly loaded low pressure turbine airfoils: measured effects of Reynolds number and turbulence intensity in steady flow, ASME Paper, GT2006-91271.

Download