Performance of pre-swirl rotating-disc systems

Karabay H., Pilbrow R., Wilson M., Owen J.

JOURNAL OF ENGINEERING FOR GAS TURBINES AND POWER-TRANSACTIONS OF THE ASME, vol.122, no.3, pp.442-450, 2000 (SCI-Expanded) identifier identifier


This paper summarizes and extends recent theoretical, computational, and experimental research into the fluid mechanics, thermodynamics, and heat transfer characteristics of the so-called cover-plate pre-swirl system. Experiments were carried out in a purpose-built rotating-disc rig, and the Reynolds-averaged Navier-Stokes equations were solved using two-dimensional (axisymmetric) and three-dimensional computational codes, both of which incorporated low-Reynolds-number k-epsilon turbulence models. The free-vortex flow, which occurs inside the rotating cavity between the disc and cover-plate, is controlled principally by the pre-swirl ratio, beta (p): this is the ratio of the tangential velocity of the air leaving the nozzles to that of the rotating disc. Computed values of the tangential velocity are in good agreement with measurements, and computed distributions of pressure are in close agreement with those predicted by a one-dimensional theoretical model. It is shown theoretically and computationally that there is a critical pre-swirl ratio, beta (p.crit), for which the frictional moment on the relating discs is zero, and there is an optimal pre-swirl ratio, beta (p.opt), where the average Nusselt number is a minimum. Computations show that, for beta (p)<(p.opt), the temperature of the blade-cooling air decreases as beta (p) increases; for beta (p) >beta (p.opt), whether the temperature of the cooling air increases or decreases as beta (p) increases depends on the flow conditions and on the temperature difference between the disc and the air. Owing to the three-dimensional flow and heat transfer near the blade-cooling holes, and to unquantifiable uncertainties in the experimental measurements, there were significant differences between the computed and measured temperatures of the blade-cooling air. In the main, the three-dimensional computations produced smaller differences than the two-dimensional computations.