Technical Program


< Go back to the papers' list


Title

Influence of Compressibility on Incidence Losses of Turbomachinery at Subsonic Operation


Topic

H3 - Fan Performance (ii)


Authors

SAUL Sebastian
Technical University Darmstadt

Darmstadt - Germany
sebastian.saul@fst.tu-darmstadt.de
STONJEK Stefan S.
Technical University Darmstadt

Darmstadt - Germany
stefan.stonjek@fst.tu-darmstadt.de
PELZ Peter F.
Technical University Darmstadt

Darmstadt - Germany
Peter.Pelz@fst.tu-darmstadt.de

Abstract

Acceptance tests on large fans to prove the performance (efficiency and total pressure rise) are sophisticated and expensive. One way to reduce costs and efforts are a better prediction of fan performance at the design point as well as at off-design conditions. The commonly used scaling-up methods are working quite well near the design point and right now none of them are taking the compressible effects into account, which are getting more and more important at higher flow velocities. The need of highly loaded radial fans make it necessary to run fans at higher rotational speed and therefore at higher flow velocities. At part- or overload conditions inertia losses are important. Other authors showed that compressible effects are becoming highly relevant for radial fans and turbocharger compressors due to a shift of the peak efficiency to other flow coefficients, which occur at high flow velocities.
This paper begins with the physically based description of a two-dimensional flow through a blade cascade (without curvature) with variable incidence angle and incoming flow velocity. The incompressible case is now complemented with the common formulas for the compressible flow. If the friction along the blades is neglected and the blades are infinitesimal thin the analytical solution of the equation of continuity and the balance of momentum and energy exist. As well is the loss coefficient of the compressible incidence loss proportional to the square of the incoming Mach number. A comparison with fundamental experimental investigations of blade cascades shows a very similar behavior of the variation of the incidence angle and the incoming Mach number. A further correction is taking the compressible Borda-Carnot loss into account to include the fact that the blades in the investigated cascade have a nonzero height. This correction shows good results in a Mach number range from zero to 0.7 and is also good up to incidence angles at which flow separations appear. These separations are the limits of the theoretical model. Additional corrections can be included for the friction along the blade or further separation models can be added.