A comprehensive study of the flowfield through a two-dimensional cascade of the
high pressure turbine blades of a jet engine is presented. The main interest is the
measurement and prediction of the mass-averaged total pressure losses. Other experiments,
such as flow visualization, are aimed at the validation of the code that
was used to obtain the numerical results and also to further knowledge about the
details of the loss generation. The experimental studies were carried out on a cascade
of eleven blades in a blow-down tunnel. Total pressure measurements were
taken upstream of the cascade and also by traversing on downstream planes. The
static pressures needed for the mass averaging and the probe bow shock correction
were obtained by pressure taps on the cascade tunnel side wall. The static pressure
was also measured on the surface of some instrumented blades. Shadowgraph pictures
were taken for study of the trailing edge shock structure and for the turbulent
transition location. A single-plate interferometer technique was used for density
field measurements. The major goal of the numerical studies was the prediction
of the mass-averaged total pressure losses, but all other measured quantities were
also generated from the computed flowfield. A critical issue was the generation of a
proper grid. For the studied type of flow, a non-periodic C-type grid turned out to
be the most advantageous. For use in the moderately compressible attached turbulent
boundary layer, a Clauser-type eddy viscosity model was developed and tested.
In the trailing edge and wake region, the Baldwin-Lomax model was used. Good
agreement of calculations and measurements was obtained for the blade surface and
cascade tunnel side wall static pressures, the trailing edge shock structure, and the
density field. The agreement between the measured and calculated total pressure
drop profiles was not quite as good; however, that quantity is known to be difficult
to predict accurately. The mass-averaged total pressure loss coefficient, calculated
from the total pressure drop profiles, was again in good agreement with the measurements.
The difference between the measured and computed total pressure drop
profiles suggested that the Baldwin-Lomax model underpredicted the eddy viscosity
in the trailing edge region.