The effects of free-steam turbulence quantities on heat transfer to turbine blading
This thesis reports upon a second phase of research
into the effects of free-stream turbulence quantities on
heat transfer to impermeable gas turbine blading. It describes
the development of a novel form of turbulence generator,
to control separately the turbulence quantities intensity
and frequency upstream of a cascade of blades at levels
typical of the gas turbine. The turbulence generator was
calibrated for these individual quantities, with a hotfilament
anemometer system combined with on-line analog and
digital signal processing.
Blade heat transfer coefficients measured by two independent
techniques are compared. A large quantity of data is
presented, taken from a first stage high pressure rotor blade
and a nozzle guide vane. These were subjected to steady
flow and turbulent streams induced by both the novel turbulence
generator and by more conventional turbulence grids.
Surface pressure measurements have also been made, to predict
the heat transfer rates by applying formulae derived from
Much of the boundary layer over the two blades was
apparently laminar. For the laminar regions, the simple
formulae for heat transfer (flat plate for example) multiplied
by a turbulence term, will provide as good a correlation
as any. The intensity Tu is the most important turbulence
quantity, but there is some evidence that the
frequency of the perturbations can effect heat transfer.
Other evidence presented would s'uggest that profile geometry
is an overriding factor, which dictates the development of
the turbulence, whatever its origin, as well as controlling
its interaction with the boundary layer.
None of the correlations available for the prediction
of boundary layer transition are applicable. On both blade
suction surfaces separation seems to have occurred, and the
analysis indicates that transition on the pressure surfaces
of modern blades will be inhibited by the high free-stream
accelerations. Beyond transition, heat transfer is little
affected by turbulence. It is now clear, that measurements
of the turbulence as it develops through the cascade must be
performed before a successful prediction procedure for all
of the boundary layer regions can emerge.