Thermal anti-icing systems are commonly used to protect aircraft leading edges from a potentially hazardous build-up of ice. Such systems have proven reliable in service and are relatively cheap and efficient. Typically, hot air is tapped from the
engine compressor and ducted (via a regulation and control system) to the surface to be protected. Ideally an optimisation process should be employed at
the design stage in order to ensure adequate anti-icing capability with minimal use of engine bleed air, since the latter represents a performance penalty.
Following submission by the author of an MSc thesis concerning thermal modelling of a hot air anti-icing system for a civil turbofan intake (Wade 67 ), it became clear that extension of the studies was necessary to
enable systematic accounting of the factors which limit ice accretion. An experimental programme was therefore carried out to investigate primarily: various exhaust geometries (through which spent
anti-icing air is emitted to join the main
engine inlet airflow and provide heating of the downstream surface; various cowl internal configurations
on a full-scale model section of a large civil
turbofan Nose Cowl. The internal geometry affects the
effectiveness of the cowl lipskin heating, and the
spent anti-icing exhaust air limits the quantity of
unevaporated water which runs back along the intake
acoustic surface downstream of the directly heated area
and freezes. The Computational Fluid Dynamics package PACE (Prediction of Aerodynamics and Combustor Emissions)
was used to model the internal, freestream and exhaust
airf lows to determine the program's potential and
usefulness for predictive purposes in this type of
application. PACE is capable of modelling two or three
dimensional, recirculating or non-recirculating flows
for simple rectangular or polar geometry. It
encompasses a suite of sub-programs to generate meshes
and to create and solve the set of coupled linear
equations representing the fluid flow. Various
parameters, including heat transfer coefficients, were
predicted in two regions: downstream of the exhaust plane to model the
mixing of the spent anti-icing air and the
freestream main engine inlet flow; inside and outside the Nose Cowl highlight
area to predict skin temperature distributions
for the three internal geometry configurations
tested.
This thesis describes the experimental work and
compares the results with the Computational Fluid
Dynamics predictions. Agreement was generally found to
be good, and it was concluded that PACE may provide a
useful modelling (design) tool, albeit with some reservations.
History
School
Mechanical, Electrical and Manufacturing Engineering