Impingement cooling in which a surface is cooled by an impinging jet
(air or liquid) is an important method for cooling electronics because
of its high efficiency. The high efficiency is the result of thin
boundary layers formed at the point of impact and which produces very
high heat transfer coefficients. The jet's boundary layer does not
follow the log-law of the wall and cannot be computed using conventional
CFD turbulence models with wall functions.
In this project a prediction by Coolit for a jet impinging on a flat
surface was compared with experimental results from the paper by
Tsubokura et al published in the Monthly Journal of Institute of
Industrial Sciences at the University of Tokyo, 1997, vol 49, No. 1.
The Coolit domain size was 20 jet widths by 10 jet heights. A symmetry
plane was utilized to reduce the computational domain size and the grid
was clustered in the vicinity of the symmetry plane and the wall.
We used Coolit's eddy viscosity model without wall functions. The
jet was modeled using a Coolit Fan with the uniform velocity. At the
jet inlet (i.e. at the Fan), the default value of the eddy viscosity was
used. The Reynolds number based on the inlet velocity and the jet width
Figure 1 shows a
schematic of the experiment. Figures
3 show the
x-velocity profile two and four jet-widths away from the jet axis,
4 shows the y-velocity distribution for several