An iteration scheme with the KIVA-II code [2] and
the HCC code was also used in the calculations for
the Caterpillar diesel engine. In the first iteration, the
KIVA-II code was executed using assumed constant
temperature wall surface boundary conditions ; then
the HCC code was run to obtain a first wall temperature
distribution prediction, using the first heat
flux data computed by KIVA-II as boundary
conditions. In the second iteration, the KIVA-II code
was executed again to compute improved heat flux
boundary conditions, using the first temperature distributions
from the HCC code as boundary conditions
; then the HCC code was used to get the second
temperature distributions, using the second heat fux
data as boundary conditions. Finally, the KIVA-II
code was used to calculate the third heat flux data,
then the HCC code was used to obtain the final (converged)
temperature distributions. More details of this
iterative procedure and a discussion of the KIVA
model used are given by Liu and Reitz [15] and Han
and Reitz [3].
The computed time-averaged temperature distributions
on each surface of the combustion chamber
are shown in Fig. 11. For the cylinder head, the first
temperature distributions, as shown in Fig. 11 (a), are
50-120 K lower than the original (assumed) constant
temperature, and have their maximum value at the
edge of the bowl. The maximum time-averaged temperature
difference at points on the surface is about
70 K. The second temperature distributions are a little
higher than the first, but still lower than the original
guess. The third temperature distributions are almost
converged to the second, and the results show that the
iterations have reached satisfactory results,
For the piston, the first temperature distributions,
as shown in Fig. 11 (b), are also lower than the original
guess used in KIVA, and form a high temperature
region on the bowl-side surface and a low temperature
region on the bowl-valley surface. The second temperature
distributions are seen to give a higher peak
temperature than the first, especially on the bowl-side
surface, where they are even higher than the original
guess value. The average value of the second temperature
distribution seems to be near the original,
and this suggests that if a multidimensional combustion
code must use constant temperature boundary
conditions, the original (guessed) piston temperature
may have been a reasonable value. The third temperature
distributions also verify that the iterations have converged. Figure 11 (c) shows that the first,
second and third temperature distributions on the cylinder
wall surface are the same as the oil temperature
adjacent to the wall for the majority of locations on
the surface, but there is a 30 K increment in temperature
in the region near the cylinder head. This
shows that combustion and heat transfer within the
cylinder do not have much influence on the temperature
distribution of the wall surface, except for in
a narrow region in the ring reversal region near the
cylinder head.
The present HCC code provides an accurate and
consistent method for obtaining the temperature distributions
within engine components and these results
would also be useful for structural analysis. In
addition, combustion chamber wall temperatures
have been shown to significantly influence engine NO,
emissions [ 15]