4-3. PREDICTION OF HEAT TRANSFER
Subsequently, heat flux on the combustion chamber walls from high temperature gas was calculated using the turbulent heat transfer model, with the distribution of heat flux from high temperature gas on the cylinder cover and piston shown in Fig.8. It can accordingly be seen that the rate of heat transfer to the cylinder cover and piston wall surface becomes higher as the direction of fuel injection is increasingly towards the outside. Also, the absolute value of the rate of heat transfer is on the order of MW/m2, and, as this is the same order as for measured results obtained by other researchers,[12] the heat flux predicted here is thought to be at an appropriate level.
4-4. THE COMPARISON OF MEASURED TEMPERATURE WITH PREDICTED HEAT FLUX
In order to verify the calculated results, it would be necessary to subject the predicted heat flux obtained via simulation to heat conductance calculation by means of 3D FEM. However, it is extremely difficult to establish linkage with the combustion simulation, and indirect evaluation was therefore employed. Correlation between the time integration values of predicted heat flux and measured wall temperature would qualitatively prove the validity of the present prediction calculations.
Measured temperature points from the cylinder cover and piston in tests using an actual UEC85LS2 engine are shown in Fig.7, while a comparison of measured temperature at the points in Fig.7 with the time integration value of estimated heat flux appears in Fig.9. For each different temperature point on the line in the graph, the three data points represent the three fuel injection directions used in analysis and testing.
It can be seen from Fig.9 that measured temperature rises in accordance with the value of estimated heat flux time differentiation. The results are particularly close for the cylinder cover. When this trend is used as calibration data, combustion simulation performed prior to engine testing can predict cylinder cover wall temperature to a certain extent.
For the piston, however, there are points that do not follow the trend for measured data. One reason for this, as noted with respect to calculations vis a vis the observation test, is thought to be the effect of incorrect prediction of flame behavior resulting from spiral flow, and future improvements in accuracy will be required to estimate piston temperature.
