Methodology to Perform Conjugate Heat Transfer Modeling for a Piston on a Sector Geometry for Direct-Injection Internal Combustion Engine Applications

The increase in computational power in recent times has led to multidimensional computational fluid dynamics (CFD) modeling tools being used extensively for optimizing the diesel engine piston design. However, it is still common practice in engine CFD modeling to use constant uniform boundary temperatures. This is either due to the difficulty in experimentally measuring the component temperatures or the lack of measurements when simulation is being used predictively. This assumption introduces uncertainty in heat flux predictions. Conjugate heat transfer (CHT) modeling is an approach used to predict the component temperatures by simultaneously modeling the heat transfer in the fluid and the solid phase. However, CHT simulations are computationally expensive as they require more than one engine cycle to be simulated to converge to a steady cycle-averaged component temperature. Furthermore, a piston design optimization study would involve large numbers of simulations and including CHT modeling would be impractical considering the computational expense. Accordingly, in the current publication, an approach to perform piston CHT simulations on sector geometries is proposed to reduce the computational time significantly with minimal impact on the prediction accuracy. The study was performed on a heavy-duty engine at a high load operating condition of 20 bar gross IMEP and engine speed of 1800 rev/min. Since the open cycle portion of the engine cycle cannot be simulated on a sector mesh, a scaling methodology was developed to account for the contribution of the open cycle to the wall heat transfer. The average and the maximum piston temperature from the sector CHT approach were predicted within 35 K and 25 K of the full geometry CHT simulation respectively. Additionally, the distribution of the temperature within the solid piston obtained from the sector CHT simulation was compared to the temperature distributions from a full CHT geometry and the results showed good agreement. The sector CHT approach resulted in a ~8x reduction in computational time on a 1/7 sector geometry.