Comparison of 1-D vs 3-D Combustion Boundary Conditions for SI Engine Thermal Load Prediction

Abstract

The thermal field generated in an engine block and cylinder head as a result of combustion loading is of paramount significance for structural durability. Computational fluid dynamics and heat transfer modeling provide strong tools; perhaps the best and most precise available for predicting thermal fields within cylinder head and engine block. However, an enduring challenge has been the temperature prediction on metal wall as a response to the time dependent fluctuations in the fluids. Fluid (coolant) flow in an engine is steady for a given engine speed and load, but combustion dynamics are inherently transient. In this study, an effective set of convective boundary condition data (as combustion load) is generated using two different approaches in a stand-alone simulation and mapped onto a decoupled Conjugate Heat Transfer (CHT) model to predict the temperature distribution in the engine. In the first approach, a predictive combustion model, tuned to dyno test data, is solved in a 1-D simulation code. This provides the cycle-averaged convective boundary condition that can be used for a CHT model as a uniform heat source. In the second, more detailed approach, in-cylinder combustion simulations involving transient piston and valve motion with flame propagation modeling are carried out using a 3-D simulation code. The 3-D methodology gives a detailed distribution of convective boundary conditions on the walls touching the combustion gases. In order to predict the gradients in heat transfer coefficient with high accuracy, the resulting temperature distribution from the CHT simulation is fed back into the combustion model to regenerate the set of convective boundary conditions. This process is repeated until a converged set of convective boundary conditions are obtained. In this paper engine temperature predictions obtained using combustion loads from both 1-D and 3-D approaches will be compared with the thermocouple data from engine dyno test.