Kikusato, A., Kusaka, J., and Daisho, Y., "A Numerical Study on Predicting Combustion Chamber Wall Surface Temperature Distributions in a Diesel Engine and their Effects on Combustion, Emission and Heat Loss Characteristics by Using a 3D-CFD Code Combined with a Detailed Heat Transfer Model," SAE Technical Paper 2015-01-1847, 2015, doi:10.4271/2015-01-1847.
A three-dimensional computational fluid dynamics (3D-CFD) code was combined with a detailed combustion chamber heat transfer model. The established model allowed not only prediction of instantaneous combustion chamber wall surface temperature distributions in practical calculation time but also investigation of the characteristics of combustion, emissions and heat losses affected by the wall temperature distributions. Although zero-dimensional combustion analysis can consider temporal changes in the heat transfer coefficient and in-cylinder gas temperature, it cannot take into account the effect of interactions between spatially distributed charge and wall temperatures. In contrast, 3D-CFD analysis can consider temporal and spatial changes in both parameters. However, in most zero-/multi- dimensional combustion analyses, wall temperatures are assumed to be temporally constant and spatially homogeneous. In reality, the wall temperature exhibits temporal and spatial distributions, thus influencing the characteristics of combustion, emissions and heat losses.In the present study, two numerical methods were developed to predict the temporal and spatial wall temperature distributions within a suitable calculation time for practical use. First, one-dimensional heat transfer calculation inside the wall along the normal direction to the wall surface was coupled to KIVA-4 with detailed chemistry. Second, ERENA, an explicit ordinary differential equation (ODE) solver developed by Morii et al., was applied to the detailed chemical kinetics.By utilizing these methods, the heat transfer calculation represented less than 1% of the total calculation time. The total calculation time was about 160 times faster than when calculated by the VODE solver with a single CPU thread on a desktop PC. The results indicate that the wall temperature and heat loss characteristics depend on the spray impingement on the walls rather than in-cylinder flow, like squish or swirl. Thus, the in-cylinder gas temperature distribution is affected by the change of wall temperatures, thereby altering the emission characteristics.