Applying Design of Experiments to Develop a Fuel Independent Heat Transfer Model for Spark Ignition Engines

Abstract

Models for the convective heat transfer from the combustion gases to the walls inside a spark ignition engine are an important keystone in the simulation tools which are being developed to aid engine optimization. The existing models are, however, inaccurate for an alternative fuel like hydrogen. This could be caused by an inaccurate prediction of the effect of the gas properties. These have never been varied in a wide range before, because air and ‘classical’ fossil fuels have similar values, but they are significantly different in the case of hydrogen. We have investigated the effect of the gas properties on the heat flux in a spark ignition engine by injecting different inert gases under motored operation and by using different alternative fuels under fired operation. Design of Experiments has been applied to vary the engine factors in a systematic way over the entire parameter space and to determine the minimum required amount of combinations. This paper presents the validation of a heat transfer model based on the Reynolds analogy for the entire experimental dataset. The paper demonstrates that the heat flux can be accurately predicted during the compression stroke with a constant characteristic length and velocity, if appropriate polynomials for the gas properties are used. Furthermore, it shows that a two-zone combustion model needs to be used as a first step to capture the effect of the flame propagation. However, alternative characteristic lengths and velocities will need to be investigated to capture the intensified convective heat transfer caused by the propagating flame front and to accurately predict the heat flux during the expansion stroke.