A physics-based design model for turbulent reaction rates in a natural gas fueled internal combustion engine

The design of conventional gasoline and diesel engines is one of engineering compromise. Most often this compromise is made after exhaustive laboratory experimentation. The application of Computational Fluid Mechanics and Heat Transfer techniques in the design of internal combustion engines is an advancing engineering discipline. The presently used tools need to be improved to lessen the need for expensive and time consuming experimentation. This work focuses on the development of a physics-based turbulent combustion model intended to improve the accuracy of such numerical analysis when applied to natural gas fired, spark ignition internal combustion engines.; Presently, the Arrhenius model is widely used for the prediction of the rate of combustion in laminar flows. The Arrhenius model uses temperature and the concentrations of fuel and oxidant as the variables that drive the rate of combustion. For turbulent flows there are a number of turbulent combustion models which predict the rate of combustion with varying degrees of success. The most commonly used one for engineering design purposes is the Magnussen model. The Magnussen model uses as variables the reactants concentration (as in the Arrhenius model) and the ratio of the dissipation of turbulent kinetic energy to the kinetic energy of turbulence.; This work applies the Reynolds averaging procedure to the Arrhenius model in conjunction with the mixing length theory to derive a new model founded on basic principles. The new turbulent combustion rate model is compared to the Magnussen one using the KIVA numerical analysis program and experimental data from a Cummins natural gas fueled engine. After calibration, both models are run and compared to the experimental data. It is found that in the majority of cases, the new model produced a closer match to the experimental data. In the numerical experiments, the peak pressures predicted by the Magnussen model show on average a 31% deviation from the experimental data while those predicted by the new model show an average 12% deviation. In addition the structure of the flame predicted by the new model is much more realistic as compared to that predicted by the Magnussen model.; It is concluded that use of the model developed in this work affords the design engineer involved in turbulent combustion analysis an improved tool to predict the rate of combustion.