Encouraged by the recent demand for eco-friendly combustion systems, advancements in the predictive capability of turbulent premixed combustion are considered to be essential. The explosion and deflagrating flame are modelled with the numerical method by applying the Large Eddy Simulation (LES) technique. It has evolved itself as a powerful tool for the prediction of turbulent premixed flames. In the LES, Sub-Grid Scale (SGS) modelling plays a pivotal role in accounting for various SGS effects. The chemical reaction rate in LES turbulent premixed flames is a SGS phenomenon and must be accounted for accurately. The Dynamical Flame Surface Density (DFSD) model which is based on the classical laminar flamelet theory is a prominent and well accepted choice in predicting turbulent premixed flames in RANS modelling. The work presented in this thesis is mainly focused upon the implementation of a dynamic flame surface density (DFSD) model for the calculation of transient, turbulent premixed propagating flames using the LES technique. The concept of the dynamism is achieved by the application of a test filter in combination with Germano identity, which provides unresolved SGS flame surface density information. The DFSD model is coupled with the fractal theory in order to evaluate the instantaneous fractal dimension of the propagating turbulent flame front. LES simulations are carried out to simulate stoichiometric propane/air flame propagating past solid obstacles in order to validate the model developed in this work with the experiments conducted by the combustion group at The University of Sydney. Various numerical tests were carried out to establish the confidence of LES. A detailed analysis has been carried out to determine the regimes of combustion at different stages of flame propagation inside the chamber. LES predictions using the DFSD model are evaluated and validated against experimental measurements for various flow configurations. The LES predictions were identified to be in strong agreement with experimental measurements. The impact of the number and position of the baffles with respect to ignition origin has also been studied. LES results were found to be in very good agreement with experimental measurements in all these cases.

The in-cylinder flow of internal combustion (IC) engines, formed during the intake stroke, is one of the most important factors that affect the quality of air-fuel mixture and combustion. The inducted airflow through the inlet valve is primarily influenced by the intake port design, intake valve design, valve lift and valve timing. Such parameters have a significant influence on the generation and development of in-cylinder flow motion. In most combustion systems the swirl and tumble motions are used to aid the air-fuel mixing with the subsequent decay of these bulk flow motions generating increased turbulence levels which then enhance the combustion processes in terms of rate of chemical reactions and combustion stability. Air motion formed inside the engine cylinder is three-dimensional, transient, highly turbulent and includes a wide spectrum of length and time scales. The significance of in-cylinder flow structures is mainly reflected in large eddy formation and its subsequent break down into turbulence kinetic energy. Analysis of the large scale and flow motions within an internal combustion engine are of significance for the improvement of engine performance. A first approximation of these flow structures can be obtained by steady state analysis of the in-cylinder flow with fixed valve lifts and pressure drops. Substantial advances in both experimental methods and numerical simulations provide useful research tools for better understanding of the effects of rotational air motion on engine performance. This study presents results from experimental and numerical simulations of in-cylinder flow structures under steady state conditions. Although steady state flow problem still includes complex three-dimensional geometries with high turbulence intensities and rotation separation, it is significantly less complex than the transient problem. Therefore, preliminary verifications are usually performed on steady state flow rig. For example, numerical investigation under steady state condition can be considered as a precondition for the feasibility of calculations of real engine cylinder flow. Particle Image Velocimetry (PIV) technique is used in the experimental investigations of the in-cylinder flow structures. The experiments have been conducted on an engine head of a pent-roof type (Lotus) for a number of fixed valve lifts and different inlet valve configurations at two pressure drops, 250mm and 635mm of H2O that correlate with engine speeds of 2500 and 4000 RPM respectively. From the 2-D in-cylinder flow measurements, a tumbling vortex analysis is carried out for six planes parallel to the cylinder axis. In addition, a swirl flow analysis is carried out for one horizontal plane perpendicular to the cylinder axis at half bore downstream from the cylinder head (44mm). Numerically, modelling of the in-cylinder flow is proving to be a key part of successful combustion simulation. The numerical simulations require an accurate representation of turbulence and initial conditions. This Thesis deals with numerical investigation of the in-cylinder flow structures under steady state conditions utilizing the finite-volume CFD package, STAR CCM+. Two turbulence models were examined to simulate the turbulent flow structure namely, Realizable k-ε and Reynolds Stress Turbulence Model, RSM. Three densities of generated mesh, which is polyhedral type, are examined. The three-dimensional numerical investigation has been conducted on the same engine head of a pent-roof type (Lotus) for a number of fixed valve lifts and both valves are opened configuration at two pressure drops 250mm and 635mm of H2O that is equivalent to engine speeds of 2500 and 4000 RPM respectively. The nature and modelling of the flow structure together with discussions on the influence of the pressure drop and valve lift parameters on the flow structures are presented and discussed. The experimental results show the advantage of using the planar technique (PIV) for investigating the complete flow structures developed inside the cylinder. It also highlighted areas where improvements need to be made to enhance the quality of the collected data in the vertical plane measurements. Based on the comparison between the two turbulence models, the RSM model results show larger velocity values of about 15% to 47% than those of the Realizable k-ε model for the whole regions. The computational results were validated through qualitative and quantitative comparisons with the PIV data obtained from the current investigation and published LDA data on both horizontal and vertical cross sections. The calculated correlation coefficient, which is above 0.6, indicated that a reasonable prediction accuracy for the RSM model. This verifies that the numerical simulation with the RSM model is a useful tool to analyse turbulent flows in complex engine geometries where anisotropic turbulence is created.

No description available.

No description available.

No description available.

No description available.

No description available.

No description available.

No description available.

No description available.