Search Results

The in cylinder air motion, fuel air mixing, evaporation, combustion and exhaust emissions have been simulated for a four stroke direct injection gasoline engine using the KIVA II code. A strong controlled tumbling air motion was created in the cylinder, through a combination of a conventional pentroof four valve cylinder head, in conjunction with a piston having a stepped crown and offset combustion bowl. A range of injection strategies were employed to optimise combustion rate and exhaust emission (NOx and unburned hydrocarbons (fuel)), at two operating conditions - one with a stoichiometric air fuel mixture and the other with a lean mixture of 30:1 air/fuel ratio. Injection directed towards the piston bowl with a hollow cone jet, in a single pulse, has shown the best results regarding burned mass fraction and level of unburned HC. Fuel concentration, air motion, combustion characteristics and pollutants level are presented for lean and stoichiometric cases.

In this paper the effect of in-cylinder crevices formed by the piston cylinder clearance, above the first ring, and the spark plug cavity, on the entrapment of unburned fuel air mixture during the late compression, expansion and exhaust phases of a spark ignition engine cycle, have been simulated using the Computational Fluid Dynamic (CFD) code KIVA II. Two methods of fuelling the engine have been considered, the first involving the carburetion of a homogeneous fuel air mixture, and the second an attempt to simulate the effects of manifold injection of fuel droplets into the cylinder. The simulation is operative over the whole four stroke engine cycle, and shows the efflux of trapped hydrocarbon from crevices during the late expansion and exhaust phases of the engine cycle.

The process of autoignition in an internal combustion engine cylinder produces large amplitude high frequency gas pressure waves accompanied by significant increases in gas temperature and velocity, and as a consequence large convective heat fluxes to piston and cylinder surfaces. Extended exposure of these surfaces to autoignition, results in their damage through thermal fatigue, particularly in regions where small clearances between the piston and cylinder or cylinder head, lie in the path of the oscillatory gas pressure waves. The ability to predict spatial and temporal' variations in cylinder gas pressure, temperature and velocity during autoignition and hence obtain reasonable estimates of surface heat flux, makes it possible to assess levels of surface fatigue at critical zones of the piston and cylinder head, and hence improve their tolerance to autoignition.

The Computational Fluid Dynamics (CFD) code KIVA II has been applied to simulate the in-cylinder mean air motion (tumble) and turbulence levels in a motored 4-stroke single cylinder engine with pentroof combustion chamber geometry, having two inlet and two exhaust valves. In-cylinder flow during intake and compression strokes were simulated and a comparison between computational and experimental results were made. The mean turbulent kinetic energy and tumble ratio variation during the compression stroke obtained with CFD, have been compared with computational and experimental data from published literature. The simulation shows general similarity of flow structure and magnitude with published data on engines with similar geometry and initial flow conditions in the cylinder.

The Computational Fluid Dynamics (CFD) Code KIVA II has been applied to model combustion pressure oscillations in the Indirect Injection Diesel Engine. These oscillations are attenuated and transmitted by the engine structure to the surroundings as noise. The computational model was used to evaluate changes in design and operating characteristics of an engine, and the effect of these on the intensity of gas pressure oscillation. The results in general corroborate the trends of published experimental measurements of combustion noise. A 40% increase in grid resolution showed minor changes in the magnitude of cylinder pressure oscillation and approximately 0.5ø crank angle phase advance in the oscillation cycle compared with the grid used for the results presented here.

Combustion noise produced by the direct injection Diesel engine is a consequence of the dynamic equilibration of the high local pressures created in the combustion space following autoignition that stimulates the resonance modes, over a range of frequencies, of the gas contents of the engine cylinder. In this paper Computational Fluid Dynamics has been used to study the effects of changes in engine design and operating parameters that, from empirical experience, are known to influence the noise output of an engine, and explanations confirmed or given for well-established behaviour.

The CFD code KIVA has been applied to the simulation of the transient air-assisted fuel injection(AAFI) process, in which air and fuel at moderate pressures are mixed in an interior chamber of the injector before passing through a pintle valve into air at near ambient pressure in a cylinder. On passage through the pintle valve fuel is atomised. Because of the small dimensions of the flow passages within the injector, a very fine computational grid structure is used to accurately resolve the flow behaviour. Adopting an axisymmetric grid structure enables symmetry to be exploited. The computational results are validated with experimental data for fuel jet penetration and spread with time, obtained using Schlieren visualisation. The simulation of air blast atomisation in an engine cannot utilise the fine grid structure above because of the large computational resources required.

Particle Image Velocimetry (PIV) here has been used to measure the instantaneous velocity field within a realistic geometry motored single cylinder engine. Through-the-piston-crown illumination of a vertical plane bisecting the inlet and exhaust valves in a four valve pent roof combustion chamber and the use of a corrective optical system has for the first time allowed the velocity field in a vertical plane within a cylindrical bore to be quantified with PIV. Techniques are described which permit accurate and repeatable camera focusing, laser to engine synchronisation and seeding density control. Large scale motion observed at 180° ATDC has been interpreted as barrel swirl. Limitations of the current technique are discussed with respect to general in-cylinder applications.

Application of the engine CFD code KIVA II with the inclusion of the SHELL model for autoignition chemistry, and the discrete transfer radiation heat transfer model, has enabled the technically important problem of non luminous radiation from the major emitting species CO2 and H2O in the combustion products within the cylinder of a spark ignition engine to be considered as a combustion diagnostic aid, and also as a method of controlling individual cylinder Air/Fuel ratio. Results from a parametric study using CFD have been found to corroborate the experimental findings of other workers over a range of operating conditions including knock.

The CFD code KIVA is used in conjunction with a one-dimensional wave action program to simulate exhaust blowdown, in a study of the scavenging and combustion at different loads and constant engine speed, in a single cylinder 4 valved 2-stroke engine configuration, using in-cylinder fuel injection. Two combustion chamber geometries -- a stepped head and a pentroof, were used in this study. The stepped head geometry has a combustion chamber recessed in the cylinder head, and contains the intake valves. The vertical intake port configuration provides a well developed reversed loop flow in the engine cylinder. The pentroof combustion chamber is similar to those used in current 4 stroke engines(1)*. The computational study focuses on the effects of injector orientation, and the subsequent interaction between the fuel spray and ‘loop swirl’ of air in the engine cylinder, and on the resulting combustion characteristics and exhaust emissions.

This paper describes the development of an interactive model to simulate a direct injection diesel engine under both steady and transient conditions, based on the application of concurrent process computing methods. Initially the engine is modelled operating under steady conditions and induction, injection, air entrainment, fuel air mixing, combustion, emission and the mechanical friction processes are considered. The fuel pump, governor, engine crankshaft and external load dynamics are incorporated in the model to study the transient behaviour of a 2.5 litre D. I. engine and its associated load. Employing a two zone combustion model enables detailed performance and exhaust emissions predictions to be produced with economic use of computing time. The model written in FORTRAN is implemented in parallel on a transputer based concurrent computer by using the transputer operating system language OCCAM as a harness. Model predictions compare favourably with experimental results.

Transputer based concurrent processing has been applied to an extensive engine based Camputational Fluid Dynamics Code - KIVA, and demonstrated to be effective in terms of accuracy and cost. Scope exists for further adaptation of the code to the hardware to enhance performance. The use of concurrent processing with Transputer hardware merits consideration as a powerful and inexpensive engineering computational aid.

The concept of using low power microwave energy to efficiently assist the regeneration of ceramic monolith diesel particulate traps is explained. A prototype Microwave Assisted Regeneration (M.A.R.) system is presented and is demonstrated to work reasonably well on both bench and engine tests. The system, which is inexpensive, reliable and controllable, requires 1 kW of overall electrical power for a short period prior to trap regeneration. The M.A.R. technique presented merits further consideration as an alternative to other proposed vehicle trap regeneration systems.

A generalized model describing oxidation in a porous substance is developed and applied to the thermal regeneration of both monolithic and fibrous diesel particulate traps. With typical engine and trap data the regeneration process is analysed using the model. A parametric study demonstrates how the exhaust gas oxygen concentration, flow rate and initial trap particulate loading affect the regeneration time and peak trap temperatures. The model is shown to be in reasonable agreement with published experimental results.

An existing phenomenological model for combustion in quiescent and swirl assisted direct injection diesel engines developed by the authors has been extended to incorporate sub-models for exhaust smoke and nitrogen oxide emissions, based on published mechanisms for the formation of these two pollutants. Prediction of exhaust smoke and NO emission over a range of engine operating conditions including the effects of exhaust gas recirculation, has been undertaken for two quiescent and one swirl assisted engine. The predictions show well the general trend in emissions behaviour when compared with experiment, giving the possibility of using the model (with some calibration) for parametric studies.

Holographic interferometry has been applied to the study of concentration and temperature distribution in transient vapourising non-burning and burning fuel sprays in a quiescent bomb. Empirical relations have been obtained to describe the axial and radial variations of concentration in the vapourising non-burning spray, and to evaluate penetration and air entrainment of the free and wall jet regions of non-burning and burning sprays. The movement of the ‘tail’ of the spray in the post injection period has been studied. Finally, data are presented for species concentration and temperature profiles within the burning spray.

The absorption and desorption of fuel by cylinder lubricating oil films has been modelled using principles of mass transfer. Henry's Law for a dilute solution of fuel in oil is used to relate gas to liquid phase fuel concentrations. Mass transfer conductances in gas and liquid phases are considered, the former via use of Reynold's Analogy to engine heat transfer data, the latter through assuming molecular diffusion through an effective penetration depth of the oil film. Oxidation of desorbed fuel is assumed complete if the mean of burned gas and lubricating oil film temperatures is greater than 1100K,. Below this value the desorbed fuel is considered to contribute to hydrocarbon emissions. Comparison with engine test data corroborate the absorption/desorption hypothesis. The model indicates the equal importance of gas and liquid phase conductances.

A phenomenological model is presented for prediction of the combustion characteristics of a Quiescent Chamber Diesel engine. Predictions with the model have shown acceptable agreement with a range of experimental data. The major physical processes controlling combustion have been characterised, and the dominant role of air entrainment and turbulent mixing confirmed quantitatively.

The importance of turbulent energy dissipation rate on smoke and unburned hydrocarbon from Diesel engines has been identified quantitatively, the latter through a quenching which is predominant at idling and load operation of the engine. A parameter has been derived to assess the contribution to unburned hydrocarbon emissions various design and operational character-of an engine. Further experimental data required to support the general validity of parameter.