Search Results

In internal combustion engines, liquid fuel injection is one of the most prevalent means of fuel delivery and air-fuel mixture preparation. The behavior of the fuel spray and wall film is a key factor in determining air-fuel mixing and hence combustion and emissions. A comprehensive model for the liquid fuel spray has been developed in conjunction with the one-dimensional gas flow code WAVE. The model includes droplet dynamics and evaporation, spray-wall impingement, wall film dynamics and evaporation. The fuel injector can be placed in the manifold, inlet port or cylinder. Liquid fuel droplets are injected with a prescribed size distribution, and their subsequent movement and vaporization are modeled via the discrete particle approach, frequently used in multi-dimensional CFD codes. This approach ensures conservation of mass, momentum and energy between the gas and liquid phases.

CFD simulations have been conducted to study Homogeneous Charge Compression Ignition (HCCI) combustion in a single-cylinder poppet valve Direct-Injection 2-stroke gasoline engine using the Ricardo engine CFD code VECTIS. Multi-domain, multi-cycle simulations were performed to determine the trapped conditions, in particular the high residual gas content and charge distribution. The HCCI ignition and combustion were modelled using the Livengood-Wu ignition integral combined with the Ricardo Two-Zone Flamelet combustion model. Good agreements have been achieved between simulation and engine test data. The ignition points and in-cylinder pressures were predicted with good accuracy. The paper demonstrates that in 2-stroke engine operation the in-cylinder charge inhomogeneity is significant. The multi-domain, multi-cycle CFD simulation is an effective approach to study the charge inhomogeneity and its effect on HCCI engine combustion.

In this paper, a new two-zone flamelet model is suggested. In the combustion model, each cell is divided by the flame front to two zones: unburned zone and burned zone. The unburned zone consists of air, fuel vapour and residual gases, whilst the burned zone contains combustion products. The unburned zone is further divided into two regions: segregate region and fully mixed region. The combustion is decoupled as two sequential events: mixing and burning. The turbulent mixing is governed by the large eddy structure, taking the effect of fuel drop spacing into consideration. The turbulent burning rate is further decomposed into three terms: laminar burning velocity for combustion chemistry, turbulence enhanced burning rate and flame strain factor for flame quenching. The turbulent burning rate is evaluated based on fractal geometry and basic dimensional analysis of turbulent flame.

For high Reynolds number flows, the fine structure turbulence is universal in the inertial subrange according to the Kolmogorov local isotropy theory; while for low-to-moderate Reynolds number flows, the turbulence tends to display a common structure for a particular type of shear flows from the Townsend structural similarity. The power laws for the turbulent flows reflect the similarity in the turbulence fine structure. The paper discusses the relationship between drop breakup and turbulence fine structure. A predictive formula for drop breakup in the turbulent flow is suggested based on the energy balance between turbulence separating kinetic energy (i.e. turbulence structure function) and surface tension. The model has been validated on a number of empirical relations for drop breakup in different turbulent flows. It is known that the drop size distribution in shear flows can be approximately expressed by a lognormal distribution.

A phenomenological model is proposed in this paper for simulating the effect of wall friction in turbulent pipe flows. The flow in this model is divided into two regions: wall region and core region. The turbulent flow in the core region obeys the law of the wake, while the flow in the wall region follows the law of the wall. In the overlapping region they interact and behave like a dynamic system of two degrees of freedom. The model can predict the variation of friction coefficient and phase shift. The model is formulated in a simple form, and can be easily integrated into one-dimensional flow calculations. The model has been validated against published measured data, including a pulsating flow without flow reversal and a reciprocating flow with flow reversal. The model predictions have shown good agreement with measurements.

A simple unified fuel spray model is proposed. The model covers the main physical processes of the fuel spray: jet spray penetration, fuel atomization and evaporation, air entrainment and mixing. The model is aimed at dealing with a wide range of the jet spray regimes: free jet, normal and oblique jet-wall impingement, wall jet and swirling jet. A new formula of the jet spray penetration is suggested based on the jet entrainment law and momentum conservation. For different jet regimes the entrainment coefficients are determined from the basic turbulence relation of the jet flow. A dynamic model of the jet spray atomization is described based on the energy conservation. The jet atomization model can evaluate the variation of jet droplet Sauter diameter with time. For the oblique jet-wall impingement an approximate solution of angular distribution of momentum for outflow jet sheet is given in a cubic polynomial form based on the mass and momentum conservation of inviscid jet flow.

This paper proposes a one-dimensional model for in-cylinder heat convection based on the boundary layer theory. The model describes the temporal variations of the velocity boundary layer and thermal boundary layer separately. It is assumed that the behaviour of the boundary layers is quasi-steady: as a whole the boundary layers change with time and wall location, while inside the boundary layers the velocity and temperature profiles follow the steady-state power law. The model integrates the full one-dimensional thin-shear-layer equations with the F-factor correction suggested by Bradshaw and the revised Kutateladze and Leont'ev relation of the velocity and thermal boundary layers. The F factor can compensate for the model error in the curved flow. The revised Kutateladze and Leont'ev relation can correctly reflect the heat transfer mechanism. The model has been validated by a simple approach, using a fixed bulk flow velocity and a surface radius of curvature.

A comprehensive general purpose engine simulation model has been successfully developed. This paper reports on an investigation undertaken to compare the accuracy and computational efficiency of four alternative methods for modelling the gas flow and performance in internal combustion engines. The comparison is based on the filling-and-emptying method, the acoustic method, the Lax-Wendroff two-stage difference method and the Harten-Lax-Leer upstream method, using a unified treatment for the boundary conditions. The filling-and-emptying method is the quickest method among these four methods, giving performance predictions with reasonably good accuracy, and is suitable for simulating engines using not highly tuned gas exchange systems. Based on the linearized Euler equations, the acoustic method is capable of describing time-varying pressure distributions along a pipe.

This paper describes the development and validation of a refined combustion model for turbulent flame propagation in SI engines. In this model the original differential equation of flame front entrainment remains, but a new difference equation of burning rate is introduced to replace the original differential equation. The model fully embodies the Tennekes-Chomiak mechanism of premixed turbulent combustion, presenting a flame thinner than the original model. A spark ignition model is also suggested. The model treats the electrical spark ignition as a diffusion wave to calculate the entrainment enhancement by the spark ignition. The simple formula of the initial flame propagation has been validated by measured results. Incorporated with this spark ignition model, the three-stage model of premixed turbulent flame velocity, published earlier by the authors, is used in the combustion model.

This paper describes a simple computer model developed for analysing the squish motion in the pentroof combustion chamber. The model is used to predict the effects of various design parameters such as the compression ratio, pentroof angles and compression plane margin on squish velocity. The results of the parametric study are presented together with the mathematical description of the model. A number of guidelines are also given for the design of pentroof chamber.

Based on dimensional reasoning of classical turbulence theory, a phenomenological turbulent flame velocity model is proposed. The model explains the whole process of flame development from the inception, through the growth stages to saturation in a simple and unified form. Two sets of mathematical formulae are presented. In formulating the model, the ratios of turbulent to laminar flame velocity are given as functions of the Reynolds numbers based on the integral scale and Taylor microscale, respectively. Comparison of fully developed turbulent flame velocity with experimental results shows good agreement. This model can be easily incorporated in phenomenological models of spark-ignition engines for simulation studies.