Search Results

A numerical investigation on a series of Diesel spray experiments in constant-volume vessels is proposed. Non reacting conditions were used to assess the spray models and to determine the grid size required to correctly predict the fuel-air mixture formation process. To this end, not only computed liquid and vapor penetrations were compared with experimental data, but also a detailed comparison between computed and experimental mixture fraction distributions was performed at different distances from the injector. Grid dependency was reduced by introducing an Adaptive Local Mesh Refinement technique (ALMR) with an arbitrary level of refinement. Once the capabilities of the current implemented spray models have been assessed, reacting conditions at different ambient densities and temperatures were considered. A Perfectly Stirred Reactor (PSR) combustion model, based on a direct integration of complex chemistry mechanisms over a homogenous cell, was adopted.

Detailed chemistry and turbulence-chemistry interaction need to be properly taken into account for a realistic combustion simulation of IC engines where advanced combustion modes, multiple injections and stratified combustion involve a wide range of combustion regimes and require a proper description of several phenomena such as auto-ignition, flame stabilization, diffusive combustion and lean premixed flame propagation. To this end, different approaches are applied and the most used ones rely on the well-stirred reactor or flamelet assumption. However, well-mixed models do not describe correctly flame structure, while unsteady flamelet models cannot easily predict premixed flame propagation and triple flames. A possible alternative for them is represented by transported probability density functions (PDF) methods, which have been applied widely and effectively for modeling turbulent reacting flows under a wide range of combustion regimes.

A new multi-zone model for the simulation of HCCI engine is here presented. The model includes laminar and turbulent diffusion and conduction exchange between the zones and the last improvements on the numerical aspects. Furthermore, a new strategy for the zone discretization is presented, which allows a better description of the near-wall zones. The aim of the work is to provide a fast and reliable model for carrying out chemical analysis with detailed kinetic schemes. A preliminary sensitivity analysis allows to verify that 10 zones are a convenient number for a good compromise between the computational effort and the description accuracy. The multi-zone predictions are then compared with the CFD ones to find the effective turbulence parameters, with the aim to describe the near-wall phenomena, both in a reactive and non-reactive cases.

Swirling flows are very dominant in applied technical problems, especially in IC engines, and their prediction requires rather sophisticated modeling. An adaptive low-pass filtering procedure for the modeled turbulent length and time scales is derived and applied to Menter' original k - ω SST turbulence model. The modeled length and time scales are compared to what can potentially be resolved by the computational grid and time step. If the modeled scales are larger than the resolvable scales, the resolvable scales will replace the modeled scales in the formulation of the eddy viscosity; therefore, the filtering technique helps the turbulence model to adapt in accordance with the mesh resolution and the scales to capture.

High-speed OH chemiluminescence imaging is used to measure the lift-off length of diesel sprays in an optical heavy-duty diesel engine of 2 L displacement operated at 1200 rpm and 5 bar IMEP. Stereoscopic images are acquired at two different wavelengths (310 and 330 nm). Subtraction of pairwise images helps reducing the background coming from natural soot incandescence in the OH chemiluminescence images. Intake air temperature (343 to 403 K), motored top dead center density (18 to 22 kg/m3), fuel injection pressure (150 to 250 MPa), intake oxygen concentration (17 to 21 %vol) and nozzle diameter (0.1 and 0.14 mm) are varied and a nonlinear regression model is derived from the experimental results to describe stabilized lift-off length as function of the experimental factors. The lift-off length follows the general trends that are known from spray vessel investigations, but the strength of the dependence on certain variables deviates strongly from those studies.

The influence of jet-flow and jet-jet interactions on the lift-off length of diesel jets are investigated in an optically accessible heavy-duty diesel engine. High-speed OH chemiluminescence imaging technique is employed to capture the transient evolution of the lift-off length up to its stabilization. The engine is operated at 1200 rpm and at a constant load of 5 bar IMEP. Decreasing the inter-jet spacing shortens the liftoff length of the jet. A strong interaction is also observed between the bulk in-cylinder gas temperature and the inter-jet spacing. The in-cylinder swirl level only has a limited influence on the final lift-off length position. Increasing the inter-jet spacing is found to reduce the magnitude of the cycle-to-cycle variations of the lift-off length.

The modeling of fuel sprays under well-characterized conditions relevant for heavy-duty Diesel engine applications, allows for detailed analyses of individual phenomena aimed at improving emission formation and fuel consumption. However, the complexity of a reacting fuel spray under heavy-duty conditions currently prohibits direct simulation. Using a systematic approach, we extrapolate available spray models to the desired conditions without inclusion of chemical reactions. For validation, experimental techniques are utilized to characterize inert sprays of n-dodecane in a high-pressure, high-temperature (900 K) constant volume vessel with full optical access. The liquid fuel spray is studied using high-speed diffused back-illumination for conditions with different densities (22.8 and 40 kg/m3) and injection pressures (150, 80 and 160 MPa), using a 0.205-mm orifice diameter nozzle.

Modeling plume interaction and collapse for direct-injection gasoline sprays is important because of its impact on fuel-air mixing and engine performance. Nevertheless, the aerodynamic interaction between plumes and the complicated two-phase coupling of the evaporating spray has shown to be notoriously difficult to predict. With the availability of high-speed (100 kHz) Particle Image Velocimetry (PIV) experimental data, we compare velocity field predictions between plumes to observe the full temporal evolution leading up to plume merging and complete spray collapse. The target “Spray G” operating conditions of the Engine Combustion Network (ECN) is the focus of the work, including parametric variations in ambient gas temperature. We apply both LES and RANS spray models in different CFD platforms, outlining features of the spray that are most critical to model in order to predict the correct aerodynamics and fuel-air mixing.

The flow field resulting from injecting a gas jet into a crossflow confined in a narrow square duct has been studied under steady regime using schlieren imaging and laser Doppler velocimetry (LDV). This transparent duct is intended to simulate the intake port of an internal combustion engine fueled by gaseous mixture, and the jet is issued from a round nozzle. The schlieren images show that the relative small size of the duct would confine the development of the transverse jet, and the interaction among jet and sidewalls strongly influences the mixing process between jet and crossflow. The mean velocity and turbulence fields have been studied in detail through LDV measurements, at both center plane and several cross sections. The well-known flow feature formed by a counter rotating vortex pair (CVP) has been observed, which starts to appear at the jet exit section and persists far downstream contributing to enhancing mixing process.

This paper is focused on the development and application of a CFD methodology that can be applied to predict the fuel-air mixing process in stratified charge, sparkignition engines. The Eulerian-Lagrangian approach was used to model the spray evolution together with a liquid film model that properly takes into account its effects on the fuel-air mixing process into account. However, numerical simulation of stratified combustion in SI engines is a very challenging task for CFD modeling, due to the complex interaction of different physical phenomena involving turbulent, reacting and multiphase flows evolving inside a moving geometry. Hence, for a proper assessment of the different sub-models involved a detailed set of experimental optical data is required. To this end, a large experimental database was built by the authors.

In this paper, an experimental and numerical analysis of combustion process and knock occurrence in a small displacement spark-ignition engine is presented. A wide experimental campaign is preliminarily carried out in order to fully characterize the engine behavior in different operating conditions. In particular, the acquisition of a large number of consecutive pressure cycle is realized to analyze the Cyclic Variability (CV) effects in terms of Indicated Mean Effective Pressure (IMEP) Coefficient of Variation (CoV). The spark advance is also changed up to incipient knocking conditions, basing on a proper definition of a knock index. The latter is estimated through the decomposition and the FFT analysis of the instantaneous pressure cycles. Contemporary, a quasi-dimensional combustion and knock model, included within a whole engine one-dimensional (1D) modeling framework, are developed. Combustion and knock models are extended to include the CV effects, too.

The fluid-mechanic behaviour of straight-sided and re-entrant chamber geometries has been studied using laser doppler anemometry (LDA) technique. Measurements have been carried out during the compression stroke in a direct injection diesel engine, representative of medium size family, operating at 1000 rpm under motored conditions. The mean motion and turbulence intensity have been computed using a filtering procedure on the LDA data. Using the second version of KIVA code, the air flow field evolution during the same crank angle period has been also computed. To perform proper comparisons between measured and computed values of mean velocity and turbulence intensity, a careful choice of the initial conditions for computations has been performed. Reasonable agreement has been found between computed and measured mean swirl velocities for both combustion chamber geometries tested. On the contrary, the computed turbulence intensities underestimate those measured.

Objective of this work is the incorporation of the flame stretch effects in an Eulerian-Lagrangian model for premixed SI combustion in order to describe ignition and flame propagation under highly inhomogeneous flow conditions. To this end, effects of energy transfer from electrical circuit and turbulent flame propagation were fully decoupled. The first ones are taken into account by Lagrangian particles whose main purpose is to generate an initial burned field in the computational domain. Turbulent flame development is instead considered only in the Eulerian gas phase for a better description of the local flow effects. To improve the model predictive capabilities, flame stretch effects were introduced in the turbulent combustion model by using formulations coming from the asymptotic theory and recently verified by means of DNS studies. Experiments carried out at Michigan Tech University in a pressurized, constant-volume vessel were used to validate the proposed approach.

The OpenFOAM® CFD methodology is nowadays employed for simulation in internal combustion engines and a lot of work has been done for an appropriate description of all complex phenomena. At the moment in the RANS turbulence models available in the OpenFOAM® toolbox the turbulence modulation is not yet included, and the present work analyzes the predictive capabilities of the code in simulating high injection pressure fuel sprays after modeling the influence of the dispersed phase on the turbulence structure. Different experiments were employed for the validation. At first, non-evaporating diesel spray was considered in a constant volume and quiescent vessel. The validation was performed via the available experimental spray evolution in terms of penetrations and spatial/temporal fuel distributions. Then the Sandia combustion chamber was chosen for diesel spray simulation in non-reacting conditions.

A three dimensional numerical simulation of the ECN “Spray A” is presented. Both primary and secondary breakup of the spray are included. The fuel is n-Dodecane. The n-Dodecane kinetic mechanism is modeled using a skeletal mechanism that consists of 103 species and 370 reactions [9]. The kinetic mechanism is computationally heavy when coupled with three dimensional numerical simulations. Multidimensional chemistry coordinate mapping (CCM) approach is used to speedup the simulation. CCM involves two-way mapping between CFD cells and a discretized multidimensional thermodynamic space, the so called multidimensional chemistry coordinate space. In the text, the cells in the discretized multidimensional thermodynamic space are called zone to discriminate them from the CFD cells. In this way, the CFD cells which are at the similar thermodynamic state are identified and grouped into a unique zone. The stiff ODEs operates only on the zones containing at least one CFD cell.

The s.i. combustion process and its corresponding pollutant formation are investigated by means of a quasiD approach and a CFD model. This work has been motivated by the need to better understand the reliability of such models and to assess their accuracies with respect to the prediction of engine performances and emissions. An extended dissertation about the fundamental mechanisms governing the pollutant formation in the turbulent premixed combustion which characterizes the s.i. engines is given. The conclusion of such analysis is the definition of a new reduced chemical scheme, based on the application of partial-equilibrium and steady-state assumptions for the radicals and the solution of a transport equation for each specie which is kinetically controlled. For this purpose the CFD code OpenFOAM [1, 2, 3] and the thermo-fluid dynamic code GASDYN [4, 5] have been applied and enhanced.

In-cylinder cycle-resolved LDV measurements have been made in a diesel engine having a high-squish re-entrant combustion chamber with compression ratio of 21:1. The engine has been motored in the range of 1000 to 3000 rpm thanks to the use of self-lubricating seeding particles. Conventional ensemble-averaging and filtering techniques have been used for analyzing instantaneous velocity data obtained at two points along a diameter located in a horizontal plane at 5 mm below the engine head. The dependence of the mean motion and turbulence on engine speed has been evaluated. The effect of cut-off frequency selection on turbulence values has been also analyzed. Moreover, the Kolmogorov's -5/3 power domain has been investigated in detail by spectral analysis on the instantaneous velocity data.

In-cylinder measurements of turbulent integral length scales, carried out during the last 60 degrees of the compression stroke at 600 and 1,000 rpm by a two-probe volume LDV system, were used to assess the capability of the k-ε model used in KIVA-II code. The objective of the paper is to address the following question: what is the most reasonable definition of turbulent length scale in the k-ε model for engine applications? The answer derived from the comparison between KIVA predictions and experiments that showed a fair agreement between the computed turbulent length scale and the measured lateral integral length scale. The agreement is a result of proper choice of the initial swirl ratio and turbulent kinetic energy at inlet valve closure (IVC) by taking into account the LDV measurements and the value of the constant Cμε in the k-ε model equations that relates the turbulent length scale to k and ε.

Tangential component of velocity and turbulence were measured in three locations in the re-entrant combustion chamber of a motored single-cylinder d.i. Diesel engine (0.435 liter, 21:1 compression ratio) using a Laser Doppler Velocimetry system. Moreover, a modified LDV system with two-probe volume was used to measure directly lateral integral length scales of the velocity tangential component at two engine speeds. The measurements were made on a horizontal plane at 5 mm below the engine head from 100 degrees before TDC to 60 degrees after TDC of both the compression and expansion strokes. The engine was motored at 1,000 and 1,500 rpm respectively. An ensemble-averaging technique was performed to analyze the instantaneous velocity information supplied by two Burst Spectrum Analyzers. The lateral integral length scale was obtained from the integral of the spatial correlation coefficient of the velocity fluctuation for different separation.

The present paper aims at developing a general method to estimate integral and microtime scales of turbulent in-cylinder flow field in reciprocating engines. The ensemble average technique was used to compute the integral time scale from the single point time autocorrelation function, whereas the microtime scale, representative of the most rapid changes that occur in the fluctuation, was computed as the intercept of the parabola that matches the autocorrelation function at the origin. Further, the microtime scale was also estimated by spectral analysis through the energy spectral density function of the ensemble turbulent fluctuation and the results obtained by the two methods were compared. The procedures were applied to the tangential component of the instantaneous velocity data collected, at different engine speeds (1,000, 1,500, 2,000 rpm), within a motored d.i. diesel engine equipped with a re-entrant combustion chamber, using the Laser Doppler Anemometry (LDA) technique.