Search Results

Turbulence modeling for fuel spray simulation plays a prominent role in the understanding of the flow behavior in Internal Combustion Engines (ICEs). Currently, a lot of research work is actively spent on Large Eddy Simulation (LES) turbulence modeling as a replacement option of standard Reynolds averaged approaches in the Eulerian-Lagrangian spray modeling framework, due to its capability to accurately describe flow-induced spray variability and to the lower dependence of the results on the specific turbulence model and/or modeling coefficients. The introduction of LES poses, however, additional questions related to the implementation/adaptation of spray-related turbulence sources and to the rise of conflicting numerics and grid requirements between the Lagrangian and Eulerian parts of the simulated flow.

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.

The paper focuses on the development of a mesh moving method based on non-conformal topologically changing grids applied to the simulation of IC engines, where the prescribed motion of piston and valves is accomplished by rigidly translating the sub-domain representing the moving component. With respect to authors previous work, a more robust and efficient algorithm to handle the connectivity of non-conformal interfaces and a mesh-motion solver supporting multiple layer addition/removal of cells, to decouple the time-step constraints of the mesh motion and of the fluid dynamics, has been implemented as a C++ library to extend the already existing classes for dynamic mesh handling of the finite-volume, open-source CFD code OpenFOAM®. Other new features include automatic decomposition of large multiple region domains to preserve processors load balance with topological changes for parallel computations and additional tools for automatic preprocessing and case setup.

The dynamics and evolution of turbulent structures inside an engine-like geometry are investigated by means of Large Eddy Simulation. A simplified configuration consisting of a flat-top cylinder head with a fixed, axis-centered valve and low-speed piston has been simulated by the finite volume CFD code OpenFOAM®; the standard version of the software has been extended to include the compressible WALE subgrid-scale model, models for the generation of synthetic turbulence, some improvements to the mesh motion strategy and algorithms for LES data post-processing. In order to study both the initial transient and the quasi- steady operating conditions, ten complete engine cycles have been simulated. Phase and spatial averages have been performed over cycles three to ten in order to extract first and second moment of velocity; these quantities have then been used to validate the numerical procedure by comparison against experimental data.

The current development to set up an automatic procedure for automatic mesh generation and automatic mesh motion for internal combustion engine simulation in OpenFOAM®-2.2.x is here described. In order to automatically generate high-quality meshes of cylinder geometries, some technical issues need to be addressed: 1) automatic mesh generation should be able to control anisotropy and directionality of the grid; 2) during piston and valve motion, cells and faces must be introduced and removed without varying the overall area and volume of the cells, to avoid conservation errors. In particular, interpolation between discrete fields is frequent in computational physics: the use of adaptive and non-conformal meshes necessitates the interpolation of fields between different mesh regions. Interpolation problems also arise in areas such as model coupling, model initialization and visualisation.

The implementation and the combination of advanced boundary conditions and subgrid scale models for Large Eddy Simulation (LES) in the multi-dimensional open-source CFD code OpenFOAM® are presented. The goal is to perform reliable cold flow LES simulations in complex geometries, such as in the cylinders of internal combustion engines. The implementation of a boundary condition for synthetic turbulence generation upstream of the valve port and of the compressible formulation of the Wall-Adapting Local Eddy-viscosity sgs model (WALE) is described. The WALE model is based on the square of the velocity gradient tensor and it accounts for the effects of both the strain and the rotation rate of the smallest resolved turbulent fluctuations and it recovers the proper y₃ near-wall scaling for the eddy viscosity without requiring dynamic procedure; hence, it is supposed to be a very reliable model for ICE simulation.

In this paper a three-dimensional time-domain CFD approach has been employed to predict and analyze the acoustic attenuation performance of complex perforated muffler geometries, where strong 3D effects limit the validity of the use of one-dimensional models. A pressure pulse has been imposed at the inlet to excite the wave motion, while unsteady flow computation have been performed to acquire the time histories of the pressures upstream and downstream of the silencer. Pressures in the time domain have been then transformed to acoustic pressures in the frequency domain, to predict the transmission loss.

This paper describes a detailed analysis of the unsteady flows in the intake and exhaust systems of a modern four-cylinder, turbocharged Diesel engine by means of advanced numerical tools and experimental measurements. In particular, a 1D-3D integrated fluid dynamic model, based on the GASDYN (1D) and Lib-ICE (3D) codes, has been developed and applied for the schematization of the geometrical domain and the prediction of the wave motion in the whole intake and the exhaust systems, including the air cleaner, the intercooler, the after-treatment devices and the silencers. Firstly, a detailed 1D simulation has been carried out to predict the pressure pulses, average pressures and temperatures in several cross-sections of the pipe systems for different speeds and loads, considering the complex geometry of the air filter, the intake manifold, the intercooler and the exhaust manifold.

In the most elementary treatment of plane-wave reflection at the open end of a duct system, it is often assumed that the ends are pressure nodes. This implies that pressure is assumed as a constant at the open end termination and that steady flow boundary condition is supposed as instantaneously established. While this simplifying assumption seems reasonable, it does not consider any radiation of acoustic energy from the duct into the surrounding free space; hence, an error in the estimation of the effects of the flow on the acoustical response of an open-end duct occurs. If radiation is accounted, a complicated three-dimensional wave pattern near the duct end is established, which tends to readjust the exit pressure to its steady-flow level. This adjustment process is continually modified by further incident waves, so that the effective instantaneous boundary conditions which determine the reflected waves depend on the flow history.

Multi-dimensional simulation of hydrodynamics in full-scale wall-flow Diesel Particulate Filters by GpenFQAM®, an open-source C++ object-oriented CFD code, is presented. A new fast and efficient parallel numerical solver has been developed by authors to simulate flows through porous media and it has been tested for the simulation of diesel particulate filters; errors caused by discretization of filter monoliths have been corrected by the formulation of a correction factor, that has been included in the solver. A set of experimental data, available from literature, has been used for code validation.

A new fast and efficient parallel numerical solver for reacting and compressible flows through porous media has been developed in the OpenFOAM® (Open Field Operation and Manipulation) CFD Toolbox. With respect to the macroscopic model for porous media originally available in OpenFOAM®, a different mathematical approach has been followed: the new implemented solver makes use of the physical normal components resulting from the velocity expansion in the unit orthogonal vector basis to compute the Darcy pressure drop across the porous medium. Also, an additional sink term to account for the increased flow friction over the porous wall has been included into the momentum equation. In the new solver, the pressure correction equation is still able to achieve a faster convergency at very low permeability of the medium, also when it is associated with grid non-orthogonality.

A computational, three-dimensional approach to investigate the behavior of diesel soot particles in the micro-channels of wall-flow Diesel Particulate Filters is presented. The KIVA3V CFD code, already extended to solve the 2D conservation equations for porous media materials [1], has been enhanced to solve in 2-D and 3-D the governing equations for reacting and compressible flows through porous media in non axes-symmetric geometries. With respect to previous work [1], a different mathematical approach has been followed in the implementation of the numerical solver for porous media, in order to achieve a faster convergency as source terms were added to the governing equations. The Darcy pressure drop has been included in the Navier-Stokes equations and the energy equation has been extended to account for the thermal exchange between the gas flow and the porous wall.

This work describes the development, application and coupling of two different numerical codes, respectively based on a 1D (Gasdyn) and 3D (OpenFOAM) schematization of the geometrical domain. They have been adopted for the prediction of the wave motion inside the intake and the exhaust systems of internal combustion engines. The HLLC Riemann solver has been implemented both in the CFD and the 1D codes to solve the Euler system of equations, in order to operate with the same solver on the different calculation domains. Moreover, the HLLC solver has been applied to treat the boundary conditions at the interface between the two domains, in such a way to allow the propagation of flow disuniformities through the domain interface, without affecting the solution accuracy. The hybrid approach was used for the simulation of two different test cases: a complex 5 into 1 pipe junction of a high performance V10 engine and a Venturi tube plus a Helmholtz resonator of a single cylinder S.I. engine.

An innovative approach for the simulation of the interaction between the software to be implemented in the Electronic Control Unit (ECU) of the vehicle and the engine is described. The aim was to perform a complete simulation of the engine coupled with the ECU, in order to support multi-disciplinary development and to enable engineers to verify and validate control models in early development stages, reducing costs by performing fewer live engine tests. Also, the simultaneous simulation of the models allows to study their interaction, thus allowing an early exploration of the possible design choices over multiple disciplines. A first prototype of the coupling has been implemented, with an emphasis on realizing a common notion of time and a proper treatment of data exchange between the control model and the engine model.

The present work focuses on the simulation of the hydrodynamics, transient filtration/loading and catalytic/NO2-assisted regeneration of Diesel after-treatment systems. A 1D unsteady model for compressible and reacting flows for the numerical simulation of the behavior of Diesel Oxidation Catalysts (DOCs) and Diesel Particulate Filters (DPFs) has been developed. The numerical model is able to keep track of the amount of soot in the flow; the increasing of back-pressure through the exhaust system (mainly due to the Diesel Particulate Filter) can be predicted by the calculation of the permeability variation of the porous wall, as the soot particles goes inside the DPF. A sub-model for the regeneration of the collected soot has been developed: the collected particulate is oxidized by the Oxygen (O2) and by the Nitrogen Dioxide (NO2).

A new approach for the fluid-dynamic simulation of the Diesel Particulate Filters (DPF) has been developed. A mathematical model has been formulated as a system of nonlinear partial differential equations describing the conservation of mass, momentum and energy for unsteady, compressible and reacting flows, in order to predict the hydrodynamic characteristics of the DPF and to study the soot deposition mechanism. In particular, the mass conservation equations have been solved for each chemical component considered, and the advection of information concerning the chemical composition of the gas has been figured out for each computational mesh. A sub-model for the prediction of the soot cake formation has been developed and predictions of soot deposition profiles have been calculated for different loading conditions. The results of the simulations, namely the calculated pressure drop, have been compared with the experimental data.

A two dimensional CFD model has been developed to study the mechanism of soot deposition on the porous wall surface in a Diesel Particulate Filter (DPF). The goal is to improve understanding of the soot deposition and its interaction with the hydrodynamic behaviour of the device. The KIVA3V CFD code and pre-processor were modified to simulate a single channel in a DPF. The code was extended to solve the conservation equations in porous media materials, to account for the sticking of particles on a porous surface and to evaluate the increasing resistance to the flow as the soot inside the trap accumulates. The code is already configured to track Lagrangian particles and these were modified to represent the soot particles in the flow. The code pre-processor was modified to allow definition of a double-symmetric geometry and to specify porous cells.