Search Results

Remote emission sensing (RES) is a non-intrusive measurement method based on absorption spectroscopy, which allows for the determination of pollutant concentrations in vehicle exhaust plumes. By measuring the absorption of the exhaust plume from the roadside using a light/laser barrier, concentration ratios of pollutants, such as nitrogen oxides to carbon dioxide, can be estimated. Computational fluid dynamics (CFD) has been employed to simulate vehicle exhaust plumes due to uncertainties in RES capabilities. In a previous study, Unsteady Reynolds-Averaged Navier-Stokes (URANS) simulations were conducted to investigate the dispersion of vehicle exhaust plumes under various ambient/driving conditions and provide insights for RES applications. However, the accuracy of these simulations can be further improved. Therefore, this study focuses on enhancing the simulation accuracy by employing large eddy simulations (LES).

A carburetted, spark ignited gasoline fuelled engine of a 5 kW rated power generator was converted to run on hydrogen. As opposed to large parts of current research, the engine conversion’s foremost goal was not to maximise efficiency and power output but rather to find a cost-effective and low-complexity conversion approach to introduce clean fuels to existing engines. To allow for the increased volumetric fuel flow, the riser of the original carburettor was enlarged. The hydrogen flow into the venturi was metered with the help of a pressure regulator from a widely available conversion kit. The effects of different hydrogen-fuel-feed pressures on engine performance, operational stability and emission levels were examined experimentally. It was found that the hydrogen-line pressure before startup has to be set precisely (±5 mbar) to allow for stable and emission free operation.

Methane abatement in the exhaust gas of natural gas engines is much more challenging in respect to the oxidation of other higher order hydrocarbons. Under steady state λ sweep, the methane conversion efficiency is high at exact stoichiometric, and decreases steeply under both slightly rich and slightly lean conditions. Transient lean to rich transitions can improve methane conversion at the rich side. Previous experimental work has attributed the enhanced methane conversion to activation of methane steam reforming. The steam reforming rate, however, attenuates over time and the methane conversion rate gradually converges to the low steady state values. In this work, a reactor model is established to predict steady state and transient transition characteristics of a three-way catalyst (TWC) mounted in the exhaust of a natural gas heavy-duty engine.

The selective catalytic reduction has seen widespread adoption as the best technology to reduce the NOx emissions from internal combustion engines, particularly for Diesels. This technology uses ammonia as a reducing agent, which is obtained injecting an ammonia carrier into the exhaust gas stream. The dosing of the ammonia carrier, usually AdBlue, is the major concern during the design and engine calibration phases, since the interaction between the injected liquid and the components of the exhaust system can lead to the undesired formation of solid deposits. To avoid this, the thermal and kinematic interaction between the spray and the components of the after treatment system (ATS) must be modeled accurately. In this work, the authors developed a Conjugate Heat Transfer (CHT) framework to model the kinetic and thermal interaction among the spray, the eventual liquid layer and the pipe walls.

The reduction of CO2 emissions in transport and power generation is currently a key challenge. One particular opportunity of CO2 reduction is the introduction of low CO2 or even CO2 neutral fuels. The combustion characteristics of such fuels are different and require engine settings modification. In addition, emissions characteristics differ significantly among different fuels. In the present study a one cylinder diesel engine was operated with conventional diesel, hydrogenated vegetable oil (HVO) and polyoxymethyl dimethyl ether (OME) as well as a series of blends. Particle filter segments were positioned in the exhaust of the engine and loaded with particles originating from the combustion of these fuels. The filter segments have been regenerated individually in a specifically designed and developed controlled temperature soot oxidation apparatus.

The accurate prediction of pollutant emissions generated by IC engines is a key aspect to guarantee the respect of the emission regulation legislation. This paper describes the approach followed by the authors to achieve a strict numerical coupling of two different 1D modeling tools in a co-simulation environment, aiming at a reliable calculation of engine-out and tailpipe emissions. The main idea is to allow an accurate 1D simulation of the unsteady flows and wave motion inside the intake and exhaust systems, without resorting to an over-simplified geometrical discretization, and to rely on advanced thermodynamic combustion models and kinetic sub-models for the calculation of cylinder-out emissions. A specific fluid dynamic approach is then used to track the chemical composition along the exhaust duct-system, in order to evaluate the conversion efficiency of after-treatment devices, such as TWC, GPF, DPF, DOC, SCR and so on.

A Euro6 gasoline light duty vehicle has been tested at the engine dynamometer and the emissions have been analyzed upstream and downstream the Three-Way-Catalyst (TWC) during a WLTC cycle. Catalyst simulations have been used for assessing the processes inside the catalytic converter using a reaction scheme based on 19 brutto reactions (direct oxidation and reduction, selective catalytic reductions with CO, C3H6 and H2, steam reforming, water-gas shift and bulk ceria as well as surface ceria reactions). The reactions have been parameterized in order to best approximate the measurements. Based on the reactions taken into account, the real vehicle emissions can be predicted with good accuracy. The simulations show that the cycle emissions comprise mainly the cold start contribution as well as discrete emission break-through events during transients. During cold start no reactions are evident in the catalyst before the temperature of the gas entering the catalyst reaches 270°C.

The transient heat transfer behavior of an automotive catalytic converter has been simulated with OpenFOAM in 1D. The model takes into consideration the gas-solid convective heat transfer, axial wall conduction and heat capacity effects in the solid phase, but also the chemical reactions of CO oxidation, based on simplified Arrhenius and Langmuir-Hinshelwood approaches. The associated parameters are the results of data in literature tuned by experiments. Simplified cases of constant flow rates and gas temperatures in the catalyst inflow have been chosen for a comprehensive analysis of the heat and mass transfer phenomena. The impact of inlet flow temperatures and inlet flow rates on the heat up characteristics as well as in the CO emissions have been quantified. A dimensional analysis is proposed and dimensionless temperature difference and space-time coordinates are introduced.

Aqueous Urea is a non-toxic and stable ammonia carrier and its injection and mixing represent the basis for the most common de-NOx technology for mobile applications. The reactant feed preparation process is defined by evaporation, thermolysis and hydrolysis of the liquid mixture upstream the Selective Catalytic Reduction reactor, and it is strongly dependent on the interaction between spray and gaseous flow. Low-pressure driven injectors are the common industrial standard for these applications, and their behavior in almost-ambient pressure cross flows is significantly different from any in-cylinder application. For this reason, two substantially different injectors in terms of geometry and design are experimentally studied, characterizing drop sizes and velocities through Phase Doppler Anemometry (PDA) and liquid mass spatial distribution through Shadow Imaging (SI).

The application of liquid aqueous Urea Solution (AUS) as reductant in SCR exhaust after-treatment systems is now a commonly accepted industry standard. Unfortunately, less acceptable are the associated difficulties caused by incomplete decomposition of the liquid, resulting in solid deposits which accumulate in the exhaust pipe downstream of the dosing components. The correct prediction of the spray pattern and, therefore, the spray impact on the walls is a key feature for the system optimization. A mechanical patternator, designed on the basis of CFD performance assessment, involving a Lagrangian representation of the dispersed liquid fully coupled with a 3D Eulerian description of the carrier phase, has been built and used to measure the spray mass distribution.

The injection process of urea-water solution (AdBlue) determines initial conditions for reactions and catalysis and is fundamentally responsible for optimal operation of selective catalytic reduction (SCR) systems. The spray characteristics of four, commercially available, injectors (one air-assisted and three pressure-driven with different nozzle-hole configurations) are investigated with non-intrusive measuring techniques. Injection occurred in the crossflow of a channel blowing preheated air in an exhaust duct similar configuration. The effect of several gas temperatures and flows on the spray propagation and entrainment has been extensively studied by shadow imaging. Shadow images, in addition, show that the spray of the pressure-driven injectors is only marginally affected by the gas crossflow. In contrast, the air assisted spray is strongly deflected by the gas, the effect increasing with increasing gas flow.

Particulate matter in diesel exhaust is captured in diesel particulate filters (DPFs). Since increased load in the filter and thus increased pressure drop deteriorates the engine performance, the filter load of the DPF has to be removed during a process referred to as regeneration. Measures for successful regeneration aim at accelerating soot oxidation and increase fuel consumption. Regeneration lay-out and thus fuel consumption increase is strongly depending on the oxidation behavior of soot. The aim of the present study is the investigation of soot oxidation characteristics. Therefore particle filters have been loaded with soot using the exhaust gas of small heavy duty vehicle operated under defined conditions on an engine dynamometer. The particle filters have been then dismantled and fragmented on their constituting segments. Each filter segment has been regenerated individually in a specifically designed test bench.

Simulations for a pressure-assisted multi-stream injector designed for urea-dosing in a selective catalytic reduction (SCR) exhaust gas system have been carried out and compared to measurements taken in an optically accessible high-fidelity flow test rig. The experimental data comprises four different combinations of mass flow rate and temperature for the gas stream with unchanged injection parameters for the spray. First, a parametric study is carried out to determine the importance of various spray sub-models, including atomization, spray-wall interaction, buoyancy as well as droplet coalescence. Optimal parameters are determined using experimental data for one reference operating condition.

Particulate matter (PM) captured in diesel particulate filters (DPF) consists of: (a) soot, the product of incomplete combustion of the fuel and (b) ash, produced by combustion of lubricating oil plus minor amounts of metal components in the fuel. Among the various types of DPFs, most efficient are the so-called wall flow filters, where the exhaust gas is forced to pass through porous walls of adjacent channels, which are plugged alternately at their opposite ends. Accumulation of PM in DPFs leads to increasing pressure drop across the filter. Since increased PM load in the filter and thus increased pressure drop across the filter deteriorates the engine performance, the filter load of the DPF has to be periodically removed during a process referred to as regeneration. During the regeneration process, soot PM captured in the DPF is expected to be oxidized. The temperature needed for oxidation of PM is usually exceeding ca. 550°C.

In the field of automotive exhaust catalysts, foam-type substrates have been proposed as alternatives to the well-established honeycomb substrates. The ceramic foams developed and manufactured in our laboratory are capable of redistributing the flow of exhaust gases, enhancing turbulence, mass transfer and species mixing, without increasing flow resistance and pressure drop to prohibitive levels. Based on the characteristics of turbulent mass and heat transfer, ceramic foam based catalysts have the potential for achieving similar pollutant conversion performances as state of art honeycomb catalysts with substantially lower precious metal requirements. In this paper we demonstrate this potential with a small Diesel powered Heavy Duty truck with a ceramic foam Diesel Oxidation Catalyst (DOC). Given the substantial differences in geometrical properties between foams and honeycombs a direct comparison with equal coating thickness, amount and precious metal amount is not feasible.

Macroscopic studies and scanning electron microscope (SEM), as well as transmission electron microscope (TEM) research were carried out to investigate the nature and properties of particulate matter (PM) deposited in three diesel particulate filters (DPFs) operating with different fuels: 100% rapeseed methyl ester (RME100), a blend of 20% RME and 80% diesel (RME20), as well as 100% diesel (RME0). The DPFs were catalytically coated with V₂O₅/TiO₂. The PM deposits were either extracted from sectioned DPFs or studied "in situ," as deposited. In the RME100-DPF, the lowest soot and highest ash depositions are found. The higher amount of ash in RME100-DPF, as well as the higher participation of the element Ca in the ash from this filter, indicates that in addition to lubricating oil, the RME fuel contributes also to ash formation. Ash is found accumulating in the plugged inlet channels only in RME100 and as a few tens of μm-thick layer on the channel walls of all three filters.