This paper examines the issues concerning particulate matter (PM) emissions measurement at the 3 mg/mi level proposed as the future LEV III standard. These issues are general in nature, but are exacerbated at the low levels contemplated for upcoming emissions standards. They are discussed in the context of gasoline direct injection (GDI) engines, where they can have an important impact on the continued development of this technology for improved fuel economy. GDI particulate emissions, just as engine-out diesel PM, contain a high fraction of soot. But the total PM mass is significantly lower than from diesel engines, and there can be significant variations in emissions rate and apparent PM composition between cold-start and running emissions. PM emissions levels depend on sampling method and location. As a result, there can be substantial differences in PM sampled and diluted directly at the exhaust pipe, as opposed to measurements from a dilution tunnel.
A laboratory study was performed to assess the effectiveness of LNT+SCR systems for NOx control in lean exhaust. The effects of the catalyst system length and the spatial configuration of the LNT & SCR catalysts were evaluated for their effects on the NOx conversion, NH₃ yield, N₂O yield, and HC conversion. It was found that multi-zone catalyst architectures with four or eight alternating LNT and SCR catalyst zones had equivalent gross NOx conversion, lower NH₃ and N₂O yield, and significantly higher net conversion of NOx to N₂ than an all-LNT design or a standard LNT+SCR configuration, where all of the SCR volume is placed downstream of the LNT. The lower NH₃ emissions of the two multi-zone designs relative to the standard LNT+SCR design were attributed to the improved balance of NOx and NH₃ in the SCR zones.
The regeneration process of a Diesel Particulate Filter (DPF) consists of an increase in the engine exhaust gas temperature by using post injections and/or exhaust fuel injection during a period of time in order to burn previously trapped soot. The DPF regeneration is usually performed during a real drive cycle, with continuously changing driving conditions. The quantity of post injection/exhaust fuel to use for regeneration is calculated using a combination of an open loop term based on engine speed, load and exhaust gas flow and a closed loop term based on an exhaust gas temperature target and the feedback from a number of sensors. Due to the nature of the system and the slow response of the closed loop term for correcting large deviations, the authority of the fuel calculation is strongly biased to the open loop. However, the open loop fuel calculation might not be accurate enough to provide adequate temperature tracking due to several disturbances in the system.
High concentrations of diesel fuel can accumulate in the engine oil, especially in vehicles equipped with diesel particle filters. Fuel dilution can decrease the viscosity of engine oil, reducing its film thickness. Higher concentrations of fuel are believed to accumulate in oil with biodiesel than with diesel fuel because biodiesel has a higher boiling temperature range, allowing it to persist in the sump. Numerous countries are taking actions to promote the use of biodiesel. The growing interest for biodiesel has been driven by a desire for energy independence (domestically produced), the increasing cost of petroleum-derived fuels, and an interest in reducing greenhouse gas emissions. Biodiesel can affect engine lubrication (through fuel dilution), as its physical and chemical properties are significantly different from those of petrodiesel. Many risks associated with excessive biodiesel dilution have been identified, yet its actual impact has not been well quantified.
Under the current emissions legislation, most of the diesel-powered vehicles have to use Diesel Particulate Filters (DPF) to remove soot particles from the exhaust gas and the accumulated soot particles have to be removed in regular intervals. To initialize the exhaust gas temperature for soot regeneration, diesel fuel is either injected into the combustion chamber in late engine cycle (e.g. post injection) or vaporized and then discharged into the exhaust gas via a dosing device (e.g. fuel vaporizer). Both approaches though require the exothermic catalyst to convert the fuel into thermal energy. For practical reasons, this paper is concentrated on describing how CFD could be used to model the fuel distribution in an aftertreatment system equipped with fuel vaporizer and the exothermic reactions in the catalysts.
A laboratory study was performed to assess the potential capability of TWC+LNT/SCR systems to satisfy the Tier 2, Bin 2 emission standards for lean-burn gasoline applications. It was assumed that the exhaust system would need a close-coupled (CC) TWC, an underbody (U/B) TWC, and a third U/B LNT/SCR converter to satisfy the emission standards on the FTP and US06 tests while allowing lean operation for improved fuel economy during select driving conditions. Target levels for HC, CO, and NOx during lean/rich cycling were established. Sizing studies were performed to determine the minimum LNT/SCR volume needed to satisfy the NOx target. The ability of the TWC to oxidize the HC during rich operation through steam reforming was crucial for satisfying the HC target.
A laboratory study was performed to assess the potential capability of passive TWC+SCR systems to satisfy the Tier 2, Bin 2 emission standards for lean-burn gasoline applications. In this system, the TWC generates the NH3 for the SCR catalyst from the feedgas NOx during rich operation. Therefore, this approach benefits from high feedgas NOx during rich operation to generate high levels of NH3 quickly and low feedgas NOx during lean operation for a low rate of NH3 consumption. It was assumed that the exhaust system needed to include a close-coupled (CC) TWC, an underbody (U/B) TWC, and an U/B SCR converter to satisfy the emission standards during the FTP and US06 tests while allowing lean operation for improved fuel economy during select driving conditions. Target levels for HC, CO, and NOx during lean/rich cycling were established. With a 30 s lean/10 s rich cycle and 200 ppm NO lean, 1500 ppm NO rich and the equivalent of 3.3 L of SCR volume were required to satisfy the NOx target.
Thermocouples are commonly used to measure exhaust gas temperature during automotive engineering experiments. In most cases, the raw measurements are used directly as an absolute indication of the actual exhaust gas temperature. However, in reality, the signal from a TC is only an indication of its own tip temperature. The TC indicated tip temperature can deviate significantly from the actual gas temperature due to factors such as thermal capacitance of the tip itself, and heat transfer to the exhaust pipe wall through conduction and radiation. A model has been developed that calculates the effects of these factors to provide an estimate of the actual exhaust gas temperature. Experiments were performed to validate the model under both transient and steady state engine dynamometer conditions utilizing three popular sizes of TCs. Good correlation among predictions for various TC sizes confirms the model's accuracy.
One-dimensional simulation tools are used extensively in the automotive industry to improve and optimize engine design for WOT performance. They are useful in target setting and in assessing the effects of certain design changes (e.g. intake manifold, valve timing, exhaust manifold, etc.). Generally the inputs to these models are “nominal” values or curves from a particular set of data and, therefore, do not take into account design or assembly variations. Often times, performance expectations are not met due to these “real world” effects and may result in significant re-design and testing efforts. The purpose of this paper is to assess the impact of typical model input variation on engine performance and to instill greater confidence in the use of these models in forecasting performance. The approach taken is to collect, analyze, and categorize actual build measurements from a 4.6L 4V Ford engine that are considered important inputs for a one-dimensional modeling.
To meet stringent 2010 NOx emissions, many manufacturers are expected to deploy urea selective catalytic reduction systems. Indications from ARB are that a threshold monitor must be developed to monitor their performance. The most capable monitoring technology at this time relies on NOx sensors. This paper assesses the capability of the NOx sensor as an SCR monitoring device. To this end, the NOx sensor must be able to distinguish between a marginal and a threshold catalyst with enough separation to allow for variability. We present the noise factors associated with the NOx conversion of the SCR system, and analyze what NOx sensor accuracy we need to preserve separation in the face of those noise factors. It is shown that a 1.75 threshold monitor is not feasible with current NOx sensor technology. We analyze the benefit of a partial volume monitor, and show there is no advantage unless the slope error of the NOx sensor is drastically reduced from current levels.
The effects of PGM zoning and washcoat staging have been investigated as a means to lower the cost and simultaneously improve the performance of a lean NOx trap system. It is shown that reverse PGM zoning can be used to reduce the cost of the LNT while essentially maintaining the NOx performance of a similarly-sized trap with a uniformly high PGM loading. In addition, the effective temperature window of the trap can be expanded by staging different NOx trap formulations that are optimized for different temperature ranges. Alternatively, LNT washcoat staging can be used to improve the hydrocarbon conversion of the trap while maintaining good NOx performance. Laboratory data and vehicle data are presented for several NOx trap system combinations that demonstrate the improved performance that can be obtained from a combination of reverse PGM zoning and washcoat staging.
This paper describes the use of an interactive NVH simulator in simulating and designing the sound character of a vehicle with a multi-mode engine and active exhaust valve. When designing a vehicle for sound quality, it is not sufficient to merely record some discreet operating conditions and modify these in a traditional sound quality program. The ability to simulate the sound quality of the vehicle over the full operating envelope is a necessity. Additionally, the ability to break down the sound contributions from intake, exhaust and other key contributors to the driver's ear, and manipulate these independently is also essential. In the case described here, an additional factor makes it mandatory that an accurate vehicle sound simulation is performed. The state of the engine and exhaust contribution, and thus the sound of the vehicle, change based on several parameters - vehicle speed, load demand and gear.
Engine and flow reactor experiments were conducted to determine the impact of biodiesel relative to ultra-low-sulfur diesel (ULSD) on inhibition of the selective catalytic reduction (SCR) reaction over an Fe-zeolite catalyst. Fe-zeolite SCR catalysts have the ability to adsorb and store unburned hydrocarbons (HC) at temperatures below 300°C. These stored HCs inhibit or block NOx-ammonia reaction sites at low temperatures. Although biodiesel is not a hydrocarbon, similar effects are anticipated for unburned biodiesel and its organic combustion products. Flow reactor experiments indicate that in the absence of exposure to HC or B100, NOx conversion begins at between 100° and 200°C. When exposure to unburned fuel occurs at higher temperatures (250°-400°C), the catalyst is able to adsorb a greater mass of biodiesel than of ULSD. Experiments show that when the catalyst is masked with ULSD, NOx conversion is inhibited until it is heated to 400°C.
This paper discusses the poisoning of a selective catalytic reduction (SCR) catalyst by trace levels of platinum originating from an upstream diesel oxidation catalyst (DOC). A diesel aftertreatment system consisting of a DOC, urea based SCR Catalyst and a DPF was aged and evaluated on a 6.4 liter diesel engine dynamometer. The SCR catalyst system consisted of an Fe-zeolite catalyst followed by a Cu-zeolite catalyst. After approximately 400 hours of engine operation at varied exhaust flow rates and temperatures, deactivation of the SCR catalyst was observed. A subsequent detailed investigation revealed that the Cu catalyst was not deactivated and the front half of the Fe-based catalyst showed severe deactivation. The deactivated portion of the catalyst showed high activity of NH3 conversion to NOx and N2O formation. The cause of the deactivation was identified to be the presence of trace Pt contamination.
This paper analyzes the potential benefit of a model based DPF leakage monitor over a conventional DPF leakage monitor that checks pressure drop after a complete regeneration. We analyze the most important noise factors involved in both approaches and demonstrate that the model based leakage monitor does not improve on the conventional leakage monitor in accuracy. It does improve on completion frequency, but at the expense of a great modeling effort.
The Ford GT is the modern re-creation of the 60's era supercar. The powertrain sound quality of the vehicle must enhance its powerful nature, meet regulatory requirements, and maintain a targeted level of refinement. The Ford GT acoustic engineering team used time domain sound decomposition and sound synthesis techniques to determine the sub-system source sounds from surrogate vehicles. The donor source sounds (e.g. exhaust system) are recombined to produce the customer perceived vehicle listening experience from these sub-systems. Target sounds are developed by modifying sub-systems by level, frequency dominance, and order balance. Proposed target sounds are verified by a jury and the results are used for early target agreement and cascading to component targets. This exercise allows development of a customer focused powertrain target sound based on realistic hardware assumptions before any prototypes are available.
This paper presents a study of urea SCR catalyst aging characteristics and implementation into an analytical model that complements laboratory based experiments for a dynamometer-aged SCR brick. The model calibration is based on measured data taken from a 120k-mile simulated dynamometer-aged base metal/zeolite SCR. Dynamometer aging led to non-uniform axial deterioration with more severe deactivation toward the front of the SCR brick compared to the rear. Data from a 120k-mile simulated hydrothermally oven-aged SCR (uniform axial aging) is used to establish baseline aged NOx performance and NH3 adsorption/desorption behavior. An axial deterioration factor is applied to the baseline model to account for differences between oven and dynamometer aging. The model is exercised using engine out vehicle data to examine how different aging processes (oven vs. dynamometer) affect overall NOx performance during the EPA FTP (Environmental Protection Agency Federal Test Procedure).
It is likely that use of urea-based selective catalytic reduction (SCR) will be needed to meet U.S. Tier 2 diesel emission standards for oxides of nitrogen (NOx). The ideal ratio of ammonia (NH3) molecules to NOx molecules (known as alpha) is 1:1 based on urea consumption and having NH3 available for reaction of all of the exhaust NOx. However, SCR efficiency can be less than 100% at low temperatures in general, and at higher temperatures with high exhaust SCR catalyst space velocities. At the low temperatures where NOx conversion efficiency is low, it may be advantageous to reduce the alpha ratio to values less than one (less NH3 than is needed to convert 100% of the NOx emissions) to avoid NH3 slip. At higher space velocities and high temperatures, the NOx conversion efficiency may be higher with alpha ratios greater than 1. There is however concern that the additional NH3 will be slipped under these conditions.
The use of urea-based selective catalytic reduction (SCR) is a promising method for achieving U.S. Tier 2 diesel emission standards for NOx. To meet the Tier 2 standards for Particulate Matter (PM), a catalyzed diesel particulate filter (CDPF) will likely be present and any ammonia (NH3) that is not consumed over an SCR catalyst would pass over the CDPF to make nitrous oxide (N2O) emissions and/or oxides of nitrogen (NOx), or exit the exhaust system as NH3. N2O is undesirable due to its high greenhouse gas potential, while NOx production from the slipped NH3 would reduce overall system NOx conversion efficiency. This paper reviews certain conditions where NH3 slip past an SCR system may be a concern, looks at what would happen to this slipped NH3 over a CDPF, and evaluates the performance of various supplier NH3 slip catalysts under varied space velocities, temperatures and concentrations of NH3 and NOx.