Abstract : This research focuses on modeling and control of PM and NOx in diesel engine exhaust using an SCR catalyst on a Diesel Particulate Filter (SCR-F). A 2D SCR-F model was developed that is capable of predicting internal states: 2D temperature, PM and NH3storage distributions and filtration efficiency, pressure drop, PM mass retained in the PM cake and substrate wall and outlet NO, NO2 and NH3concentrations. The SCR-F model was used to simulate a DOC + SCR-F + DOC + SCR ultra-low NOx system that can achieve > 99.5% NOx conversion efficiency. 99.5% NOx conversion efficiency. The model was calibrated with experimental data from a Johnson Matthey SCRF® with a Cummins 2013 ISB engine. The impact of SCR reactions on passive PM oxidation rate and PM loading on SCR reactions was determined. A comparison of the experimental and model data for different ammonia to NOx ratios, PM loading, and passive oxidation conditions is presented. A 2D SCR-F state estimator was developed by combining a simplified version of the 2D SCR-F model with pressure drop, outlet thermocouple and NOx sensormeasurements using an Extended Kalman Filter. The temperature, PM mass retained and NH3coverage fraction states were predicted which can be used to develop fuel dosing and urea dosing strategies for the SCR-F. A 2D SCR-F + 1D SCR system model was used to simulate the experimental data collected on a SCR-F + SCR system from a Cummins 2013 ISB engine. The NO2/NOxratioat the SCR-F and SCR inlet was found to be limiting factor for NOx conversion efficiency of this system. An ultra-low NOxsystem was developed with a DOC downstream of the SCR-F that maintains an optimum NO2/NOxratio of 0.5 at the downstream SCR inlet by using 2 urea injectors. This system was simulated with a combination of 1D DOC, 2D SCR-F, and 1D SCR models and it was found to be capable of > 99.5% NO 99.5% NOxconversion efficiency, a 90% increase in PM oxidation rate compared to an SCR-F + SCR system with 1 injector for typical engine operating conditions.

Urea-SCR Technology for deNOx After Treatment of Diesel Exhausts presents a complete overview of the selective catalytic reduction of NOx by ammonia/urea. The book starts with an illustration of the technology in the framework of the current context (legislation, market, system configurations), covers the fundamental aspects of the SCR process (catalysts, chemistry, mechanism, kinetics) and analyzes its application to useful topics such as modeling of full scale monolith catalysts, control aspects, ammonia injections systems and integration with other devices for combined removal of pollutants.

Abstract : Reduction of emissions and improving the fuel consumption are two prime research areas in Diesel engine development. The present after-treatment systems being used for emissions control include diesel oxidation catalyst (DOC) for NO, HC and CO oxidation along with catalyzed particulate filters for PM (particulate matter) and selective catalytic reduction (SCR) for controlling NOx emissions. Recently an after-treatment system called SCR catalyst on a DPF capable of simultaneously reducing both NOx and PM emissions has been developed in order to reduce the overall size of the after-treatment system. The goal of this proposed research is to create a state estimator that is capable of estimating the internal states of temperature distribution, PM distribution, NH3 storage faction as well as pressure drop across the filter and outlet concentration of NO, NO2 and NH3 for different operating conditions. This would help in achieving an optimal urea dosing strategy during NOx reduction as well as an optimum fuel dosing strategy during active regeneration for the SCR catalyst on a DPF. The motivation for this research comes from the desire to quantify the interaction of SCR reactions and PM oxidation in the SCR catalyst on a DPF and to use the mathematical model created in the process to develop a state estimator that can provide optimal control and onboard diagnostics of combined SCR catalyst on a DPF devices. In the initial phase of the research a high-fidelity SCR-F model is being developed in MATLAB/Simulink which is capable of predicting the filtration efficiency, temperature distribution, PM distribution, pressure drop across the filter and outlet concentrations of NO, NO2 and NH3. This model will be calibrated using experimental data collected on a Cummins 2013 ISB SCRF®. After the validation of the SCR-F model, the high-fidelity SCR-F model developed will be used with an existing 1D SCR model to perform NOx reduction studies on a system consisting of SCRF® + SCR using experimental data. This step will be followed by development of a reduced order SCR-F model using a coarser mesh (e.g. 5x5 vs 10x10) and simplified governing equations which will also be used as the mathematical model for the state estimator. SCR-F state estimator will be developed to accurately predict the internal states of NH3 coverage fraction, temperature distribution, PM distribution and pressure drop across the SCR catalyst on the DPF. The estimator will be validated using experimental data.

Abstract : The heavy-duty diesel engines use a Diesel Oxidation Catalyst (DOC), a Catalyzed Particulate Filter (CPF), a Selective Catalytic Reduction (SCR) with urea injection and a Ammonia Oxidation Catalyst (AMOX), to meet the US EPA 2010/2013 particulate matter (PM) and NOx emission standards. However, it is not possible to achieve the 2015 California low NOx standards with this arrangement. Hence, there is a need to improve the existing aftertreatment system. This can be achieved by coating the SCR catalyst on a diesel particulate filter (DPF), thus combining the PM filtration and NOx reduction functionality into a single device. This reduces the overall volume/weight of the system and provides opportunity for packaging flexibility and improved thermal management along with the possibility of higher NOx reduction with a downstream SCR system. The SCR catalyst on a DPF used in this study is known as a SCRF℗ʼ which was supplied by Johnson Matthey and Corning. Previous research on the CPF and SCRF℗ʼ at MTU highlighted that the reactivity of PM retained in the CPF and SCRF℗ʼ is higher during loading conditions compared to passive oxidation conditions i.e. when the flow rate of PM entering the CPF or SCRF℗ʼ is higher in loading conditions compared to the low flow rate and higher PM reaction rate during passive oxidation conditions. A 2013 Cummins ISB engine with a DOC-SCRF℗ʼ arrangement was used to perform twelve tests (eight tests without urea injection and four tests with urea injection) in order to determine the NO2 assisted passive oxidation performance of the SCRF℗ʼ under loading conditions with and without urea injection. The primary focus of this study was to carry out Loading Tests with and without Urea injection and measure species concentrations, PM mass retained, exhaust flowrates, substrate temperature distributions, pressure drop across the filter, and to determine the kinetics of NO2 assisted PM oxidation under loading conditions and compare it with kinetics under passive oxidation conditions. The NO2 assisted passive oxidation performance of the SCRF℗ʼ was experimentally studied by running the engine at 2400 RPM and four different loads at nominal and reduced rail pressure for 5.5 hours in two stages of loading. These conditions were intended to span the SCRF℗ʼ inlet temperatures in the range of 264-364oC and inlet NO2 concentrations in the range of 52-120 ppm. Four conditions out of these eight conditions were repeated with the injection of urea in the form of diesel exhaust fluid at a target ammonia to NOx ratio of 1.0 to investigate both the NOx reduction performance, as well as the effect of urea on the NO2 assisted passive oxidation performance. From the conclusions of the study based on the experimental data, it was found that the cumulative percentage of PM oxidized in the SCRF℗ʼ increases with the increase in engine load due to higher SCRF℗ʼ temperatures and NO2 concentrations. On average, the reactions rates with urea injection during loading conditions in the SCRF℗ʼ are 25% lower compared to the reaction rates without urea injection. The reactivity of PM under loading conditions with and without urea injection is higher compared to the reactivity of PM under passive oxidation with and without urea injection. For a lumped PM oxidation model, a higher pre-exponential for NO2 assisted oxidation is needed for loading as compared to passive oxidation conditions. It was not possible to determine the kinetics of NO2 assisted oxidation of PM under loading conditions from the experimental data using a standard Arrhenius model which lead to the development of a different model for PM oxidation. A PM oxidation model was developed based on the shrinking core model which keeps the identity of the incoming PM masses in the SCRF℗ʼ as compared to SCR-F model being developed at MTU which is lumped model for PM oxidation. The PM oxidation model was calibrated to simulate PM oxidation in the SCRF℗ʼ with a single set of kinetics under wide range of conditions including loading and passive oxidation conditions. The reaction rate results from the PM oxidation model were then applied to the SCR-F model to simulate the pressure drop across SCRF℗ʼ and the PM retained in the SCRF℗ʼ for the loading conditions used in this study. The SCR-F model was calibrated using experimental data from Loading Tests w/o Urea to simulate the PM retained within ℗ł2 g and pressure drop across SCRF℗ʼ within ℗ł0.5 kPa of the experimental data at the end of the test. The calibrated SCR-F model was also used to estimate the cake, wall and channel pressure drop and the PM retained in the cake and wall for the Loading Tests w/o Urea to check the integrity of experimental data and the consistency of the model. The NO2 assisted kinetics for PM oxidation in the SCRF℗ʼ without urea injection using the SCR-F model resulted in an activation energy of 96 kJ/gmol and pre-exponential factor of 2.6 m/K-s for the cake and 1.8 m/K-s for the wall. An analysis of the results from the SCR-F model suggests that for all the conditions, 84-92% of the total PM retained was in the PM cake layer and the oxidation in the PM cake layer accounted for 72-84% of the total PM mass oxidized during loading.

The research carried out on this project developed experimentally validated Diesel Oxidation Catalyst (DOC), Diesel Particulate Filter (DPF), and Selective Catalytic Reduction (SCR) high-fidelity models that served as the basis for the reduced order models used for internal state estimation. The high-fidelity and reduced order/estimator codes were evaluated by the industrial partners with feedback to MTU that improved the codes. Ammonia, particulate matter (PM) mass retained, PM concentration, and NOX sensors were evaluated and used in conjunction with the estimator codes. The data collected from PM experiments were used to develop the PM kinetics using the high-fidelity DPF code for both NO2 assisted oxidation and thermal oxidation for Ultra Low Sulfur Fuel (ULSF), and B10 and B20 biodiesel fuels. Nine SAE papers were presented and this technology transfer process should provide the basis for industry to improve the OBD and control of urea injection and fuel injection for active regeneration of the PM in the DPF using the computational techniques developed. This knowledge will provide industry the ability to reduce the emissions and fuel consumption from vehicles in the field. Four MS and three PhD Mechanical Engineering students were supported on this project and their thesis research provided them with expertise in experimental, modeling, and controls in aftertreatment systems.

Abstract: A systematic nonlinear control methodology for urea-SCR systems applicable for light-to-heavy-duty Diesel engine platforms in a variety of on-road, off-road, and marine applications is developed and experimentally validated in this dissertation. Urea selective catalytic reduction (urea-SCR) systems have been proved of being able to reduce more than 90% of Diesel engine-out NOx emissions and have been favored by the automotive industry in recent years. Urea-SCR systems utilize ammonia, converted from 32.5% aqueous urea solution (AdBlue) injected at upstream of the SCR catalyst, as the reductant for NOx reductions. Because ammonia is considered a hazardous material, urea injection should be systematically controlled to avoid undesired tailpipe ammonia slip while achieving a sufficient level of SCR NOx reduction. The novelty of the control methodology is to regulate the ammonia storage distribution along the axial direction of a SCR catalyst to a staircase profile and thus to simultaneously realize high NOx reduction efficiency and low ammonia emissions. To achieve this control objective, several relevant subjects are studied, including: 1) aftertreatment system control-oriented modeling, 2) online NOx sensor ammonia cross-sensitivity correction, 3) SCR catalyst ammonia coverage ratio estimation, as well as 4) adaptive urea dosing controller design. A unique SCR system which consists of a urea injector and two SCR catalysts connected in-series with several NOx and NH3 sensors is used for the study of the proposed urea-SCR control methodology. Such a SCR system is integrated with a state-of-the-art Diesel engine and aftertreatment system (DOC-DPF). The US06 test cycle experimental results show the proposed control methodology, in comparison to a conventional control strategy, is capable of improving the SCR NOx reduction by 63% and reducing the tailpipe ammonia slip amount by 74%. The contributions of this research to the art include: 1) A novel, efficient, and generalizable urea-SCR dosing control methodology; 2) Diesel engine-DOC-DPF NO/NO2 ratio control-oriented models and observer-based estimations; 3) SCR catalyst ammonia coverage ratio estimation methods; 4) An online correction approach for NOx sensor ammonia cross-sensitivity elimination; and 5) An improved SCR control-oriented model.