Modelling of selective catalytic reduction systems using discrete-time Linear Parameter Varying models

Abstract. Isocyanic acid (HNCO), an acidic gas found in tobacco smoke, urban environments, and biomass-burning-affected regions, has been linked to adverse health outcomes. Gasoline- and diesel-powered engines and biomass burning are known to emit HNCO and hypothesized to emit precursors such as amides that can photochemically react to produce HNCO in the atmosphere. Increasingly, diesel engines in developed countries like the United States are required to use selective catalytic reduction (SCR) systems to reduce tailpipe emissions of oxides of nitrogen. SCR chemistry is known to produce HNCO as an intermediate product, and SCR systems have been implicated as an atmospheric source of HNCO. In this work, we measure HNCO emissions from an SCR system-equipped diesel engine and, in combination with earlier data, use a three-dimensional chemical transport model (CTM) to simulate the ambient concentrations and source/pathway contributions to HNCO in an urban environment. Engine tests were conducted at three different engine loads, using two different fuels and at multiple operating points. HNCO was measured using an acetate chemical ionization mass spectrometer. The diesel engine was found to emit primary HNCO (3–90 mg kg fuel−1) but we did not find any evidence that the SCR system or other aftertreatment devices (i.e., oxidation catalyst and particle filter) produced or enhanced HNCO emissions. The CTM predictions compared well with the only available observational datasets for HNCO in urban areas but underpredicted the contribution from secondary processes. The comparison implied that diesel-powered engines were the largest source of HNCO in urban areas. The CTM also predicted that daily-averaged concentrations of HNCO reached a maximum of ∼ 110 pptv but were an order of magnitude lower than the 1 ppbv level that could be associated with physiological effects in humans. Precursor contributions from other combustion sources (gasoline and biomass burning) and wintertime conditions could enhance HNCO concentrations but need to be explored in future work.

<div class="htmlview paragraph">The tightening legislation in the on-road heavy-duty diesel area means that pollution control systems will soon be widely introduced on such engines. A number of different aftertreatment systems are currently being considered to meet the incoming legislation, including Diesel Particulate Filters (DPF), Diesel Oxidation Catalysts (DOC) and Selective Catalytic Reduction (SCR) systems.</div> <div class="htmlview paragraph">Relatively little is known about the interactions between lubricant-derived species and such aftertreatment systems. This paper describes the results of an experimental program carried out to investigate these interactions within DPF, DOC and SCR systems on a state-of-the-art 9 litre engine.</div> <div class="htmlview paragraph">The influence of lubricant composition and lube oil ash level was investigated on the different catalyst systems. In order to reduce costs and to speed up testing, test oil was dosed into the fuel. Tests without dosing lubricant into the fuel were also run.</div> <div class="htmlview paragraph">Driving distances of up to 115,000 km (based on oil consumption) were simulated in these experiments. Following such treatments, no significant change in the activity of the CR-DPF oxidation catalyst, Diesel Oxidation Catalysts or SCR catalysts were observed. In each case, the distribution of oil-derived species was studied as a function of distance down the catalyst. It was found that the Ca, Zn and P were generally present in higher concentrations at the front of the catalyst, decreasing down the unit. In contrast, sulfur is distributed evenly throughout the length of each catalyst. This is because sulfur can be deposited from the gas phase and is more mobile on the catalyst surface.</div> <div class="htmlview paragraph">Interestingly, when doping high ash oil into the fuel it was found that the back pressure of the DPF system increased rapidly. This is particularly important since it is known that used engine oil is sometimes poured back into the fuel tank for disposal. These results imply that this practice may strongly affect the efficiency of DPF devices.</div> <div class="htmlview paragraph">This rapid increase in back pressure is a consequence of doping the fuel with the high ash oil, since operation of the engine with the high ash oil without doping, or with a low ash oil with doping, resulted in only a gradual increase in back pressure. In addition, the distribution of the ash in the filter was strongly affected when doping the high ash oil into the fuel, in that the ash was uniformly distributed down the length of the filter. In the other experiments carried out here, and in measurements of field-aged filters, the ash thickness increases with distance into the filter. These observations reveal that care needs to be taken when doping oil into the fuel to accelerate the observation of oil-derived effects in DPF systems.</div>

In this paper, we investigate the cycle-based sizing optimization for diesel-engine two-cell selective catalytic reduction (SCR) systems. In modeling the SCR system, it is generally assumed that the states inside the catalyst can are homogeneous to simplify the modeling work. However, the actual states are not homogeneous, and the modeling error becomes large if the catalyst length is long. To reduce the modeling error, the states are assumed piecewise homogeneous, i.e., the catalyst is sliced into multicells and the states for each cell are assumed homogeneous. In this paper, we focus on the two-cell SCR systems where the two cells are connected in series and where the chemical reaction rates are assumed identical for the two cells. Considering the main chemical reactions and assuming that each cell is a continuous stirred tank reactor, a nonlinear system model with six states is obtained. To simultaneously reduce the tailpipe $\mbox{NO}_{x}$ and ammonia slip, it is quite interesting to study the optimal control of the urea dosing such that the $\mbox{NO}_{x}$ emissions can be maximally converted and where the ammonia slip is constrained. A dynamic programming (DP) algorithm that has a tolerable computational load is proposed. Since the whole catalyst is sliced into two cells, the next motivation is how to divide the catalyst such that the $\mbox{NO}_{x}$ conversion efficiency can be maximized while meeting the tailpipe ammonia constraint. The best sizing ratio of the SCR system in our application for the US06 test cycle is found to be 60%–40%.

Development and experimental studies of a control-oriented SCR model for a two-catalyst urea-SCR system

Large two-stroke marine diesel engines are used as the prime mover in the majority of ocean going commercial vessels and are currently facing new stringent NOx emission reductions. Selective catalytic reduction is an aftertreatment technology which is able to reduce emitted NOx at allowable levels. In most marine applications, the selective catalytic reduction system is placed on the high pressure side of the turbine, due to limitations at the exhaust gas temperature at the inlet of selective catalytic reduction system. In this article, the operation of a large two-stroke marine diesel with a selective catalytic reduction aftertreatment system, is investigated in heavy weather conditions. Operation of a vessel in heavy weather results in increased ship resistance, wave-induced ship motions, and a highly varying flow field in the propeller due to ship motions. This results in a fluctuation of propeller demanded torque and hence a fluctuation in engine load and exhaust gas temperature which may affect engine and selective catalytic reduction performance significantly. To investigate this phenomenon and taking into account the engine–propeller interaction, the entire propulsion system was modeled, namely the propulsion engine, the high pressure selective catalytic reduction system, the directly driven propeller, and the ship’s hull. A propeller model was employed to simulate the transient propeller torque demand and torque fluctuations due to ship motions. A three-dimensional six degrees of freedom panel code was used to calculate ship motions and wave added resistance in regular waves. The coupled model of the marine propulsion plant was validated against available measured data from a ship-board propulsion system on good weather conditions. The model was then used to simulate the behavior of the system during transient loading conditions in the presence of regular waves.

The purpose of this study was to perform both experimental and computational investigations on the selective catalytic reduction (SCR) system in passenger cars equipped with compression ignition (CI) engines. The study involves a comparison of results obtained for two separate SCR systems: an existing one and a newly developed system. The newly designed SCR system is intended for implementation in the spare parts market (Aftermarket) and includes the creation of a custom mixer design. This research analyzed multiple SCR systems and mixers under varying operating conditions. Various factors were considered, including the examination and evaluation of gas distribution and nitrogen oxide reduction. The multiphase computational fluid dynamics analyses were conducted using the ANSYS Fluent software. A detailed assessment was carried out for the sequential processes occurring within the system. The final version of the replacement SCR system was analyzed in relation to the original system supplied by the original equipment manufacturer (OEM). The implementation of the new mixer in the replacement SCR system led to slightly reduced NOX emissions, as validated by emission tests (NEDC) performed in a car on a chassis dynamometer within a certification unit.

Mixing and evaporation analysis of a high-pressure SCR system using a hybrid LES-RANS approach

This study proposes a data-driven optimization framework to enhance emission control performance in diesel engine selective catalytic reduction (SCR) systems under transient operating conditions. A one-dimensional SCR model was constructed in GT-Power, and simulation datasets were generated using experimentally measured inputs from the World Harmonized Transient Cycle (WHTC), with representative emission responses obtained by varying fixed ammonia-to-NOx (A/N) ratios. Building on these datasets, a hybrid prediction model combining Long Short-Term Memory (LSTM) networks and multi-head attention mechanisms was developed to accurately forecast SCR outlet NH3 leakage and NOx emissions. The model exhibited high predictive accuracy, achieving R2 values exceeding 0.977 and low RMSE across training, validation, and test sets. Based on the model predictions, a constrained dynamic multi-objective optimization strategy was implemented to adaptively adjust ammonia dosing, aiming to simultaneously minimize NH3 leakage and NOx emissions. The optimized NH3 injection profiles were validated through reapplication in the GT-Power simulation environment. Compared to the baseline fixed-ratio control strategy, the proposed approach reduced NH3 leakage and NOx emissions by 34.40% and 11.15%, respectively, as determined for the transient segment of the WHTC cycle. These results demonstrate the effectiveness of integrating physics-based simulation, deep learning prediction, and dynamic optimization for improving aftertreatment adaptability and emission compliance in real-world diesel engine applications. All reported values are based on a single simulated WHTC cycle without statistical uncertainty analysis.

The heavy-duty diesel engines have created effective method to reduce nitrogen oxide (NOx) pollutions with selective catalytic reduction (SCR) system. This study deals with problems in the urea evaporation and decomposition process, the ammonia gas distribution in SCR system and ammonia pattern at inlet of catalyst. The test system used a 12,000 cc heavy-duty diesel engine for parameter and two types of urea injectors, an L-type and an I-type. The uniformity of the ammonia was sampled at the catalyst inlet using a gas sensor, and samples of NOx from each urea injector were compared by experiments and simulations using STAR-CCM+ software. The results elucidate the saturation phenomena, vaporization phenomena, urea distribution patterns, and NOx reduction efficiency value from the two types of urea injectors. The study on the extent of NOx reduction at the catalyst inlet demonstrate the different chemical mechanisms governing the ammonia concentration distribution and the mixing process quality with NOx.

Urea-based selective catalytic reduction (SCR) system coupled with a downstream ammonia oxidation catalyst (AMOX), has become a standard NO x reduction device for Diesel engines. On-board diagnostics of tailpipe NO x and ammonia (NH3) emissions and closed-loop controls using only tailpipe NO x sensor, are critical for SCR systems to achieve high NO x conversion efficiency and low tailpipe NH3 slip. However, due to commercial NO x sensor NH3 cross-sensitivity issue, the tailpipe NH3 emissions can be misinterpreted as NO x emissions, which may lead to erratic diagnostics and unstable control systems. The purpose of this study is to develop a high-fidelity control-oriented SCR-AMOX model and a model-based estimation algorithm using extended Kalman filter (EKF) for effectively decoupling NO x and NH3 concentrations from the mixed tailpipe NO x sensor signals for an SCR-AMOX system. The proposed model and EKF-based decoupling algorithm were validated using the experimental data collected from a Diesel engine platform during both steady-state and transient driving cycles. Experimental verification results demonstrated high accuracy of the control-oriented model and proved the efficacy of the proposed decoupling algorithm in meeting the preset thresholds. The robustness of the decoupling algorithm against the uncertainty from catalyst aging was also demonstrated and verified. Such EKF-based robust decoupling algorithm can be instrumental in the diagnostics and controls of urea-based SCR systems.

Opacity and nitric oxides (NOx), as indicators of exhaust emissions of a diesel engine, are good cases of the harm it is being made to the environment and the population. AdBlue® injected in a selective catalytic reduction (SCR) system seeks to decrease NOx emissions. The goal of this work is to test the variations in both parameters in static conditions and engine operating temperature of 80°C with different injection frequencies and with different engine speeds. It is reported that a reduction of 25-30% in NOx emissions can be achieved with a frequency of injection of 30 ms. However, the opacity was demonstrated to increase to some extent.

We Energies’ Pleasant Prairie Power Plant (P4) is installing a wet flue gas desulfurization (WFGD) system on steam generating Units 1 and 2 for SO2 control and a selective catalytic reduction (SCR) system on steam generating Unit 1 for NOx control. Unit 2 already has a SCR system that was installed in 2002. This retrofit project also includes installation of a new chimney, limestone and gypsum handling equipment, and wastewater treatment equipment. This environmental improvement project, now under way at P4, is part of the Company’s plan for meeting current and pending air quality improvement commitments and environmental regulations. P4 consists of two 617 MW Riley PC subcritical boilers firing PRB coal. The two units currently share a common chimney. These units were commissioned in the early to mid 80s. When the current project is completed, P4 will have the lowest SO2 and NOx emission rates of any coal fired Wisconsin power plant. The new equipment will be capable of reducing SO2 emissions by up to 95 percent and NOx emissions by up to 90 percent. The current schedule has Unit 1’s WFGD and SCR system on line by December of 2006 and Unit 2’s WFGD system on line by May of 2007. This paper will outline the execution plan used to launch this multi-million dollar project and to maintain its schedule. The paper will also review the technical drivers that lead to the selection and performance specifications of the WFGD and an overview of the major equipment parameters including the materials of construction chosen for this project.

<div class="section abstract"><div class="htmlview paragraph">Selective Catalytic Reduction (SCR) systems are crucial for automotive emissions control, as they are essential to comply with stringent emissions regulations. Model-based SCR controls are used to minimize NOx emissions in a broad range of real-word driving scenarios, constantly adapting the urea injection to diverse load and temperature operating conditions, also accounting for different catalyst ageing status. In this framework, Neural Networks (NN) based models offer a promising alternative to reduced-order physical models or map-based controls.</div><div class="htmlview paragraph">This study introduces a hybrid modeling approach for SCR systems, leveraging the integration of machine learning techniques with detailed physics-based models. A high fidelity 1D-CFD plant model of a SCR catalyst, previously calibrated on experimental data, was used as digital twin of the real component. A standardized simulation protocol was defined to virtually characterize the SCR thermal and chemical behavior under the full range of operating conditions typically covered during the real operation of the system. The generated dataset, including hard-to-measure physical quantities such as the catalyst wall temperature and the ammonia storage, was used to train and validate the neural network models. In particular, Recurrent Neural Networks (RNN) were used to mimic catalyst wall temperature and ammonia storage temporal trends, while Feed Forward Neural Networks (FFNN) were applied to model SCR outlet temperature and species concentrations such as NOx and NH3.</div><div class="htmlview paragraph">The NN-based SCR model predictive capabilities were assessed against experimental driving cycles data. The results show that the NN model is able to accurately capture the non-linear characteristics of the system behavior, even under the highly transient conditions typical of real-driving scenarios, thus confirming the reliability of the proposed methodology.</div></div>

This paper presents a control-oriented model describing the dynamics of oxygen concentration through a Diesel engine aftertreatment system that includes a Diesel oxidation catalyst (DOC) and a Diesel particulate filter (DPF). Exhaust gas oxygen concentration is important for catalysts such as NOx conversion efficiencies of selective catalytic reduction (SCR) systems and lean NOx traps (LNT). In the presence of low-pressure loop exhaust gas recirculation (EGR), the exhaust gas oxygen concentration after-DPF also influences combustion. Due to the chemical reactions occurring inside DOC and DPF, the exhaust gas oxygen concentration considerably varies through the aftertreatment systems. Directly measuring the exhaust gas oxygen concentrations at different locations through the exhaust gas aftertreatment system is costly and unreliable. A dynamic model is thus needed in order to design model-based observers to estimate the exhaust gas oxygen concentrations at various locations. The oxygen-related reactions within a DOC and a DPF are investigated in this study. A lumped-parameter, control-oriented DOC-DPF oxygen concentration dynamic model was developed by a multi-objective optimization method and validated with experimental data obtained on a medium-duty Diesel engine equipped with full aftertreatment systems. Experimental results show that the model can well capture the oxygen dynamics across the Diesel engine aftertreatment systems.

The plant of the selective catalytic reduction (SCR) system is characterized by significant nonlinearity, time delay and temperature sensitivity. In order to control the urea injection accurately, the (nonlinear auto regressive moving average) NARMA-L2 model based control is applied to the SCR system. In this paper, a data-based technique is taken and a model of the plant is identified on the basis of input-output data. Then the identified model is used to the design of a NARMA-L2 controller. Simulation of the NARMA-L2 model based control for the SCR system is presented to demonstrate the effectiveness and superiority. The comparison results show better performance over the traditional PID control.