Printer Friendly

The Effect of N[O.sub.2]/N[O.sub.x] Ratio on the Performance of a SCR Downstream of a SCR Catalyst on a DPF.

Introduction

A combination of aftertreatment devices consisting of a diesel oxidation catalyst (DOC), catalyzed particulate filter (CPF), and selective catalytic reduction (SCR) is being used to meet the present N[O.sub.x] standard of 0.2 g/bhp-hr. for on-highway vehicles. Reduction of N[O.sub.x] emissions from diesel engine exhaust to meet the lower N[O.sub.x] standards being proposed by the EPA and CARB has been explored in this work using a combination of a SCR-F with a flow through downstream SCR. The N[H.sub.3] injected at the inlet of the SCR-F was used for N[O.sub.x] reduction in both the SCR-F and SCR. In order to collect the experimental data for this system, a Johnson Matthey SCRF[R] was used along with a production SCR on a Cummins 2013 6.7L ISB engine. The N[H.sub.3] slip from the SCRF[R] was the inlet N[H.sub.3] for the SCR.

A modeling study using a combination of a 2D SCR-F model [1] and a 1D SCR model [2] was performed. The SCR kinetics of the 1D SCR model and 2D SCR-F model were obtained from the experimental data described in References [3, 4]. The major objectives of the article are as follows:

1. Simulate the SCR-F + SCR system N[O.sub.x] conversion and PM oxidation performance.

2. Determine the N[O.sub.x] conversion performance of the SCR at different temperature, flow rate, and inlet N[O.sub.2]/N[O.sub.x] ratio conditions.

3. Quantify the impact of the upstream SCR-F on the N[O.sub.x] conversion performance by analyzing the contribution of each of the three SCR reactions in the SCR.

4. Determine the reduction in PM oxidation rate in the SCR-F for the given inlet conditions in the SCR-F + SCR system.

The article is divided into six sections. The literature review section presents the literature on SCR-F modeling, experimental work, and development of ultra-low N[O.sub.x] after-treatment systems. The SCR-F and SCR Models section describes the 1D SCR and 2D SCR-F models followed by 2D SCR-F + 1D SCR model development. The experimental data overview section presents the baseline SCR and SCR-F + SCR data used for the simulations.

The calibration section of models deals with the calibration process used for the 1D SCR model followed by the results section that presents N[O.sub.x] reduction performance of the SCR and SCR-F + SCR. An analysis of the factors that affect N[O.sub.x] conversion and N[H.sub.3] slip such as space velocity of the exhaust gas, reaction rate, and local N[O.sub.2]/N[O.sub.x] ratio is presented. The summary and conclusion section describes the factors that limit the system performance in terms of N[O.sub.x] conversion efficiency.

Literature Review

The SCR-F modeling and experimental literature primarily focuses on the impact of the SCR reactions on PM oxidation rate and the impact of PM loading on N[O.sub.x] reduction efficiency. The N[O.sub.x] reduction efficiency under PM loading conditions is an important factor that impacts the performance of the SCR-F + SCR system that was studied. A brief literature review of major SCR-F modeling and experimental research is described in this section followed by a description of studies directed at developing aftertreatment systems that meet the 0.02 g/bhp-hr. N[O.sub.x] standard. This literature review is an extension of the earlier research at MTU [1, 5].

SCR-F Modeling

The interaction of SCR reactions and PM oxidation rate takes place internally in the PM cake and substrate wall layers [1]. This interaction takes place by reaction-diffusion mechanism which cannot currently be quantitatively measured by any device. This motivates the development of numerical models that can simulate the change in chemical species due to the SCR reactions and PM oxidation across the PM cake and substrate wall layers.

An inhibition in the SCR reaction rate due to PM loading is primarily caused by a mass transfer limitation due to the PM deposited in the substrate wall, change in local N[O.sub.2]/N[O.sub.x] ratio across the PM cake and substrate wall, and competition for N[O.sub.2] between PM oxidation and the SCR reactions. Park et al. [6, 7] developed a 1D SCR-F model that can simulate the change in local N[O.sub.2]/N[O.sub.x] ratio across the PM cake and inhibition of the SCR reactions by mass transfer limitation caused by PM loading using the effectiveness factor concept. Colombo et al. [8] developed a model that also found a change in the SCR reaction rate due to local N[O.sub.2]/N[O.sub.x] ratio change in the PM cake. These models have also taken into account the impact of the competition for N[O.sub.2] between the PM oxidation and the SCR reactions. The model developed by Yang et al. [9] found that the PM deposited in the substrate wall leads to a mass transfer limitation of chemical species from the gas stream to the catalyst surface.

Modeling the number of catalytic sites to simulate the maximum storage capacity is another important aspect of the SCR-F modeling along with the deterioration of the catalyst from hydrothermal aging and sulfur poisoning. Lopez et al. [10] developed a vanadium-based SCR-F model that models the number of active sites proportional to the washcoat loading on the substrate wall. Dosda et al. [11] developed a Cu-Ze model that studied the deterioration of catalyst performance due to thermal aging. They found that CuO accumulation in the catalyst decreased the number of active sites leading to lower N[H.sub.3] storage and reduced N[O.sub.x] reduction performance. Tan et al. [12] developed a model that showed that for a fresh SCR-F, a 30% reduction in N[H.sub.3] storage is observed under PM loading while this effect was negligible for an aged SCR-F.

Tronconi et al. [13] developed a two-site storage model based on Axisuite software that studied the reduction of PM oxidation rate due to forward diffusion of N[O.sub.2] from the PM cake layer to the substrate wall during the SCR reactions. Also, pressure drop and filtration characteristics of the SCR-F were modeled in this work. Watling et al. [14] developed a 1D SCR-F model with reactor kinetics that found negligible change in N[H.sub.3] storage during PM loading conditions. A 5[degrees]C increase in the exhaust gas temperature was found in this research during N[O.sub.x] reduction.

SCR-F Experimental Studies

SCR-F experimental studies in the literature focus on determining the impact of SCR reactions on PM oxidation rate, change in N[O.sub.x] reduction performance in the presence of PM, and impact of N[O.sub.2]/N[O.sub.x] ratio on N[O.sub.x] reduction performance as a function of inlet exhaust gas temperature and ANR. Studies on hydrothermal aging, sulfur poisoning and low temperature nitrate formation have also been conducted.

N[O.sub.x] reduction performance of the SCR-F was studied in the work by Naseri [15] and Cavataio et al. [16] where a catalyzed soot filter (CSF) + SCR system was compared against a SCR-F system. The SCR-F was found to have higher N[O.sub.x] conversion performance compared to the CSF + SCR system in both transient cycle and steady-state conditions. Tang et al. [17] conducted similar studies related to N[O.sub.x] reduction performance during transient and steady-state conditions in the presence of >395 ppm sulfur fuel. The presence of sulfur significantly reduced the N[O.sub.x] reduction performance. The SCR-F performance was recovered by desulification at temperatures greater than 500[degrees]C. Lee et al. [18] conducted US06-based transient experiments on a SCR-F and found that increased mileage reduced N[O.sub.x] conversion due to catalyst aging which led to reduced N[H.sub.3] storage performance. Mihai et al. [19, 20] studied the reduction in N[O.sub.x] conversion with PM loading for a hydrothermally aged Cu-Ze SCR-F. An increase in ammonia storage with PM loading from 424 to 494 gmol was observed in the SCR-F. Also, formation of two types of nitrates at temperatures less than 250[degrees]C was observed.

Lasitha et al. [21] conducted a comparative study between a CSF and a SCR-F in terms of PM oxidation rate during active regeneration. For temperatures greater than 270[degrees]C, the CSF was found to have higher passive PM oxidation rate compared to the SCR-F. Also, during active regeneration, the CSF had significantly higher PM oxidation rate due to back diffusion of N[O.sub.2] from the substrate wall to the PM cake compared to the SCR-F.

Ultra-Low NOx Aftertreatment Systems

The modeling and experimental studies on SCR-Fs indicate lower N[O.sub.x] conversion across the SCR-F due to low inlet exhaust gas temperature (<250[degrees]C), increased PM loading, thermal aging, sulfur poisoning, and unfavorable N[O.sub.2]/N[O.sub.x] ratio < 0.5 during engine operation. In order to ensure a N[O.sub.x] reduction efficiency of greater than 99.0% required to meet the 0.02 g/bhp-hr. standard, a combination of SCR-F with a SCR is required to mitigate the impact of reduced N[O.sub.x] conversion of the SCR-F. Experimental studies designed to meet the cold start and hot cycle N[O.sub.x] standards have been reported in the literature on different combinations of SCR-F, SCR, and passive N[O.sub.x] adsorber (PNA).

Strots et al. [22] developed a system model of a SCR-F with various catalysts and a urea dosing injector to determine the interaction of the SCR-F with the SCR in terms of N[O.sub.x] reduction performance. A 1D SCR-F model was used along with the DOC and SCR models. A World Harmonized Transient Cycle (WHTC) cycle based on a 6-cylinder 255 kW Euro 5 engine simulation was used for the work. Two designs of the DOC + DPF + SCR + ammonia oxidation catalyst (AMOX) were compared to a DOC + SCR-F + SCR + AMOX system. A faster light-off of the SCR-F compared to the SCR in the DPF + SCR system was observed during cold startup due to the lower system thermal inertia caused by the upstream DPF in the DPF + SCR system. A higher operating temperature >8[degrees]C compared to the SCR during the hot portion of the cycle was observed for the SCR-F. The importance of N[O.sub.2] concentration profile caused by reaction-diffusion interaction with the fast SCR reaction in the substrate wall was identified for future work.

Sharp et al. [23, 24, 25] studied different combinations of aftertreatment devices that can achieve the 0.02 g/bhp-hr. N[O.sub.x] emissions. It was determined that in order to achieve this level for a cycle consisting of l/7th cold start and 6/7th hot start, a composite of 99.4% N[O.sub.x] reduction efficiency is required. A final configuration consisting of PNA + mini burner (MB) + SCR-F + SCR + ammonia slip catalyst (ASC) was identified as a plausible system that can meet this standard. Significant cold start Federal Test Procedure (FTP) emissions reduction is required to achieve the 0.02 g/bhp-hr. level. It was concluded that a combination of additional external heat, reduction of the thermal mass of the system, and positioning of catalyst is required to achieve the objective of achieving Ultra-Low N[O.sub.x] emissions.

Georgiadis et al. [26] designed a system that can significantly reduce the nonuniformity of the N[H.sub.3] coverage fraction in the SCR-F leading to lower N[H.sub.3] slip. A control system that can reduce N[H.sub.3] slip by maximizing N[H.sub.3] utilization in the SCR-F during real-world operation was developed in order to eliminate the need for an AMOX downstream of the SCR-F.

In the system-level studies in the literature consisting of a SCR-F, the role of external heating and lower thermal mass along with placement of catalyst were explored. The impact of local N[O.sub.2]/N[O.sub.x] ratio and N[H.sub.3] adsorption rate in the SCR-F and the SCR as a function of temperature and flow rate of the exhaust needs to be studied. The contribution of each of the SCR reactions at different temperatures and flow rate conditions in both the SCR-F and the SCR is also important in determining the system performance. These aspects have been studied in this work while taking into account the impact of PM loading on N[O.sub.x] reduction performance and the change in the N[O.sub.2]/N[O.sub.x] ratio across the SCR-F during PM loading.

SCR-F and SCR Models

A system consisting of a 2D SCR-F model and 1D SCR model has been used for this study. The 2D SCR-F model described in Reference [1] is capable of simulating the exhaust gas outlet temperature, PM oxidation, SCR reactions, PM mass retained, and pressure drop characteristics of the SCR-F. The outlet concentrations of NO, N[O.sub.2], and N[H.sub.3] from the SCR-F model were used as the input to the 1D SCR model described in Reference [2]. Both the models have been calibrated with experimental data described in References [3, 4]. The SCR-F and the SCR model were used to simulate the N[O.sub.x] conversion and N[H.sub.3] slip in the SCR-F + SCR system.

SCR-F Model

The 2D SCR-F model is capable of simulating the 2D temperature, PM mass, and N[H.sub.3] coverage fraction distribution. The SCR and PM oxidation kinetics were calibrated based on engine data collected on a Johnson Matthey SCRF[R] with the exhaust from a Cummins 6.7L ISB engine. A two-site model was used for N[H.sub.3] storage. The first site participates in both N[H.sub.3] storage and SCR reactions. The second site only participates in N[H.sub.3] storage.

A re action-diffusion scheme was employed to simulate the fast and slow SCR reactions, N[H.sub.3] oxidation, and the PM oxidation reactions. The change in PM oxidation rate due to the SCR reactions was accounted for by using forward diffusion of N[O.sub.2] from the PM cake layer to the substrate wall leading to an approximate 70% reduction in N[O.sub.2]-assisted PM oxidation reaction rate. The inhibition of SCR reactions due to the mass transfer limitation by PM deposited in the wall was simulated using the effectiveness concept [6]. The list of reactions in the 2D SCR-F model are listed in Table 1. The (*) in Table 1 represents the adsorbed N[H.sub.3] on the catalyst site and was added as a footnote to Table 1. The governing equations used for simulating the chemical species in the SCR-F model are described in Equations 1 to 5.

[mathematical expression not reproducible] Eq. (1)

[mathematical expression not reproducible] Eq. (2)

[mathematical expression not reproducible] Eq. (3)

[mathematical expression not reproducible] Eq. (4)

[mathematical expression not reproducible] Eq. (5)

Equations 1 and 3 are used for computing the mass transfer of chemical species by convection and diffusion mechanisms in the inlet and outlet channels of the SCR-F. Equation 2 is used to simulate the change in chemical species in the PM cake and substrate wall layers by convection, diffusion, and the SCR and the PM oxidation reactions. Equations 4 and 5 are used to track the change in coverage fraction of the two N[H.sub.3] storage sites.

The SCR-F model uses two types of catalytic active sites for N[H.sub.3] storage known as sites 1 and 2. Site 1 is responsible for both N[H.sub.3] storage and N[O.sub.x] reduction by SCR reactions and N[H.sub.3] oxidation reaction. Site 2 is used for only N[H.sub.3] storage which simulates the nonzero N[H.sub.3] slip during transition to ANR = 0. In the SCR-F model the SCR catalytic coating was assumed to be uniformly deposited inside the substrate wall.

As part of the 2D SCR-F model development, the N[O.sub.2]/N[O.sub.x] ratio was identified as an important parameter that determines the N[O.sub.x] reduction performance of the SCR-F. The PM oxidation reaction leads to a change in local N[O.sub.2]/N[O.sub.x] ratio across the PM cake layer that impacts the contribution of fast and standard SCR reactions.

Equations 6 to 8 are used to solve for the 2D temperature distribution in the SCR-F which is coupled with the chemical species equation.

[mathematical expression not reproducible] Eq. (6)

[mathematical expression not reproducible] Eq. (7)

[mathematical expression not reproducible] Eq. (8)

The SCR-F model simulates the change in the N[O.sub.2] concentration by both N[O.sub.2]-assisted oxidation of PM and SCR reactions. In the species model the impact of N[O.sub.2] consumption by the PM cake is taken into account in simulating the SCR reactions in the substrate wall by using the changed local N[O.sub.2]/N[O.sub.x] ratio in the wall. This aspect plays an important role in determining the final N[O.sub.x] conversion rate across the SCR-F.

The change in cake permeability and the resultant change in pressure drop across the SCR-F during PM oxidation were also simulated in this research. The calibrated SCR-F model was able to simulate the outlet exhaust gas temperature to within [+ or -]5[degrees]C, pressure drop to within [+ or -]0.1 kPa, PM mass retained to [+ or -]2 gm, and outlet N[O.sub.x] and N[H.sub.3] concentrations to within [+ or -]20 ppm of the experimental data. This accuracy made the model suitable for use in the system simulations.

SCR Model

The 1D SCR model can simulate the change in NO, N[O.sub.2], and N[H.sub.3] concentrations across a flow through SCR. A two-site storage model for N[H.sub.3] with fast, slow, and standard SCR reactions was employed to simulate N[O.sub.x] reduction and N[H.sub.3] slip. The SCR model is described in detail in Reference [2]. It was calibrated to work within a temperature range of 200-450[degrees]C.

The model is capable of simulating the outlet concentrations of NO, N[O.sub.2], and N[H.sub.3] to within [+ or -]20 ppm of the experimental data. The set of governing equations used to simulate the chemical species concentrations in the flow through SCR are described in Equations 9 to 12

[mathematical expression not reproducible] Eq. (9)

[mathematical expression not reproducible] Eq. (10)

i = NO, N[O.sub.2], N[H.sub.3]

j = ads, des, Fast SCR, Standard SCR, Slow SCR, N[H.sub.3] oxid.

[mathematical expression not reproducible] Eq. (11)

[mathematical expression not reproducible] Eq. (12)

Equations 9 and 10. are used for mass transport of chemical species by convection from the gas phase in the channel to the surface phase and from the surface phase to the catalyst surface as described in Reference [2]. The N[H.sub.3] adsorbs onto the catalyst surface by Eley-Rideal mechanism simulated by Equations 11 and 12. The first site participates in both N[H.sub.3] storage and SCR reactions where the stored N[H.sub.3] on the catalyst surface active site reacts with NO and N[O.sub.2] in the solid phase. The second storage site only participates in N[H.sub.3] storage. A detailed procedure used to determine the kinetics of all the reactions involved is described in the model calibration section.

SCR-F + SCR Model Architecture

The 2D SCR-F + 1D SCR model shown schematically in Figure 1 was developed in MATLAB/Simulink based on the ODE15s variable time step solver that solved for temperature, PM mass retained, and N[H.sub.3] coverage fraction states of the system components. The outlet temperature and chemical species concentrations of the SCR-F model are used as inputs for the SCR model.

This combined model uses the kinetics and parameters from the calibration process that was applied to each model described in References [1, 3]. This model was able to simulate the chemical species concentrations at the SCR outlet to within [+ or -]20 ppm of the experimental data. The 2D SCR-F + SCR model was able to simulate the experimental data at speedup of 30 times real time (1-h data takes 120 s) on a laptop computer with 16 GB RAM and quad-core i7 processor.

Experimental Data Overview

The 2D SCR-F model was calibrated using experimental data described in References [3, 4] and the process of calibration is described in Reference [1]. The 1D SCR model was calibrated using data collected in Reference [3]. The data collected on the SCR-F + SCR system was simulated using the parameters obtained from these datasets.

SCR Data

The data used for the 1D SCR model calibration was collected on the production aftertreatment system shown in Figure 2 on a Cummins 2013 6.7L ISB engine. The setup consists of DOC + CPF + SCR shown in Figure 2, and the two SCR-A devices are identical and have the same specifications. Each experiment consisted of three stages. Prior to each experiment in cylinder fuel dosing at 600[degrees]C temperature, at the SCR-F inlet was used to remove residual PM and adsorbed N[H.sub.3] from the CPF and SCR. The first two stages were run at the baseline condition (1660 rpm and 475 N-m) to load the CPF with PM. In the third stage a urea dosing cycle was used at N[O.sub.x] reduction test point engine conditions shown in Table 2. The emission samples at upstream DOC (UDOC), downstream DOC (DDOC), upstream SCR (USCR), and downstream SCR (DSCR) were collected for NO, N[O.sub.2], and N[H.sub.3]. In this system, the SCR-A refers to a brick with only SCR catalyst coating, whereas SCR-B is a SCR brick with AMOX coating in the last 10% of the brick length. The SCR-B was replaced with a SCR-A in order to collect the experimental N[H.sub.3] slip data which would be oxidized by the AMOX catalyst in SCR-B if it was not removed.

Seven engine conditions from Table 2 were used in the dataset with a urea dosing cycle consisting of inlet ANR values 0.3, 0.5, 0.8, 1.0, 1.2, 1.0 repeat, 0.8 repeat, and 1.2 repeat as shown in Figure 3. The data at different ANR values from 0.3 to 1.2 were used to calibrate the SCR kinetics and N[H.sub.3] storage site parameters to predict the outlet NO, N[O.sub.2], and N[H.sub.3] concentrations. ANR values 0.8, 1.0, and 1.2 repeat were used to validate repeatability of the SCR performance. The eight test points were selected to represent a wide range of exhaust temperature (208-455[degrees]C) and flow rate (4.4-16.4 kg/min) conditions that represent the operating range of the engine.

SCRF[R] + SCR Data

The data used to simulate the SCRF[R] + SCR system performance and to analyze the underlying SCR reaction rates in the SCRF[R] and SCR will be described. The schematic of the setup used for collecting these data is shown in Figure 4 consisting of the DOC + SCRF[R] + SCR. The specifications of the aftertreatment components are shown in Table 3. Hereafter, the SCRF[R] will be referred to as SCR-F.

The test procedure consisted of SCR-F cleanout and PM loading at an engine condition of 2400 rpm, 200 Nm designated as stages 1 and 2. This was followed by a ramp up stage at the same engine condition as stage 1 in order to bring the temperature of the substrate to the same value as stage 2 after weighing the filter. This was followed by the passive oxidation condition that was carried out at one of the six test point engine conditions used for the dataset as shown in Table 4. Passive oxidation is followed by stages 3 and 4 with the same engine condition as stage 2. During the passive oxidation condition, the urea was dosed into the exhaust with a target ANR range of 1.02-1.13 which was determined for each engine condition based on the SCR-F inlet N[O.sub.x]. The detailed procedure of the experiments is described in References [3, 27]. Figure 5 shows the test cycle used for the SCR-F + SCR experiments. The NO, N[O.sub.2], and N[H.sub.3] concentrations were measured at upstream of DOC (UDOC), downstream of DOC (DDOC), upstream of SCR-F (USCRF), downstream of SCR-F (DSCRF), upstream of SCR(USCR), and downstream of SCR (DSCR).

Calibration of Models

The SCR model uses a set of time-varying inputs, physical constants, and calibration parameters to simulate the SCR outlet chemical species concentrations. The input quantities used are

1. Exhaust mass flow rate

2. Exhaust gas temperature and pressure at the SCR inlet

3. Chemical species concentration of NO, N[O.sub.2], N[H.sub.3], [O.sub.2], [H.sub.2]O, and C[O.sub.2] at the SCR inlet

The calibration parameters of the 1D SCR model are found using the experimental data collected on the 2013 Cummins ISB production SCR [3]. The primary aim of the calibration process was to determine one set of kinetics that can simulate the SCR performance for all the engine conditions. The experimental data consisted of seven experiments that were conducted over a wide range of space velocity, exhaust gas temperature, and N[O.sub.2]/N[O.sub.x] ratio conditions to simulate the engine operating conditions. The kinetics and storage parameters from the Cummins ISB 2010 engine SCR from Reference [1] were used as initial values for the calibration. The cost function used for the calibration is given by Equations 13 and 14.

[mathematical expression not reproducible] Eq. (13)

[mathematical expression not reproducible] Eq. (14)

[Cost.sub.i] is the cost function with i = NO, N[O.sub.2], and N[H.sub.3]. [t.sub.0] and [t.sub.end] are the start and end times for the simulation in seconds. [e.sub.i] is the error between experimental and model concentrations. [C.sub.i, model] and [C.sub.i, exp] are the SCR outlet concentrations of the chemical species i from the model and the experimental data.

The cost function consisting of the integral of the squared error is supplied to the numerical optimizer based on fmincon function in MATLAB/Simulink which changes the calibration parameters to minimize the value of the cost function by reducing the deviation in the model and experimental outlet concentrations of NO, N[O.sub.2], and N[H.sub.3].

The SCR model consists of three SCR reactions, two adsorption, two desorption reactions, N[H.sub.3] oxidation, and the [N.sub.2]O formation reaction. These nine reactions each consists of an activation energy and pre-exponential parameters from the Arrhenius form used to model the N[O.sub.x] reduction across the SCR. These 18 parameters are found by comparing experimental NO, N[O.sub.2], and N[H.sub.3] concentrations at the SCR outlet to the 1D SCR model outlet concentration values. The parameters are listed in Table 5.

The cost function shown in Equations 13 and 14 was used in the numerical optimization scheme using the process shown in Figure 6. The activation energies of all the reactions are kept constant, and pre-exponentials of the reactions are updated for individual experiments using the numerical optimization scheme. Based on the pre-exponentials obtained, the rate constant for each reaction are calculated. These rate constants are then used in Arrhenius plots to obtain a common set of kinetics for all the reactions. The updated activation energies are used in the next step with the numerical optimizer to further improve the calibration. This iterative procedure is continued until the set of kinetics is obtained which is able to simulate the outlet NO, N[O.sub.2], and N[H.sub.3] concentrations to within [+ or -]20 ppm of the experimental data for all the experiments.

The steps used in the calibration of the SCR model are shown in Figure 7. Based on NO and N[O.sub.2] concentration data at SCR outlet, from experiments with inlet temperature less than 350[degrees]C, set 1 parameters are obtained. Set 1 consists of the kinetics of the three SCR reactions, adsorption, and desorption reactions of the first site. Once the kinetic parameters are found using the numerical optimization and Arrhenius plots, set 2 is found in the next step. Set 2 consists of the N[H.sub.3] oxidation reaction and three SCR reactions. These kinetics are found using NO, N[O.sub.2], and N[H.sub.3] outlet concentrations from experiments with inlet temperatures greater than 350[degrees]C where significant N[H.sub.3] oxidation reaction is observed.

In the final step, set 3 kinetics consisting of adsorption and desorption reactions of the second storage site, and maximum storage of the two storage sites are found. N[H.sub.3] slip data from the experiments with inlet temperature less than 350[degrees]C are used for this step. The adsorption and desorption kinetics are found based on the steady-state N[H.sub.3] slip value at ANR values 0.8 to 1.2. The transient change in the N[H.sub.3] slip pattern during transition from ANR 1.0 to 1.2 was used to find the final value of the maximum storage of the two storage sites.

The steps shown in Figure 7 are performed iteratively. Whenever a condition for a step is satisfied, then the next step is followed. If the next step fails to converge, then all the previous steps are repeated since the parameters are coupled. This iterative process is continued till a common set of kinetic parameters that satisfy all the conditions is obtained. Table 6 shows the list of kinetic parameters obtained after the calibration procedure.

The SCR-F model used the SCR-F calibration parameters obtained in the calibration process followed in Reference [1]. The list of SCR kinetics and PM oxidation kinetics used for the SCR-F are shown in Table 6. The SCR-F model PM oxidation kinetics were found using the passive oxidation experimental data collected on SCR-F without urea injection as described in Reference [4]. The PM kinetics were able to simulate the PM mass retained to within [+ or -]2 gm of the experimental value for all the seven experiments in the SCR-F + SCR data. The PM kinetics include pre-exponential and activation energy of N[O.sub.2]-assisted PM oxidation ([A.sub.NO2,oxid] and [E.sub.NO2,oxid]) and thermal PM oxidation ([A.sub.O2,oxid] and [E.sub.O2,oxid]) reactions.

The SCR-F + SCR system data were used to validate the kinetics from the two models. The experimental outlet species concentration data were simulated to within [+ or -]20 ppm using the two models with no modification to the calibration parameters of either of the individual models.

Results

The SCR-F + SCR system consists of the two major components which were simulated using the 2D SCR-F and 1D SCR model. The performance of the SCR-F (1) impacts the SCR performance downstream in terms of N[O.sub.x] reduction efficiency and N[H.sub.3] slip. The SCR-F + SCR model was able to capture these aspects of the system using calibration parameters found for the individual components. A comparison of the experimental and model outlet NO, N[O.sub.2], and N[H.sub.3] outlet concentrations is presented in Appendix A.

SCR-F + SCR System N[O.sub.x] Reduction Efficiency

The N[O.sub.x] reduction efficiency of the SCR-F + SCR was simulated using data from six experiments in the temperature range of 210-367[degrees]C. The N[O.sub.x] reduction efficiency was affected by a change in local N[O.sub.2]/N[O.sub.x] ratio in the SCR-F substrate wall due to passive oxidation of the PM cake and a change in SCR inlet N[O.sub.2]/N[O.sub.x] ratio due to N[O.sub.x] reduction in the SCR-F. Both of these factors impacted the system N[O.sub.x] efficiency significantly.

Figure 8 shows the N[O.sub.x] conversion efficiency of the SCR and SCR-F for all the SCR-F + SCR experiments. In all the cases, the model conversion efficiency values were simulated to within 1.6% of experimental values. As can be observed from the figure, the SCR conversion efficiency values are limited to a maximum of 60% due to the unfavorable N[O.sub.2]/N[O.sub.x] ratio at the SCR inlet (near 0) leading to only the standard SCR reaction being effective, which has a maximum N[O.sub.x] conversion efficiency of 60% for the given conditions.

The SCR-F conversion efficiency is around 97% for all the cases. The combined efficiency of the SCR-F and the SCR in the system is a maximum of 97.7%. The system efficiency does not increase significantly with the addition of the SCR because it is limited by the lack of N[O.sub.2] at the SCR inlet. The N[O.sub.x] conversion and N[H.sub.3] slip data for all the SCR-F + SCR system experiments is given in Appendix A.

Figure 9 compares the model and experimental SCR outlet N[H.sub.3] concentrations for all engine conditions. In all the engine conditions, the model was able to simulate the experimental N[H.sub.3] slip to within 20 ppm of the experimental data. For temperatures greater than 350[degrees]C, a 3-4% drop in N[O.sub.x] conversion due to significant N[H.sub.3] oxidation was observed.

In order to capture and explain the N[O.sub.x] reduction efficiency and N[H.sub.3] slip characteristics of this system, various internal variables related to the SCR-F and SCR were analyzed. The following subsections describe some of the important phenomena that explain these aspects of the system.

SCR-F N[O.sub.x] Reduction Efficiency

Figure 10 shows the SCR-F + SCR simulated change in N[O.sub.x] reduction efficiency of the system as a function of inlet exhaust gas temperature at ANR 1.1. The N[O.sub.x] conversion in all the cases was 97[+ or -]1% at ANR 1.1. The N[O.sub.x] conversion efficiency was limited by the inlet N[O.sub.2]/N[O.sub.x] ratio rather than by inlet temperature in the SCR-F due to a decrease in the local N[O.sub.2]/N[O.sub.x] ratio at the substrate wall inlet.

Figure 11 shows the simulated change in N[O.sub.x] conversion efficiency of Test C for the SCR-F + SCR and for the SCR-F as a function of inlet N[O.sub.2]/N[O.sub.x] ratio. The left y-axis shows the N[O.sub.x] conversion efficiency and right y-axis shows the inlet SCR-F N[O.sub.2]/N[O.sub.x] ratio. The ratio was changed in increments of 0.1 from 0.2 to 1.0 keeping PM loading of 2 g/L and inlet temperature of 339[degrees]C constant.

The N[O.sub.x] conversion efficiency at 0.3 hr. and N[O.sub.2]/N[O.sub.x] ratio was observed to be 94%. With an increase in N[O.sub.2]/N[O.sub.x] ratio, the conversion efficiency increases to 97.5% at N[O.sub.2]/N[O.sub.x] = 0.5, time t = 1.2 hr. Upon further increase in the ratio, the conversion efficiency decreases to a value of 84% at N[O.sub.2]/N[O.sub.x] = 1.0. The increase in N[O.sub.x] conversion efficiency is due to an increase in the fast SCR reaction rate from 0 to 0.5 hr. At values above 0.5 hr., the fast SCR reaction rate decreases and the slow SCR reaction rate starts increasing; since the slow SCR reaction is slower than the standard SCR reaction, the decrease in conversion efficiency for values >0.5 hr. is higher compared to <0.5 hr. (94% at 0.2 hr. vs 84% at 1.0 hr.). This change in conversion efficiency of the SCR-F coupled with the downstream SCR N[O.sub.x] conversion efficiency plays an important role in determining the SCR-F+SCR system performance.

SCR Baseline Results

The baseline SCR results, for the experiments and model, describe the SCR performance when it is present downstream of the CPF. ANR values of 0.3 to 1.2 were used to determine the N[O.sub.x] reduction and N[H.sub.3] slip performance. The N[O.sub.x] reduction performance increased with an increase in inlet ANR value. The N[H.sub.3] slip was observed to be near zero for ANR values < 1.0 and a significant increase in N[H.sub.3] slip at ANR 1.2 was observed. N[O.sub.x] conversion, SCR reaction rate, and N[H.sub.3] slip characteristics as a function of SCR N[O.sub.2]/N[O.sub.x] ratio were developed.

Figures 12 and 13 show the experimental and 1D SCR model inlet, outlet concentrations of NO, N[O.sub.2], and N[H.sub.3] versus time for Tests 1 and 8. Test 1 was conducted at inlet temperature of 218[degrees]C, space velocity (SV) of 12 k/hr., inlet N[O.sub.x] of 632 ppm, and inlet N[O.sub.2]/N[O.sub.x] ratio of 0.27. Due to the low temperature, the adsorption and desorption rates for the two storage sites were observed to be low leading to a slow response time of the SCR to changes in inlet ANR. N[H.sub.3] oxidation was observed to be negligible in this case with the second site being responsible for the N[H.sub.3] slip characteristics observed in Figure 12.

The outlet NO, N[O.sub.2], and N[H.sub.3] concentrations were simulated to within [+ or -]20 ppm of the experimental data in this experiment. An increase in N[O.sub.x] conversion efficiency with an increase in ANR is observed with maximum efficiency of 88% observed at ANR 1.2 and time t = 130 min. The N[H.sub.3] slip remains near zero for ANR values 0 to 1.0 with a significant increase to 180 ppm at time t = 120 min, inlet ANR 1.2.

The data in Figure 13 for Test 8 was conducted at inlet temperature of 447[degrees]C and SV of 45 k/hr. with inlet N[O.sub.x] of 510 ppm and N[O.sub.2]/N[O.sub.x] ratio of 0.18. Significant N[H.sub.3] oxidation was observed in this experiment leading to a lower N[H.sub.3] slip than expected at ANR 1.0 and 1.2. Due to the higher temperature, the response time of the SCR to changes in inlet ANR values was fast and the second site had a low amount of N[H.sub.3] storage. Significant N[H.sub.3] slip was observed from ANR 0.3 to 1.2 due to high desorption rate caused by the high inlet temperature and low inlet N[O.sub.2]/N[O.sub.x] ratio. The NO, N[O.sub.2], and N[H.sub.3] concentrations were simulated to [+ or -]15 ppm of the experimental data. The transient change in N[H.sub.3] slip was also captured as a result of the SCR kinetics of adsorption and desorption reactions for the two N[H.sub.3] storage sites. The following section explores the various internal variables that resulted from the simulation of the seven experiments used in this dataset.

N[O.sub.x] Reduction Performance of the SCR The N[O.sub.x] reduction performance of the SCR increases with an increase in the SCR inlet temperature up to a temperature of 350[degrees]C. This increase in N[O.sub.x] reduction performance can be attributed to an increase in the fast SCR reaction rate between 200 and 350[degrees]C. At temperatures above 350[degrees]C, an increase in the N[H.sub.3] oxidation leads to a decrease in the N[O.sub.x] reduction performance of the SCR.

Figure 14 shows the simulated trend of change in N[O.sub.x] reduction performance with SCR inlet temperature at ANR = 1.0. A maximum N[O.sub.x] reduction of 96% was observed at an inlet temperature of 350[degrees]C. In Figure 14, each of the seven test points was run using the SCR model with inlet temperature of 200 to 470[degrees]C in increments of 35[degrees]C in order to obtain the continuous curves. These curves intersect with the model N[O.sub.x] conversion efficiency values shown by red circles at the seven test points. These values are comparable to the experimental data shown in black circles. For all the cases, the N[O.sub.x] conversion efficiency of the model was within [+ or -]3% of the experimental data.

Table 7 compares the experimental and model N[O.sub.x] conversion efficiency for all the experiments in the SCR baseline data. A maximum N[O.sub.x] conversion efficiency of 96% was obtained from the SCR at ANR 1.0.

N[H.sub.3] Storage Characteristics of SCR Figure 15 shows the coverage fraction of sites 1 and 2 for the seven experiments as a function of SCR inlet temperature. A maximum storage capacity of 43 and 42 gmol/[m.sup.3], respectively, were identified for the two storage sites 1 and 2 in the SCR.

The coverage fraction increases with an increase in the ANR value. The coverage fraction of the second site increases significantly for inlet ANR values > 1.0. The first-site coverage fraction increases with an increase in ANR up to ANR 1.0 leading to a corresponding increase in N[O.sub.x] reduction performance of the SCR.

With an increase in SCR inlet temperature, the coverage fraction of both sites decreases due to an increased desorption and consumption of N[H.sub.3] by the SCR reactions. The contribution of the second storage site compared to the first storage site decreases with an increase in temperature. This change in the second storage site leads to a faster response time of the SCR for changes in inlet ANR at higher temperatures.

SCR N[H.sub.3] Slip Characteristics The N[H.sub.3] slip characteristics of the SCR are a function of storage capacity of the two storage sites, exhaust gas temperature, inlet ANR, and N[O.sub.2]/N[O.sub.x] ratio. The experimental N[H.sub.3] slip values were simulated to within [+ or -]20 ppm of the experimental data for all the experiments. This calibration was obtained based on the N[H.sub.3] slip characteristics of the SCR in the low temperature (<300[degrees]C) experiments.

Figure 16 shows the comparison of the experimental and model N[H.sub.3] slip values at ANR = 1.0 for all the seven experiments. The N[H.sub.3] slip for experiments with T > 350[degrees]C are impacted by the N[H.sub.3] oxidation reaction which significantly reduces the N[H.sub.3] slip at inlet ANR > 1.0. The transient response of the N[H.sub.3] slip as a function of inlet ANR is a function of exhaust gas temperature due to the change in the desorption reactions at both storage sites.

The N[H.sub.3] slip was found to be a strong function of the inlet N[O.sub.2]/N[O.sub.x] ratio as can be observed in experiments at SCR inlet temperatures 340 and 352[degrees]C in Figure 16 where a near-zero slip at ANR 1.0 is observed. Figure 17 shows the model-simulated relation between N[O.sub.2]/N[O.sub.x] ratio vs N[H.sub.3] slip for conditions at test point 5 at ANR = 1.2. A decrease in N[H.sub.3] slip is observed in N[O.sub.2]/N[O.sub.x] from 0 to 0.5 followed by an increase in N[H.sub.3] slip from 0.5 to 0.9. This change in N[H.sub.3] slip can be attributed to increased N[H.sub.3] consumption by the fast SCR reaction. The red dot represents experimental N[H.sub.3] slip at this engine condition. For temperatures greater than 350 C, a 3-4% drop in N[O.sub.x] conversion due to significant N[H.sub.3] oxidation was observed along with reduced N[H.sub.3] slip.

Contribution of the Three SCR Reactions Figure 18 shows the change in contribution of each of the three SCR reactions with a change in inlet N[O.sub.2]/N[O.sub.x] ratio obtained from the SCR model for the conditions in the seven baseline SCR experiments. At low N[O.sub.2]/N[O.sub.x] ratio conditions in experiments Tests 8, 1, and 3 (<0.35), the standard SCR reactions led to 30-60% of the N[O.sub.x] reduction. The fast SCR reaction was responsible for 40-70% of the N[O.sub.x] reduction with the standard SCR reaction being less than 10% due to the absence of available N[O.sub.2] for the reaction.

At N[O.sub.2]/N[O.sub.x] values above 0.35, the fast SCR reaction is predominantly responsible for the SCR reactions (Tests 2, 4, 5, and 6). Up to 80% of N[O.sub.x] reduction takes place due to the fast SCR reaction and standard SCR reaction leading to a further 10-20% N[O.sub.x] reduction. The slow SCR reaction starts having significant impact on N[O.sub.x] reduction (>10%) at N[O.sub.2]/N[O.sub.x] ratio 0.35.

The decrease in the standard SCR reaction with increase in N[O.sub.2]/N[O.sub.x] is due to the consumption of NO by the fast SCR reaction. Due to the rate constants of the SCR reactions, the standard SCR reaction is only able to consume excess NO left after the fast SCR reaction. This excess NO is a function of N[O.sub.2]/N[O.sub.x] ratio which for these engine conditions leads to higher NO compared to N[O.sub.2] at the SCR inlet.

The slow SCR reaction is limited by the lower N[O.sub.2] concentration. At N[O.sub.2]/N[O.sub.x] ratio of 0-0.25 due to a lower fast SCR reaction, the slow SCR reaction has a contribution of 8-10%. At 0.25 to 0.35 due to the increased fast SCR reaction, the slow SCR reaction rate is zero. Above 0.35, there is an increase in available N[O.sub.2] that is left after consumption of N[O.sub.2] by the fast SCR reaction. This enables an increase of the slow SCR reaction to 10%.

The trends observed in the plots indicate a further increase in the fast SCR reaction with an increase in N[O.sub.2]/N[O.sub.x] ratio to 0.5 and a decrease in the standard SCR reaction with the standard SCR reaction being <10% for these values. Also, the slow SCR reaction shows an upward trend in terms of N[O.sub.2]/N[O.sub.x] ratio.

Figure 19 shows the change in N[O.sub.x] conversion efficiency for the seven experiments with a change in N[O.sub.2]/N[O.sub.x] ratio obtained from the 1D SCR model. The N[O.sub.2]/N[O.sub.x] ratio of each experiment was changed in increments of 0.1 from 0 to 1 keeping other conditions constant.

The trends observed in Figure 19 indicate a strong dependency of N[O.sub.x] reduction performance of the SCR on the SCR inlet N[O.sub.2]/N[O.sub.x] ratio. Maximum N[O.sub.x] conversion efficiency in all cases was observed at N[O.sub.2]/N[O.sub.x] ratio of 0.5 which indicates the importance of the fast SCR reaction since the fast SCR reaction reaches a maximum value at a N[O.sub.2]/N[O.sub.x] ratio of 0.5.

Contribution of SCR Reactions in SCR-F and SCR The change in N[O.sub.2]/N[O.sub.x] ratio has significant impact on N[O.sub.x] conversion performance of the system as observed in Figure 19. Table 8 shows the change in N[O.sub.2]/N[O.sub.x] ratio as the exhaust gas passes through PM cake and substrate wall in the SCR-F and SCR. At the inlet of the SCR-F, the N[O.sub.2]/N[O.sub.x] ratio is a function of the DOC NO conversion efficiency which in turn is a function of exhaust gas temperature and space velocity. An inlet N[O.sub.2]/N[O.sub.x] ratio of 0.29 to 0.48 was observed for the six engine conditions in the SCR-F + SCR experiments. As the exhaust gas passes through the PM cake, a significant decrease in N[O.sub.2] concentration takes place due to passive oxidation of PM. This leads to a decrease in the N[O.sub.2]/N[O.sub.x] ratio at the SCR-F wall inlet. The change in N[O.sub.2]/N[O.sub.x] ratio across the PM cake is variable and dynamic, and it depends on the exhaust gas temperature, PM cake thickness, and available N[O.sub.2] in the PM cake. As a result, the effective local N[O.sub.2]/N[O.sub.x] ratio changes to a range of 0.22 to 0.39 for the six experiments. At the SCR-F exit, the N[O.sub.2] concentration for all cases was observed to be zero leading to a N[O.sub.2]/NO ratio of zero for all cases at the SCR inlet.

Figure 20 shows the contribution of the three SCR reactions to the overall N[O.sub.x] conversion in the SCR-F and SCR. These values are a function of N[O.sub.2]/N[O.sub.x] ratio from Table 7. For the SCR-F, the fast SCR reaction starts at 70% at a ratio of 0.29 and increases to 82% for an inlet ratio of 0.48. The standard SCR reaction starts at 30%, but at values greater than 0.38, a significant decrease occurs to less than 20%. The slow SCR reaction is zero at 0 to 0.38, and beyond this value, the slow SCR reaction reaches a value of 10%.

In the SCR where the N[O.sub.2]/N[O.sub.x] ratio is zero for all the cases, the fast and slow SCR reactions shown in black and blue dotted lines remain near zero due to a lack of available N[O.sub.2]. The standard SCR reaction contributes to nearly 100% of all the N[O.sub.x] conversion efficiency in the SCR for all experiments.

N[H.sub.3] Slip from SCR-F and SCR The N[H.sub.3] slip characteristics of SCR-F and SCR were observed to be a strong function of N[O.sub.2]/N[O.sub.x] ratio and inlet temperature. Figure 21 shows the change in SCR-F N[H.sub.3] slip as a function of inlet N[O.sub.2]/N[O.sub.x] ratio for test point C with increments in N[O.sub.2]/N[O.sub.x] ratio of 0.1 from the 2D SCR-F model. The N[H.sub.3] slip of the SCR-F decreases from 40 to 20 ppm with an increase in N[O.sub.2]/N[O.sub.x] ratio from 0.1 to 0. From 0.5 to 1.0, the slip increases to 30 ppm. The change in NO and N[O.sub.2] outlet concentrations is also shown in the plot. N[O.sub.2]/N[O.sub.x] ratio of 0.5 leads to maximum N[O.sub.x] conversion and minimum N[H.sub.3] slip. The N[H.sub.3] slip value at N[O.sub.2]/N[O.sub.x] ratio of 1 is lower than at ratio of 0.1 due to the higher N[H.sub.3] consumption per mole of N[O.sub.x] reduced by the slow SCR reaction compared to the standard and fast SCR reactions.

Impact of Local N[O.sub.2]/N[O.sub.x] Ratio on System Performance Local N[O.sub.2]/N[O.sub.x] ratio at the SCR inlet was found to be the important parameter that impacts the system performance in terms of N[O.sub.x] reduction and N[H.sub.3] slip characteristics. The lack of N[O.sub.2] at the inlet to the SCR leads to a condition where the fast SCR reaction is near zero leading to low N[O.sub.x] reduction (<70%) and low adsorption of N[H.sub.3] leading to excess N[H.sub.3] slip compared to the case with optimal N[O.sub.2]/N[O.sub.x] ratio of 0.5 at the SCR inlet. This impact is higher at low temperatures (<300[degrees]C) where the standard SCR reaction rate is low. Figures 22a and 22b show the change in N[O.sub.x] conversion efficiency of the downstream SCR for all the experiments for N[O.sub.2]/N[O.sub.x] ratio = 0 and 0.5 at the SCR inlet using the SCR-F model.

Figure 22a compares different variables at the inlet of the SCR and internal variables from the SCR model used to simulate the SCR performance for SCR-F + SCR experiments at N[O.sub.2]/N[O.sub.x] = 0. The top plot shows the experimental and model N[O.sub.x] conversion efficiency which were simulated to within [+ or -]3% of experimental data. The second plot shows the inlet N[O.sub.x] and N[H.sub.3] concentration for each experiment on the left y-axis in parts per million. The right y-axis shows the inlet SCR ANR for each experiment which was found to be >1 for all experiments. The third plot on the left y-axis shows the space velocity in k/hr. which was observed to be less than 50 k/hr. for all experiments. The left axis shows the adsorption rate of N[H.sub.3] onto the catalyst surface in kmol/[m.sup.3]s. This value was found to be low for N[O.sub.2]/N[O.sub.x] = 0 cases leading to excess N[H.sub.3] slip and low N[O.sub.x] conversion efficiency (<50%) for most cases.

Figure 22b based on the SCR-F + SCR model shows the SCR-F + SCR model N[O.sub.x] conversion efficiency for N[O.sub.2]/N[O.sub.x] = 0.5 with SCR-F + SCR experimental inlet conditions in the top plot in Figure 22a. The second plot in Figure 22b shows inlet N[O.sub.x] and N[H.sub.3] concentrations and inlet ANR similar to Figure 22a. The bottom plot has similar space velocity values as Figure 22a: however the N[H.sub.3] adsorption rate for a given experiment was observed to be 2-5 times higher compared to N[O.sub.2]/N[O.sub.x] = 0 case due to lower coverage fraction and higher N[H.sub.3] consumption by fast SCR reaction leading to >90% N[O.sub.x] conversion in all the cases except for T > 360[degrees]C, where the N[H.sub.3] oxidation leads to a lower N[O.sub.x] conversion rate of 87%.

The main reasons for the observed trends in Figure 22a and 22b are as follows:

1. Low inlet concentration of N[H.sub.3] in Figure 22a into the SCR led to low adsorption rate caused by mass limitation for the given

flow rate conditions. This mass transfer limitation led to <50% N[O.sub.x] conversion for the SCR-F + SCR experiments which is consistent with the experimental data.

2. For N[O.sub.2]/N[O.sub.x] = 0.5, the mass transfer limitation was overcome due to the fast SCR reaction which increased the N[H.sub.3] adsorption to a value where the mass transfer limitation was not observed to be a limitation leading to N[O.sub.x] conversion >90% as shown in Figure 22b.

3. At T > 350[degrees]C, the N[H.sub.3] oxidation led to a decrease in N[O.sub.x] conversion for the N[O.sub.2]/N[O.sub.x] = 0.5 case as shown in Figure 22b.

It can be observed that for the low temperature experiments (T < 300[degrees]C), up to a 70% increase in N[O.sub.x] conversion efficiency can be achieved with an optimum N[O.sub.2]/N[O.sub.x] ratio; while for experiments with T > 350[degrees]C, a 30-50% increase is expected. This increase in downstream SCR performance leads to an increase in system N[O.sub.x] conversion performance from 97.7% to 99.5% which is required for a potential system that can achieve the ultra-low N[O.sub.x] standard.

Increasing the N[O.sub.x] conversion in the SCR by optimum N[O.sub.2]/N[O.sub.x] ratio could increase the low temperature performance of the system significantly since the fast SCR reaction rate is higher compared to the standard SCR reaction at temperatures less than 350[degrees]C. Table 9 shows the experimental and model N[O.sub.x] conversion efficiencies for all the SCR-F + SCR experiments. In all cases the model was able to simulate the outlet N[O.sub.x] to within 15 ppm of the experimental data and N[O.sub.x] conversion to within 1.6% of experimental data.

Summary and Conclusions

A system model consisting of a 2D SCR-F and 1D SCR model was developed that is capable of simulating the N[O.sub.x] conversion across the system consisting of a SCRF[R] with a downstream SCR. The 1D SCR model was calibrated using data from the baseline production SCR with a 2013 Cummins ISB engine to simulate the N[O.sub.x] conversion and N[H.sub.3] storage characteristics of the device for a range of engine conditions.

The combined model was run with experimental data collected on the 2013 ISB Cummins engine with a Johnson Matthey SCRF[R] and a downstream production SCR-A brick. The model was able to simulate the NO, N[O.sub.2], and N[H.sub.3] SCR outlet concentrations to within 15 ppm of the experimental values for all the eight experiments used. The major findings from this work are

1. The baseline SCR downstream of the CPF had a maximum N[O.sub.x] conversion efficiency of 97.5% at a N[O.sub.2]/N[O.sub.x] ratio of 0.5.

2. The N[H.sub.3] slip from the baseline SCR increased significantly for inlet ANR values > 1.0.

3. The N[O.sub.x] conversion efficiency in the SCR-F in the SCR-F + SCR system is impacted by the N[O.sub.2]-assisted oxidation rate in the PM cake resulting in a decrease in local N[O.sub.2]/N[O.sub.x] ratio by 0.01 to 0.16 in the substrate wall compared to PM cake causing lower N[O.sub.x] reduction efficiency.

4. For the engine conditions used in the SCR-F+SCR system experiments, the outlet N[O.sub.2]/N[O.sub.x] ratio was observed to be near zero for all cases. Low N[O.sub.2]/N[O.sub.x] ratio at the SCR inlet limits its N[O.sub.x] reduction performance to a maximum value of 60%. At low temperatures (<300[degrees]C), the N[O.sub.x] conversion decreases further to less than 50%.

5. The combined efficiency of the SCR-F + SCR system is limited to 97.7% and decreases to 97% for inlet temperatures less than 350[degrees]C where the SCR in effect provides negligible N[O.sub.x] reduction.

6. The SCR-F + SCR system performance can be improved by increasing the N[O.sub.2]/N[O.sub.x] ratio at the SCRF[R] inlet resulting in favorable N[O.sub.x] conversion in the SCR.

7. N[H.sub.3] slip in the SCR-F + SCR system downstream SCR is significant although the inlet N[H.sub.3] concentrations are low due to the low N[O.sub.x] conversion and low residence time of the exhaust gas in the SCR compared to the N[H.sub.3] adsorption rate.

The N[O.sub.2] at the inlet of SCR can be potentially increased by increasing the N[O.sub.2]/N[O.sub.x] ratio at SCR-F inlet to greater than 0.5. In order to achieve this the performance of the DOC needs to be improved significantly. Possible ways of achieving this objective are

1. Increase the length of DOC

2. Increased DOC catalyst loading

3. Improved catalyst with better N[O.sub.x] conversion especially at low temperatures

Contact Information

Venkata Rajesh Chundru

Michigan Technological University

Houghton, MI, USA 49931

Phone: +19062818993

vrchundr@mtu.edu

Definitions/Abbreviations

1D - One dimensional

2D - Two dimensional

AMOX - Ammonia oxidation catalyst

ANR - Ammonia to N[O.sub.x] ratio

ASC - Ammonia slip catalyst

[C.sub.12][H.sub.24] - Dodecane

CARB - California Air Resources Board

CO - Carbon monoxide

C[O.sub.2] - Carbon dioxide

CPF - Catalyzed particulate filter

CSF - Catalyzed soot filter

CuO - Cupric oxide

Cu-Ze - Copper zeolite

DPF - Diesel particulate filter

DOC - Diesel oxidation catalyst

EPA - Environmental Protection Agency

FTP - Federal test procedure

[H.sub.2]O - Water

MB - Mini burner

[N.sub.2]O - Nitrous oxide

N[H.sub.3] - Ammonia

N[O.sub.2] - Nitrogen dioxide

N[O.sub.x] - Oxides of nitrogen

[O.sub.2] - Oxygen

PM - Particulate matter

PNA - Passive N[O.sub.x] adsorber

SCR - Selective catalytic reduction

SCR-A - SCR brick

SCR-B - SCR brick with AMOX coating at the end

SCR-F - SCR catalyst on a DPF

SCRF[R] - Johnson Matthey SCR-F

WHTC - World Harmonized Transient Cycle

[epsilon] - Void fraction

[A.sub.g] - Geometric surface area

a - Channel width

[[beta].sub.j] - Mass transfer coefficient

[C.sub.1,l] - Inlet channel concentration of species

[C.sub.1s,l] - Inlet channel - wall boundary concentration

[C.sub.2,l] - Concentration in the outlet channel

[C.sub.2s,l] - Concentration at outlet channel - wall boundary

[C.sub.c] - Specific heat of cake

[C.sub.f,l] - Concentration of species 1 in wall and PM cake

[C.sup.n.sub.g] - Concentration of species in gas phase at time n

[c.sub.p] - Constant pressure specific heat of exhaust gas

[c.sub.v] - Constant volume specific heat of exhaust gas

[C.sub.w] - Specific heat of substrate wall

[D.sub.1] - Diffusion rate

[e.sub.NH3] - Error in N[H.sub.3] concentration

[e.sub.NO] - Error in NO concentration

[??] - Error in N[O.sub.2] concentration

[f.sub.co] - Partial oxidation factor for thermal PM oxidation

[g.sub.co] - Partial oxidation factor for passive oxidation

[K.sub.1] - Mass transfer coefficient in inlet channel

[K.sub.2] - Mass transfer coefficient in outlet channel

l - Species index

m - Index for reactions

[m.sub.1,in] - Mass flow rate inlet channel

[m.sub.2,in] - Mass flow rate in outlet channel

[[xi].sub.l], m - Stoichiometric coefficient

[[ohm].sub.1] - Maximum storage capacity of storage site 1

[[ohm].sub.2] - Maximum storage at storage site 2

[Q.sub.1] - Convection heat transfer rate in inlet channel

[Q.sub.2] - Convection heat transfer rate in outlet channel

[Q.sub.f] - Heat transfer rate in filter by conduction, convection, energy release by reactions and heat loss to ambient

[[rho].sub.c] - Density of PM cake

[[rho].sub.g] - Density of exhaust gas

[[rho].sub.w] - Density of substrate wall

[R.sub.ads,1] - Adsorption reaction rate at site 1

[R.sub.ads,2] - Adsorption reaction at site 2

[R.sub.des,1] - Desorption reaction rate at site 2

[R.sub.des,2] - Desorption reaction at site 2

[R.sub.fst] - Fast SCR reaction rate

[R.sub.m] - Reaction rate

[R.sub.oxid] - N[H.sub.3] oxidation reaction rate

[R.sub.slo] - Slow SCR reaction rate

[R.sub.std] - Standard SCR reaction rate

[[theta].sub.1] - Coverage fraction of storage site 1

[[theta].sub.2] - Coverage fraction at storage site 2

[T.sub.1] - Temperature in the inlet channel

[T.sub.2] - Temperature in the outlet channel

[V.sub.1] - Volume of inlet channel

[V.sub.f] - Volume of filter

[V.sub.2] - Volume of outlet channel

[v.sub.1] - Exhaust gas velocity in inlet channel

[v.sub.2] - Exhaust gas velocity in the outlet channel

[v.sub.f] - Exhaust gas velocity in the substrate wall

References

[1.] Chundru, V. R., Mahadevan, B. S., Johnson, J. H., Parker, G. G., and Shahbakhti, M., "Development of a 2D Model of a SCR Catalyst on a DPF," J. Emiss. Control. Sci Technol. 5:133-171, 2019, doi:10.1007/s40825-019099115-4.

[2.] Song, X., "A SCR Model Based on Reactor and Engine Experimental Studies for a Cu-Zeolite Catalyst," Ph.D. dissertation, Michigan Technological University, 2013.

[3.] Kadam, V., "An Experimental Investigation of the Effect of Temperature and Space Velocity on the Performance of a Cu-Zeolite Flow-through SCR and a SCR Catalyst on a DPF with and without PM Loading," M.S. thesis, Michigan Technological University,2016.

[4.] Gustafson, E. A., "An Experimental Investigation into N[O.sub.2] Assisted Passive Oxidation with and without Urea Dosing and Active Regeneration of Particulate Matter for a SCR Catalyst on a DPF," M.S. thesis, Michigan Technological University, 2016.

[5.] Song, X., Johnson, J. PL, and Naber, J. D., "A Review of the Literature of Selective Catalytic Reduction Catalysts Integrated into Diesel Particulate Filters," International journal of Engine Research 16(6):738-749, 2015.

[6.] Park, S.-Y., Narayanaswamy, K., Schmieg, S. J., and Rutland, C. J., "A Model Development for Evaluating Soot-NOx Interactions in a Blended 2-Way Diesel Particulate Filter/Selective Catalytic Reduction," Industrial & Engineering Chemistry Research 51(48):15582-15592, 2012.

[7.] Park, S. Y., Rutland, C. J., Narayanaswamy, K., Schmieg, S. J. et al., "Development and Validation of a Model for Wall-Flow Type Selective Catalytic Reduction System," Proceedings of the Institution of Mechanical Engineers, Part D: Journal of Automobile Engineering 225(12):1641-1659, 2011.

[8.] Colombo, M., Koltsakis, G., and Koutoufaris, I., "A Modeling Study of Soot and De-NOx Reaction Phenomena in SCRF Systems," SAE Technical Paper 2011-37-0031, 2011, doi:10.4271/2011-37-0031.

[9.] Yang, Y, Cho, G., and Rutland, C, "Model Based Study of DeNOx Characteristics for Integrated DPF/SCR System over Cu-Zeolite," SAE Technical Paper 2015-01-1060, 2015, doi:10.4271/2015-01-1060.

[10.] Lopez-De Jesus, Y M., Chigada, P. I., Watling, T. C, Arulraj, K. et al., "NOx and PM Reduction from Diesel Exhaust Using Vanadia SCRF[R]," SAE International Journal of Engines 9(2):1247-1257, 2016.

[11.] Dosda, S., Berthout, D., Mauviot, G., and Nogre, A., "Modeling of a DOC SCR-F SCR Exhaust Line for Design Optimization Taking into Account Performance Degradation due to Hydrothermal Aging," SAE International Journal of Fuels and Lubricants 9(3):621-632, 2016.

[12.] Tan, J., Solbrig, C, and Schmieg, S. J., "The Development of Advanced 2-Way SCR/DPF Systems to Meet Future Heavy-Duty Diesel Emissions," SAE Technical Paper 2011-01-1140, 2011, doi:10.4271/2011-01-1140.

[13.] Tronconi, E., Nova, I., Marchitti, F., Koltsakis, G. et al., "Interaction of NOx Reduction and Soot Oxidation in a DPF with Cu-Zeolite SCR Coating," Emission Control Science and Technology 1(2):134-151, 2015.

[14.] Watling, T. C, Ravenscroft, M. R., and Avery, G., "Development, Validation and Application of a Model for an SCR Catalyst Coated Diesel Particulate Filter," Catalysis Today 188(1):32-41, 2012.

[15.] Naseri, M., Chatterjee, S., Castagnola, M., Chen, H.-Y et al., "Development of SCR on Diesel Particulate Filter System for Heavy Duty Applications," SAE International Journal of Engines 4(1):1798-1809, 2011, doi:10.4271/2011-01-1312.

[16.] Cavataio, G., Girard, J. W., and Lambert, C. K., "Cu/Zeolite SCR on High Porosity Filters: Laboratory and Engine Performance Evaluations," SAE Technical Paper 2009-01-0897, 2009, doi:10.4271/2009-01-0897.

[17.] Tang, W., Youngren, D., Santa Maria, M., and Kumar, S., "On-Engine Investigation of SCR on Filters (SCRoF) for HDD Passive Applications," SAE International Journal of Engines 6(2):862-872, 2013, doi:10.4271/2013-01-1066.

[18.] Lee, J. H., Paratore, M. J., and Brown, D. B., "Evaluation of Cu-Based SCR/DPF Technology for Diesel Exhaust Emission Control," SAE International Journal of Fuels and Lubricants 1(1):96-101, 2009, doi:10.4271/2008-01-0072.

[19.] Mihai, O., Tamma, S., Stenfeldt, M., and Olsson, L.,"Effect of Soot on the SCR Reactions in an Integrated SCR Coated DPF," in 24th North American Catalysis Society Meeting, Pittsburgh, PA, 2015.

[20.] Mihai, O., Tamm, S., Stenfeldt, M., Wang-Hansen, C, and Olsson, L., "Evaluation of an Integrated Selective Catalytic Reduction-Coated Particulate Filter," Industrial & Engineering Chemistry Research 54(47):11779-11791, 2015.

[21.] Cumaranatunge, L., Chiffey, A., Stetina, J., McGonigle, K. et al., "A Study of the Soot Combustion Efficiency of an SCRF[R] Catalyst vs a CSF during Active Regeneration," Emission Control Science and Technology 3(1):93-104, 2017.

[22.] Strots, V, Kishi, A., Adelberg, S., and Kramer, L., "Application of Integrated SCR/DPF Systems in Commercial Vehicles," in JSAE Annual Congress, 454-20145174, 2014.

[23.] Sharp, C, Webb, C. C, Neely, G., Carter, M. et al., "Achieving Ultra Low NOX Emissions Levels with a 2017 Heavy-Duty On-Highway TC Diesel Engine and an Advanced Technology Emissions System - NOx Management Strategies," SAE International Journal of Engines 10:1736-1748, 2017, doi:10.4271/2017-01-0958.

[24.] Sharp, C, Webb, C. C, Neely, G., Carter, M. et al., "Achieving Ultra Low N[O.sub.x] Emissions Levels with a 2017 Heavy-Duty On-Highway TC Diesel Engine and an Advanced Technology Emissions System - Thermal Management Strategies," SAE International Journal of Engines 10:1697-1712, 2017, doi:10.4271/2017-01-0954.

[25.] Sharp, C., Webb, C. C, Yoon, S., Carter, M., and Henry, C, "Achieving Ultra Low NOX Emissions Levels with a 2017 Heavy-Duty On-Highway TC Diesel Engine - Comparison of Advanced Technology Approaches," SAE International Journal of Engines 10:1722-1735, 2017, doi:10.4271/2017-01-0956.

[26.] Georgiadis, E., Kudo, T., Herrmann, O., Uchiyama, K., and Hagen, J., "Real Driving Emission Efficiency Potential of SDPF Systems without an Ammonia Slip Catalyst," SAE Technical Paper 2017-01-0913, 2017, doi:10.4271/2017-01-0913.

[27.] Sharma, S., "The Emission and Particulate Matter Oxidation Performance of a SCR Catalyst on a Diesel Particulate Filter with a Downstream SCR," M.S. report, Michigan Technological University, 2017.

Appendix A: 2D SCR-F + 1D SCR Model Results

Tables A.1 to A.3 compare the NO, N[O.sub.2], and N[H.sub.3] emissions from the 2D SCR-F + 1D SCR model and SCRF[R] + SCR experimental data.

Appendix B: 2D SCR-F Model Test C Results

Figures A.1 to A.3 show the 2D temperature, PM mass retained, and coverage fraction distribution from the 2D SCR-F model. The 2D SCR-F model for the SCRF[R] was able to simulate the experimental temperature distribution in the SCRF[R] to within 5[degrees]C of the experimental data and the resultant distributions of PM and N[H.sub.3] coverage fraction govern the outlet NO and N[O.sub.2] concentrations of the SCRF[R] which were used as input for the 1D SCR model to simulate the SCRF[R] + SCR system.

In Figure A.3, the storage site 1 distribution determines the N[O.sub.x] conversion performance of the SCR-F which impacts the SCR inlet N[O.sub.2]/N[O.sub.x] ratio and SCR-F + SCR system performance. The coverage fraction of site 1 is a function of temperature distribution leading to higher coverage fraction on outer edges at radius >100 mm due to lower temperature. A decrease in axial coverage fraction at length of 200 mm to 0.07 is due to diffusion of chemical species in the axial direction in the inlet channel leading to lower reaction rate at length >200 mm. The coverage fraction of site 2 follows a similar trend as site 1 with much lower coverage fraction value. The second site distribution impacts the SCR-F outlet N[H.sub.3] concentrations. The coverage fraction of site 1 also impacts the PM oxidation rate and thus PM distribution shown in Figure A.2 by impacting the available N[O.sub.2] in the cake.

Venkata Rajesh Chundru, Gordon G. Parker, and John H. Johnson, Michigan Technological University, USA

History

Received: 08 Mar 2019

Revised: 27 May 2019

Accepted: 06 Jun 2019

e-Available: 14 Jun 2019

doi:10.4271/04-12-02-0008

(1) The "Results" presented do not use as figures any of the 2D output of temperature, PM mass retained, and N[H.sub.3] coverage fractions, but the authors felt it was important to show these 2D data in Appendix B for Test C in order to give the reader a feel of the SCR-F model capability which is described in Reference [1].
TABLE 1 2D SCR-F model reactions.

Reaction name                Reaction (1)

N[H.sub.3] adsorption/       N[H.sub.3] [left right arrow] N[H.sub.3]*
desorption at site 1 and 2
Fast SCR                     N[H.sub.3]* + 1/2NO + 1/2N[O.sub.2] [right
                             arrow] [N.sub.2] + 3/2[H.sub.2]O
Slow SCR                     N[H.sub.3]* + 3/4N[O.sub.2] [right arrow]
                             7/8[N.sub.2] + 3/2[H.sub.2]O
Standard SCR                 N[H.sub.3]* + NO + 1/4[O.sub.2]
                             [right arrow] [N.sub.2] + 3/2[H.sub.2]O
N[H.sub.3] oxidation         N[H.sub.3]* + 3/4[O.sub.2] [right arrow]
                             1/2[N.sub.2] + 3/2[H.sub.2]O
NO oxidation                 NO + 1/2[O.sub.2] [right arrow] N[O.sub.2]
Passive PM oxidation         C+(2 - [g.sub.CO])
                             N[O.sub.2] [right arrow] [g.sub.CO]CO +
                             (1-[g.sub.CO]) C[O.sub.2] +
                             (2 - [g.sub.CO])NO
Thermal PM oxidation         C + [(1 - [f.sub.CO]/2)][O.sub.2]
                             [right arrow] [f.sub.CO]CO + (1 -
                             [f.sub.CO]) C[O.sub.2]
CO oxidation                 CO + 1/2[O.sub.2] [right arrow] C[O.sub.2]
HC oxidation                 [C.sub.12][H.sub.24] + [18O.sub.2] [right
                             arrow] 12 C[O.sub.2] + 12 [H.sub.2]O

(1) The (*) represents the adsorbed N[H.sub.3] on the catalyst site.

TABLE 2 SCR production data exhaust conditions.

       Exhaust      Inlet
Test   flow rate    temperature   Inlet N[O.sub.x]   Inlet N[O.sub.2]
[-]    [kg/min]     [C]           [ppm]              [ppm]

1       4.4         219            648               175
2       6.3         238            279               103
3       9.7         307            291                90
4       9.7         327            342               157
5       7.8         354            552               226
6       6.2         352           1730               692
8      16.4         447            542                98

TABLE 3 Aftertreatment system component specifications.

Substrate                          DOC_ (*)     SCR-A_ (*)   SCRF[R]
Material                           Cordierite   Cordierite   Cordierite

Diameter (in)                         9.0         10.5         10.5
Length (in)                           4           12           12
Cell geometry                       Square       Square       Square
Total volume (L)                      4.17        17.04        17.04
Open volume (L)                       3.5         14.4         10.2
Cell density/[in.sup.2]             400          400          200
Cell width (mil)                     46           46           55
Filtration area ([in.sup.2])        NA           NA           11,370
Open frontal area ([in.sup.2])       26.92        73.29        25.9
Channel wall thickness (mil)          4            4            16
Wall density (g/[cm.sup.3])           0.91         0.91         -
Mean pore size ([micro]m)           NA           NA           16
Number of in cells               25,447       34,636       8659
Weight of substrate + can          5155         7044         18,140
at room temp. (g)

(*) Production components on 2013 Cummins ISB Diesel Engine.

TABLE 4 SCR-F + SCR engine test condition data.

       Exhaust flow   Inlet
Test   rate           temperature   Inlet N[O.sub.x]   Inlet N[O.sub.2]
[-]    [kg/min]       [C]           [ppm]              [ppm]

A       5.6           267            590               215
C       6.9           339            689               290
E       7.1           342           1450               584
B       3.7           256           1580               758
D      12.5           366            450               161
1       5.2           203            625               182

TABLE 5 List of parameters in the SCR model.

Parameter                                     Symbol

Pre-exponential of first site adsorption      [A.sub.ads,1]
Pre-exponential of first site desorption      [A.sub.des,1]
Activation energy of first site desorption    [E.sub.des,1]
Pre-exponential of second site adsorption     [A.sub.ads,2]
Pre-exponential of second site desorption     [A.sub.des,2]
Activation energy of second site desorption   [E.sub.des,2]
Pre-exponential of standard SCR               [A.sub.std,scr]
Activation energy of standard SCR             [E.sub.std,scr]
Pre-exponential of fast SCR                   [A.sub.fst,scr]
Activation energy of fast SCR                 [E.sub.fst,scr]
Pre-exponential of slow SCR                   [A.sub.slo,scr]
Activation energy of slow SCR                 [E.sub.slo,scr]
Pre-exponential of N[H.sub.3] oxidation       [A.sub.NH3,oxid]
Activation energy of N[H.sub.3] oxidation     [E.sub.NH3,oxid]
Pre-exponential of [N.sub.2]O formation       [A.sub.N2O]
Activation energy of [N.sub.2]O formation     [E.sub.N2O]

TABLE 6 SCR kinetics obtained from calibration of 1D SCR model with
baseline experimental data and SCR and PM oxidation kinetics for the 2D
SCR-F model.

Parameter           SCR         SCRF[R]    Units

Omega               0.043       0.180      kmol/[m.sup.3]
Omega 2             0.042       0.092      kmol/[m.sup.3]
[A.sub.ads,1]       1.1         9.00E+03   [m.sup.3]/gmol.s
[E.sub.ads,1]       -10.2       6.00E+01   kJ/gmol
[A.sub.des,1]       4.9E+04     1.91E+02   1/s
[E.sub.des,1]       67.5        1.83E+02   kJ/gmol
[A.sub.ads,2]       2.11E+01    1.14E+03   [m.sup.3]/gmol.s
[E.sub.ads,2]       -7.6        1.24E+03   kJ/gmol
[A.sub.des,2]       9.58E+05    9.74E+06   1/s
[E.sub.des,2]       72.4        8.54E+01   kJ/gmol
[A.sub.std,1]       1E+05       2.50E+08   [m.sup.3]/gmol.s
[E.sub.std,1]       82.3        6.76E+01   kJ/gmol
[A.sub.std,2]       6.17E+06    2.15E+09   [m.sup.3]/gmol.s
[E.sub.std,2]       68.4        4.58E+01   kJ/gmol
[A.sub.fst]         1.01E+08    2.69E+09   [m.sup.6]/[gmol.sup.2].s
[E.sub.fst]         41.2        1.08E+02   kJ/gmol
[A.sub.slo]         7.13E+09    3.45E+13   [m.sup.3]/gmol.s
[E.sub.slo]         109         2.00E+02   kJ/gmol
[A.sub.oxid]        2.33E+05    1.83E+02   1/s
[E.sub.oxid]        91.1        1.14E+03   kJ/gmol
[A.sub.NO2,oxid]    N/A         605.9      m/K-s
[E.sub.NO2,oxid]    N/A         116.5      kJ/gmol
[A.sub.O2,oxid]     N/A         486.9      m/K-s
[E.sub.O2,oxid]     N/A         197.8      kJ/gmol

TABLE 7 N[O.sub.x] reduction efficiency of SCR (model vs experimental)
at ANR = 1.0.

         N[O.sub.x] conversion   N[O.sub.x] conversion
Test     experimental            model
[-]      [%]                     [%]

Test 1   81.4                    83.1
Test 2   82.3                    85.0
Test 3   92.3                    95.6
Test 4   96.3                    95.9
Test 5   96.8                    95.9
Test 6   95.9                    95.5
Test 8   84.1                    81.6

TABLE 8 N[O.sub.2]/N[O.sub.x] at SCR-F inlet, SCR-F wall inlet and SCR
inlet.

                                     SCR-F wall
       SCR-F inlet                   inlet N[O.sub.2]/N[O.sub.x]
Test   N[O.sub.2]/N[O.sub.x] ratio   ratio
[-]    [-]                           [-]

A      0.44                          0.31
C      0.44                          0.29
E      0.37                          0.25
B      0.48                          0.39
D      0.38                          0.22
1      0.29                          0.28

       SCR inlet
Test   N[O.sub.2]/N[O.sub.x]
[-]    ratio [-]

A      0
C      0
E      0
B      0
D      0
1      0

TABLE 9 N[O.sub.x] conversion efficiency from experimental data for the
SCR-F + SCR system.

                  N[O.sub.x] conversion   N[O.sub.x] conversion
Test              experimental            model

Test 1            97.7                    97.2
Test B            97.4                    97.5
Test A            95.9                    97.5
Test C            97.5                    97.3
Test C with SCR   97.8                    96.4
Test C w/o SCR    96.7                    97.1
Test D            97.1                    97.4
Test E            97.6                    97.9

TABLE A.1 Outlet NO concentration from the SCRF[R] and SCR
(experimental and modeling) studies.

           Outlet SCRF[R]            Outlet SCR
           NO        NO              NO      NO
           exp.      model           exp.    model
Test       (ppm)     (ppm)    Diff   (ppm)   (ppm)   Diff

Test 1      7        12       -5      8      11      -3
Test B      5        10       -5      9       8       1
Test A     18        15        3     24       9      15
Test C      4        12       -8     17       8       9
Test C     25        32       -7     13      25      -8
with SCR
Test C     15        24       -9     22      20      -2
w/o SCR
Test D      3        11       -8     13       7       6
Test E      8         5        3      6       0       6

TABLE A.2 Outlet N[O.sub.2] concentration from the SCRF[R] and SCR
(experimental and modeling) studies.

            Outlet SCRF[R]                Outlet SCR
            N[O.sub.2]  N[O.sub.2]        N[O.sub.2]  N[O.sub.2]
            exp.        model             exp.        model
Test        (ppm)       (ppm)       Diff  (ppm)       (ppm)       Diff

Test 1      0           0           0     0           0           0
Test B      0           0           0     0           0           0
Test A      1           0           1     0           0           0
Test C      0           0           0     0           0           0
Test C      3           0           3     2           0           2
with SCR
Test C w/o  1           0           1     1           0           1
SCR
Test D      0           0           0     0           0           0
Test E      1           0           1     0           0           0

TABLE A.3 Outlet N[H.sub.3] concentration from the SCRF[R] and SCR
(experimental and modeling) studies.

           Outlet SCRF[R]                 Outlet SCR
           N[H.sub.3]  N[H.sub.3]         N[H.sub.3]  N[H.sub.3]
           exp.        model              exp.        model
Test       (ppm)       (ppm)       Diff   (ppm)       (ppm)       Diff

Test 1      1          15          -14     2           8          -6
Test B     10          17           -7     4           9          -5
Test A     30          44          -14    28          25           3
Test C     17          30          -13    18          20          -2
Test C     19           5           14    12           0          12
with SCR
Test C     32          25            7    27          15          12
w/o SCR
Test D     41          37            4    36          30           6
Test E     29          38           -9    32          28           4
COPYRIGHT 2019 SAE International
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2019 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Chundru, Venkata Rajesh; Parker, Gordon G.; Johnson, John H.
Publication:SAE International Journal of Fuels and Lubricants
Date:Jun 1, 2019
Words:13983
Previous Article:Assessment of Hydrotreated Vegetable Oil (HVO) Applicability as an Alternative Marine Fuel Based on Its Performance and Emissions Characteristics.
Next Article:Investigation into the Tribological Properties of Biodiesel-Diesel Fuel Blends Under the Run-in Period Conditions.
Topics:

Terms of use | Privacy policy | Copyright © 2020 Farlex, Inc. | Feedback | For webmasters