Modelling and Operation of Diesel Engine Exhaust Gas Cleaning Systems

Andreas Åberg

Research output: Book/ReportPh.D. thesisResearch

1321 Downloads (Pure)

Abstract

Diesel engine exhaust gases contain several harmful substances. The main pollutants are carbon monoxide (CO), hydrocarbons (HC), particulate matter (PM), and nitrous gases such as nitrogen oxide (NO) and nitrogen dioxide (NO2) (together NOx). Reducing the emission of these pollutants is of great importance due to their effect on urban air quality, and because of new legislation. In modern heavy-duty applications, the exhaust gases are typically treated with four different catalysts: a Diesel Oxidation Catalyst (DOC) which oxidises HC and CO into H2O and CO2, and NO into NO2, a Diesel Particulate Filter (DPF) which filters PM, a Selective Catalytic Reduction (SCR) catalyst which removes NO and NO2 through reaction with NH3, and an Ammonia Slip Catalyst (ASC) which removes excess ammonia (NH3) before the gases are released to the atmosphere.
SCR is a widely used technology to reduce NOx to N2. Challenges with this technology include dosing the appropriate amount of urea to reach sufficient NOx conversion, while at the same time keeping NH3- slip from the exhaust system below the legislation. This requires efficient control algorithms.
The focus of this thesis is modelling and control of the SCR catalyst. A single channel model for a heavy-duty SCR catalyst was derived based on first principles. The model considered heat and mass transfer between the channel gas phase and the wash coat phase. Four simplified models were derived, with simplifications related to mass and heat transfer. The model parameters were estimated using bench-scale monolith isothermal data. Validation was done by simulating the out-put from a full-scale SCR monolith that was treating real engine gases from the European Transient Cycle (ETC). Results showed that the models were successfully calibrated, and that some of the models could predict the ETC output satisfactorily. The models’ predictive capabilities were investigated in relation to the simplifications, and results showed that the simplifications related to mass transfer resulted in the smallest information loss.
A methodology to analyse the NOx-NH3 trade-off for different urea dosing con-trollers was developed, and applied to P, PI, PD, and PID controllers, both with and without Ammonia-NOx-Ratio (ANR) based feedforward. Simulation results showed that the PI controller with feedforward had the best NOx-NH3 trade-off, and that feedforward coupled with feedback outperformed the other control structures. The results were experimentally verified by implementing the tested controllers on a full-scale engine setup, and the results showed that coupling feedback with ANR based feedforward was yielding better performance. The PD controller showed good performance in the experimental validation.
Finally, a methodology for creating a modular simulation tool was developed. The methodology goes through the steps that are required to integrate individual models so that they can be used for the tool. The methodology is demonstrated by applying it to four models from literature, and simulating the system.
Original languageEnglish
PublisherDanmarks Tekniske Universitet (DTU)
Number of pages181
Publication statusPublished - 2017

Cite this

Åberg, A. (2017). Modelling and Operation of Diesel Engine Exhaust Gas Cleaning Systems. Danmarks Tekniske Universitet (DTU).
Åberg, Andreas. / Modelling and Operation of Diesel Engine Exhaust Gas Cleaning Systems. Danmarks Tekniske Universitet (DTU), 2017. 181 p.
@phdthesis{5084dd1582ea4216990738b0974e32f1,
title = "Modelling and Operation of Diesel Engine Exhaust Gas Cleaning Systems",
abstract = "Diesel engine exhaust gases contain several harmful substances. The main pollutants are carbon monoxide (CO), hydrocarbons (HC), particulate matter (PM), and nitrous gases such as nitrogen oxide (NO) and nitrogen dioxide (NO2) (together NOx). Reducing the emission of these pollutants is of great importance due to their effect on urban air quality, and because of new legislation. In modern heavy-duty applications, the exhaust gases are typically treated with four different catalysts: a Diesel Oxidation Catalyst (DOC) which oxidises HC and CO into H2O and CO2, and NO into NO2, a Diesel Particulate Filter (DPF) which filters PM, a Selective Catalytic Reduction (SCR) catalyst which removes NO and NO2 through reaction with NH3, and an Ammonia Slip Catalyst (ASC) which removes excess ammonia (NH3) before the gases are released to the atmosphere.SCR is a widely used technology to reduce NOx to N2. Challenges with this technology include dosing the appropriate amount of urea to reach sufficient NOx conversion, while at the same time keeping NH3- slip from the exhaust system below the legislation. This requires efficient control algorithms.The focus of this thesis is modelling and control of the SCR catalyst. A single channel model for a heavy-duty SCR catalyst was derived based on first principles. The model considered heat and mass transfer between the channel gas phase and the wash coat phase. Four simplified models were derived, with simplifications related to mass and heat transfer. The model parameters were estimated using bench-scale monolith isothermal data. Validation was done by simulating the out-put from a full-scale SCR monolith that was treating real engine gases from the European Transient Cycle (ETC). Results showed that the models were successfully calibrated, and that some of the models could predict the ETC output satisfactorily. The models’ predictive capabilities were investigated in relation to the simplifications, and results showed that the simplifications related to mass transfer resulted in the smallest information loss.A methodology to analyse the NOx-NH3 trade-off for different urea dosing con-trollers was developed, and applied to P, PI, PD, and PID controllers, both with and without Ammonia-NOx-Ratio (ANR) based feedforward. Simulation results showed that the PI controller with feedforward had the best NOx-NH3 trade-off, and that feedforward coupled with feedback outperformed the other control structures. The results were experimentally verified by implementing the tested controllers on a full-scale engine setup, and the results showed that coupling feedback with ANR based feedforward was yielding better performance. The PD controller showed good performance in the experimental validation.Finally, a methodology for creating a modular simulation tool was developed. The methodology goes through the steps that are required to integrate individual models so that they can be used for the tool. The methodology is demonstrated by applying it to four models from literature, and simulating the system.",
author = "Andreas {\AA}berg",
year = "2017",
language = "English",
publisher = "Danmarks Tekniske Universitet (DTU)",
address = "Denmark",

}

Åberg, A 2017, Modelling and Operation of Diesel Engine Exhaust Gas Cleaning Systems. Danmarks Tekniske Universitet (DTU).

Modelling and Operation of Diesel Engine Exhaust Gas Cleaning Systems. / Åberg, Andreas.

Danmarks Tekniske Universitet (DTU), 2017. 181 p.

Research output: Book/ReportPh.D. thesisResearch

TY - BOOK

T1 - Modelling and Operation of Diesel Engine Exhaust Gas Cleaning Systems

AU - Åberg, Andreas

PY - 2017

Y1 - 2017

N2 - Diesel engine exhaust gases contain several harmful substances. The main pollutants are carbon monoxide (CO), hydrocarbons (HC), particulate matter (PM), and nitrous gases such as nitrogen oxide (NO) and nitrogen dioxide (NO2) (together NOx). Reducing the emission of these pollutants is of great importance due to their effect on urban air quality, and because of new legislation. In modern heavy-duty applications, the exhaust gases are typically treated with four different catalysts: a Diesel Oxidation Catalyst (DOC) which oxidises HC and CO into H2O and CO2, and NO into NO2, a Diesel Particulate Filter (DPF) which filters PM, a Selective Catalytic Reduction (SCR) catalyst which removes NO and NO2 through reaction with NH3, and an Ammonia Slip Catalyst (ASC) which removes excess ammonia (NH3) before the gases are released to the atmosphere.SCR is a widely used technology to reduce NOx to N2. Challenges with this technology include dosing the appropriate amount of urea to reach sufficient NOx conversion, while at the same time keeping NH3- slip from the exhaust system below the legislation. This requires efficient control algorithms.The focus of this thesis is modelling and control of the SCR catalyst. A single channel model for a heavy-duty SCR catalyst was derived based on first principles. The model considered heat and mass transfer between the channel gas phase and the wash coat phase. Four simplified models were derived, with simplifications related to mass and heat transfer. The model parameters were estimated using bench-scale monolith isothermal data. Validation was done by simulating the out-put from a full-scale SCR monolith that was treating real engine gases from the European Transient Cycle (ETC). Results showed that the models were successfully calibrated, and that some of the models could predict the ETC output satisfactorily. The models’ predictive capabilities were investigated in relation to the simplifications, and results showed that the simplifications related to mass transfer resulted in the smallest information loss.A methodology to analyse the NOx-NH3 trade-off for different urea dosing con-trollers was developed, and applied to P, PI, PD, and PID controllers, both with and without Ammonia-NOx-Ratio (ANR) based feedforward. Simulation results showed that the PI controller with feedforward had the best NOx-NH3 trade-off, and that feedforward coupled with feedback outperformed the other control structures. The results were experimentally verified by implementing the tested controllers on a full-scale engine setup, and the results showed that coupling feedback with ANR based feedforward was yielding better performance. The PD controller showed good performance in the experimental validation.Finally, a methodology for creating a modular simulation tool was developed. The methodology goes through the steps that are required to integrate individual models so that they can be used for the tool. The methodology is demonstrated by applying it to four models from literature, and simulating the system.

AB - Diesel engine exhaust gases contain several harmful substances. The main pollutants are carbon monoxide (CO), hydrocarbons (HC), particulate matter (PM), and nitrous gases such as nitrogen oxide (NO) and nitrogen dioxide (NO2) (together NOx). Reducing the emission of these pollutants is of great importance due to their effect on urban air quality, and because of new legislation. In modern heavy-duty applications, the exhaust gases are typically treated with four different catalysts: a Diesel Oxidation Catalyst (DOC) which oxidises HC and CO into H2O and CO2, and NO into NO2, a Diesel Particulate Filter (DPF) which filters PM, a Selective Catalytic Reduction (SCR) catalyst which removes NO and NO2 through reaction with NH3, and an Ammonia Slip Catalyst (ASC) which removes excess ammonia (NH3) before the gases are released to the atmosphere.SCR is a widely used technology to reduce NOx to N2. Challenges with this technology include dosing the appropriate amount of urea to reach sufficient NOx conversion, while at the same time keeping NH3- slip from the exhaust system below the legislation. This requires efficient control algorithms.The focus of this thesis is modelling and control of the SCR catalyst. A single channel model for a heavy-duty SCR catalyst was derived based on first principles. The model considered heat and mass transfer between the channel gas phase and the wash coat phase. Four simplified models were derived, with simplifications related to mass and heat transfer. The model parameters were estimated using bench-scale monolith isothermal data. Validation was done by simulating the out-put from a full-scale SCR monolith that was treating real engine gases from the European Transient Cycle (ETC). Results showed that the models were successfully calibrated, and that some of the models could predict the ETC output satisfactorily. The models’ predictive capabilities were investigated in relation to the simplifications, and results showed that the simplifications related to mass transfer resulted in the smallest information loss.A methodology to analyse the NOx-NH3 trade-off for different urea dosing con-trollers was developed, and applied to P, PI, PD, and PID controllers, both with and without Ammonia-NOx-Ratio (ANR) based feedforward. Simulation results showed that the PI controller with feedforward had the best NOx-NH3 trade-off, and that feedforward coupled with feedback outperformed the other control structures. The results were experimentally verified by implementing the tested controllers on a full-scale engine setup, and the results showed that coupling feedback with ANR based feedforward was yielding better performance. The PD controller showed good performance in the experimental validation.Finally, a methodology for creating a modular simulation tool was developed. The methodology goes through the steps that are required to integrate individual models so that they can be used for the tool. The methodology is demonstrated by applying it to four models from literature, and simulating the system.

M3 - Ph.D. thesis

BT - Modelling and Operation of Diesel Engine Exhaust Gas Cleaning Systems

PB - Danmarks Tekniske Universitet (DTU)

ER -

Åberg A. Modelling and Operation of Diesel Engine Exhaust Gas Cleaning Systems. Danmarks Tekniske Universitet (DTU), 2017. 181 p.