Modeling and Control of the Heavy Duty Automotive SCR Catalyst
Andreas Åberg*, Anders Widd**, Jens Abildskov*, Jakob Kjøbsted Huusom*^
* CAPEC-PROCESS, Department of Chemical and Biochemical Engineering, Technical University of Denmark, Søltofts Plads, Building 229, DK-2800 Kgs. Lyngby, Denmark
** Haldor Topsøe A/S, Nymøllevej 55, 2800 Kgs. Lyngby, Denmark
^ Corresponding author: jkh@kt.dtu.dk
Diesel engine exhaust gases contain amongst other things nitrous gases such as NO and NO2 (together NOx). Reducing the amount of these gases is of great importance due to new legislations, because of their effect on urban air quality [1, 2]. A promising and widely used technology for this is based on Selective Catalytic Reduction (SCR) of the gases, with ammonia as a reducing agent [3, 4]. Challenges with this technology include dosing the right amount of urea for sufficient NOx conversion, while keeping the ammonia slip below the legislative limit. This requires efficient dosing algorithms. To this end, model-based control is a promising strategy. A model based on first principles that can predict the complicated dynamics of the SCR catalyst will be useful both for system understanding, and as a benchmark for new control algorithms, to reduce the required experimental work.
A first principles single channel model of high complexity has been developed and coupled with a kinetic model. A second model of slightly lower complexity has also been developed and coupled with the same kinetic model. The kinetic parameters have been calibrated using steady-state data from reactor tests with bench-scale equipment under isothermal conditions. The model has been validated with transient data from a full-scale monolith and a real engine following the European Transient Cycle (ETC) [5], which shows satisfactory results. Results from the validation can be seen in Fig. 1.
To standardize evaluation of control algorithms for the automotive exhaust gas cleaning system an objective function based on current legislations for the automotive system has been developed. A simplified form of the objective function can be used for on-line optimization in for example a model predictive controller. Three different controllers, a feedforward, a feedback and a combined feedforward/feedback have been implemented and tested with the derived model. The controllers have been evaluated using the objective function.
Figure 1. NOx outlet variations during full scale validation. Full line represents data and dotted line represents model predictions.
References:
[1] A. Fritz, V. Pitchon, The current state of research on automotive lean NOx catalysts, Appl. Catal. B 13 (1997).
[2] R. M. Heck, R. J. Farrauto, Automotive exhaust catalysis, Appl. Catal. A 221 (2001).
[3] M. Koebel, M. Elsener, M. Kleeman, Urea-SCR: a promising techniquie to reduce NOx emissions from automotive diesel engines, Catal. Today 59 (2000).
[4] P. Gabrielsson, Urea-SCR in automotive applications, Top. Catal. 28 1-4 (2004).
[5] DieselNet hompage. [Online]. Available: http://www.dieselnet.com/standards/cycles/etc.php
See more of this Group/Topical: Environmental Division