In English

Calibration and Validation of a 1D model for Exhaust Aftertreatment System in Heavy Duty Diesel Engines

Hongda Cen
Göteborg : Chalmers tekniska högskola, 2018. Examensarbete - Institutionen för mekanik och maritima vetenskaper; 2017:81, 2018.
[Examensarbete på avancerad nivå]

Diesel engines are faced with more stringent legislation for higher fuel efficiency and lower emissions. New technologies have been designed to meet these needs. Volvo Penta has started a project called virtual test cell for increasing the effectiveness of verification and validation process. The goal is to reduce the testing and calibration time in physical test cells and increase the quality of products by moving part of activities to the virtual test cell. A model in the virtual test cell should be fast and have a shorter lead time. The current thesis was performed as a part of the Virtual Calibration project in collaboration with Chalmers University of Technology. It aims to refine an existing 1D model for exhaust gas aftertreatment system for Volvo Penta’s configuration. One of the key objectives of this thesis is to carry out the calibration work on a 1D Diesel oxidation catalyst (DOC) and Diesel particulate filter (DPF) chemical model in MATLAB/ Simulink. The model predicts temperature, pressure drop, soot loading and concentrations of NOx, CO and HC. The work includes tuning the reaction rate parameters such as activation energies and pre-exponent factors in the purpose of minimizing the difference between experimental and simulation results. The other objective is to validate the model to evaluate if a 1D model is good enough to be used in a virtual test rig. The model was calibrated to a steady state cycle called part load map (PLM). Temperatures and mass flow were high in the beginning and going down to lower as time elapsed. The result shows that the model accuracy is improved after the calibration. The temperature deviation becomes larger at high torque, and the errors of CO and HC concentrations are higher at a lower temperature. The residual of NO concentrations is mainly due to the temperature deviation and unknown initial soot loading in the DPF. The error of pressure drops across the DPF also comes from the unknown initial soot loading in the DPF. The model was then validated to a nonroad transient cycle (NRTC). The model errors become quite high at the transient points. In the future, further calibration against transient data is needed to fully calibrate the model.

Nyckelord: Calibration, Validation, Diesel oxidation catalyst, Diesel particulate filter, Kinetic modelling, 1D simulation, MATLAB Simulink, Parameter estimation

Publikationen registrerades 2019-03-27. Den ändrades senast 2019-03-27

CPL ID: 256605

Detta är en tjänst från Chalmers bibliotek