Modeling of NH3 Storage in Vanadia-Based SCR Catalyst for Urea-Dosing Control in a Dieselelectric Hybrid Car
The Urea-SCR system is responsible for reducing the engine-out NOx using NH3, which is injected into the system as Diesel Exhaust Fluid or DEF. An optimum amount of DEF needs to be injected into the system to maximize the NOx reduction and minimize the NH3 slip. Since the amount of NH3 stored into the catalyst is an important part of the system dynamics, an NH3 storage-based urea-dosing control must be developed to inject optimum quantity of DEF. However, it is not possible to measure the amount of NH3 stored in the catalyst. So, the goal of this thesis was to estimate it. This goal was broken down into two sub-objectives of parameter identification and full-state observer design. In this thesis, simulations were developed, based on lumped-parameter zero dimensional three-state and four-state models, to test the parameter identification methods and observer designs. The steady-state testing protocols from previous research were also modified such that the experiments for parameter identification become repeatable. A parameter identification method was then formulated and tested on the simulations and the experiment data. The identified parameters showed reasonable fit with the measurements of emitted NOx and NH3 slip. These fits further improve when the data is fed to the full-state feedback-based observer designed in this work. Though the storage estimates of the observer could not be verified for the real data, it was shown, using the simulations, that the reference tracking performance of a simple proportional controller can be maintained by replacing the system's storage by the observer estimates in the controller's feedback.
Meckl, Purdue University.
Automotive engineering|Mechanical engineering
Off-Campus Purdue Users:
To access this dissertation, please log in to our