Model-Based On-Board Diagnostics for SCR-ASC

Kaushal Kamal Jain, Purdue University

Abstract

Selective Catalytic Reduction (SCR) and Ammonia Slip Catalyst (ASC) are important components of the diesel engine aftertreatment. SCR reduces the engine-out NOx into harmless N2 and H2O using NH3, which is injected into the system as Diesel Exhaust Fluid or DEF. ASC is responsible for oxidizing SCR-out NH3. Thus, SCR-ASC system minimizes tailpipe NOx and NH3 emissions. A major challenge with the SCR-ASC system is degradation or aging of the SCR catalyst with time, which leads to increase in tailpipe NOx emissions. Therefore, it is important for diesel-engine vehicles to be equipped with effective on-board diagnostics (OBD), which can monitor and report catalyst degradation before it degrades beyond acceptable levels. The primary objective of this work is to develop a robust model-based non-intrusive on-board diagnostics algorithm that can monitor catalyst health using commercial NOx sensors under real-world on-road driving conditions. Cummins Inc. has generously provided on-road data for four trucks, and test-cell data for cold Federal Test Procedure (cFTP), hot Federal Test Procedure (hFTP), and Ramped Mode Cycle (RMC) cycles for degreened (DG) and end-of-useful-life (EUL) catalysts. This thesis presents a diagnostics-oriented aging model for combined SCR-ASC system, along with two model-based OBD methods applied to both test-cell data and on-road data from commercial trucks. The key challenge with model development was unavailability of NOx and NH3 measurements between SCR and ASC. Since it would have been very difficult to calibrate both SCR and ASC dynamics without any measurements between SCR and ASC, therefore ASC was modeled using static look-up tables to determine ASC's NH3 conversion efficiency and its selectivity to NOx and N2O as a function of temperature and flow rate. The traditional three-state single-cell ordinary differential equation (ODE) model was used for SCR. Hot FTP was used to calibrate the model. Cold FTP and RMC were used for validation. Results show that the SCR-ASC model can capture the aging signatures in tailpipe NOx, NH3, and N2O reasonably well for cFTP, hFTP, and RMC cycles in the test-cell data. After slight re-calibration and combining with a simple model for commercial NOx sensor's cross-sensitivity to NH3, the model works reasonably well for on-road data from commercial trucks. Two model-based on-board diagnostic (OBD) methods are presented with enable conditions designed to detect operating conditions suitable for detecting aging signatures, while minimizing false positives and false negatives. It was demonstrated that the enable conditions increase the robustness of the OBD methods to model inaccuracy, uncertainty in initial NH3 storage, and NOx sensor's cross-sensitivity to NH3. The first OBD method does a binary classification at each sample point to label it as either degreened or EUL. This OBD method is applied to both test-cell and real-world truck data with commercial NOx sensors. Results on test-cell data showed that this method is capable of correctly identifying the aging levels of degreened and EUL catalysts with zero false positives and very few false negatives. The method was also shown to be robust to NOx sensor's cross-sensitivity to NH3 and measurement noise when the tailpipe NOx and NH3 signals in test-cell data were combined with Gaussian noise to simulate worst-case cross-sensitivity and measurement noise. Results on truck data show encouraging trends between relative degradation level and the number of miles on the four trucks. The trends reported by the method on truck data were shown to be robust to uncertainty in the initial value of NH3 storage. A drawback with this method was that very few sample points were selected from the test-cell data after applying the enable conditions, which demonstrate the challenge with designing model-based enable conditions that are robust to false positives and false negatives but still lead to good In-use monitoring performance ratio (IUMPR). Unlike the first method, the second OBD method assigns a non-binary value to each sample point, which is proportional to the probability of that point belonging to a degreened or an EUL catalyst. This method uses a stochastic version of the proposed SCR-ASC model, which is derived using a simplified version of the Bayesian approach for model calibration. This method results in a much better IUMPR than the first one and can still correctly classify the degreened and EUL catalysts for all three cycles in the test-cell data. Even though this work presents a preliminary implementation of the stochastic OBD method, the detailed framework presented in this thesis along with a complete set of enable conditions lays a strong foundation for developing a more detailed version of the method in future based on a rigorous implementation of the Bayesian approach.

Degree

Ph.D.

Advisors

Meckl, Purdue University.

Subject Area

Automotive engineering|Chemical engineering|Mechanical engineering

Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server
.

Share

COinS