Design and experimental validation of three way catalyst age estimator and Fisher information analysis for optimal sensor selection. (July 2021)

• Record Type: Journal Article
• Authors: Gelmini, S.; Hoffmann, M.A.; Onori, S.
• Is Part Of: Control engineering practice. Volume 112(2021)
• Publication Date: 2021-07
• Subjects: Three-way catalyst -- Extended Kalman Filter -- Fisher information analysis -- Rapid control prototyping; Automatic control -- Periodicals; 629.89
• DOI: 10.1016/j.conengprac.2021.104805
• Languages: English
• ISSNs: 0967-0661

Abstract

Abstract: Three way catalysts (TWCs) are installed downstream of internal combustion engines to mitigate the engine out pollutants generated from combustion. Accurate engine control systems are needed to achieve high conversion efficiency during vehicle operation and optimize TWC inputs. Accounting for TWC internal dynamics within the exhaust emission control design is key to enable improved engine and catalyst performance. However, since internal TWC dynamics cannot be directly measured, modeling and estimation of the converter's dynamics are essential for the development of high performing exhaust emission control strategies. In this paper, a Dual Extended Kalman Filter (dEKF) is designed and experimentally validated for the estimation of both the TWC oxygen storage level, ϕ, and age-dependent oxygen storage capacity, O S C . The estimator design is based on an experimentally validated physics-based TWC model developed by the authors in a previous work. Observer validation and performance evaluation are conducted over transient drive cycles in a chassis dynamometer. Moreover, observability properties and related estimation performance are analyzed using different sensor technologies. It is found out that using switch-type oxygen sensors in place of wideband oxygen sensors leads to a loss of observability within the model, preventing real-time OSC determination and, therefore, aging estimation. A Fisher information quantity study is developed and presented that provides quantitative guidelines for the optimization of TWC sensor design. Finally, the observer is tested on real hardware via rapid control prototyping.