Kinetic parameter estimation by standard optimization methods in catalytic converter modeling

Abstract

The application of mathematical models to the prediction of the performance of automotive catalytic converters is gaining increasing interest, both for gasoline and diesel engined-vehicles. This article addresses converter modeling in the transient state under realistic experimental conditions. The model employed in this study relies on Langmuir-Hinshelwood kinetics, and a number of apparent kinetic parameters must be tuned to match the behavior of each different catalyst formulation. The previously applied procedure of manually tuning kinetics parameters requires significant manpower. This article presents a methodology for kinetic parameter estimation that is based on standard optimization methods. The methodology is being applied in the exploitation of synthetic gas experiments and legislated driving cycle tests and the assessment of the quality of information contained in the test results. Although the optimization technique employed for parameter estimation is well known, the development of the specific parameter estimation methodology that employs the results of the available types of experiments is novel and required significant development. Application of this refined tuning methodology increases the quality and reliability of prediction and also greatly reduces the required manpower, which is important in the specific engineering design process. The parameter estimation procedure is applied to the example of modeling of a diesel catalytic converter with adsorption capabilities, based on laboratory experiments and vehicle driving cycle tests.