A machine learning approach to modeling and identification of automotive three-way catalytic converters

The working of three-way catalytic converters (TWC's) is based on chemical reactions whose rates are nonlinear functions of temperature and reactant c oncentrations all along the device. Unfortunately, the choice of suitable e xpressions and the tuning of their parameters is particularly difficult in dynamic conditions. In this paper we introduce a hybrid modeling technique which allows us to preserve the most important features of an accurate dist ributed parameter TWC model, while it circumvents both the structural and t he parameter uncertainties of "classical" reaction kinetics models, and sav es computational time. In particular, we compute the rates within the TWC d ynamic model by a neural network which, thus, becomes a static nonlinear co mponent of a larger dynamic system. A purposely designed genetic algorithm, in conjunction with a fast ad hoc partial differential equation integratio n procedure, allows us to train the neural network, embedded in the whole m odel structure, using currently available measurement data and without comp uting gradient information.