Groupe PSA and IFPEN have jointly developed a new method to model engine emissions
It performs in many specific situations, and provides opportunities for design and set-up steps. A relevant case, developed here, is an engine-out NOx emissions virtual sensor for diesel applications. The model, based on a neural network and empirical correlations, deals with all real driving conditions and lifetime constraints, like ambient air temperature, humidity and altitude variations. Its development and validation process is based on an extensive series of tests.