Abstract
Carbon dioxide emissions from vehicles are a particular focus and challenge for automotive designers and manufacturers due to increasingly stringent engine emissions legislation. In addition to the potential environmental impacts, the rate of carbon dioxide production is strongly indicative of the efficiency and therefore fuel economy of an engine at its different operating conditions.In this paper, a neural network model is developed in order to predict the carbon dioxide production rate from a number of engine variables including engine speed, torque, temperature and parameters controlling fuel injection timing. The model structure accurately predicts the rate of carbon dioxide production and has applications in future efficiency and emissions optimisation during engine design and also in online engine control. © 2012 University of Bath.