Modelling of modern automotive petrol engine performance using Support Vector Machines

Modern automotive petrol engine performance is significantly affected by effective tune-up. Current practice of engine tune-up relies on the experience of the automotive engineer, and tune-up is usually done by trial-and-error method and then the vehicle engine is run on the dynamometer to show the actual engine performance. Obviously the current practice involves a large amount of time and money, and then may even fail to tune up the engine optimally because a formal performance model of the engine has not been determined yet. With an emerging technique, Support Vector Machines (SVM), the approximate performance model of a petrol vehicle engine can be determined by training the sample engine performance data acquired from the dynamometer. The number of dynamometer tests for an engine tune-up can therefore be reduced because the estimated engine performance model can replace the dynamometer tests to a certain extent. In this paper, the construction, validation and accuracy of the model are discussed. The study showed that the predicted results agree well with the actual test results. To illustrate the significance of the SVM methodology, the results were also compared with that regressed using multilayer feedforward neural networks.