A Model Based Design Approach For Knock Control in Internal Combustion Engines Using Machine Learning

This work presents an alternative approach to address knock occurrence control modeling by delivering an alternative logistic regression model that takes in consideration knock intensity's dependence on factors such as in-cylinder pressure, mixture temperature and engine speed. The model is developed based on a 1.4 liter spark ignited engine model, also part of this work. The alternative model is validated through frequentist inference guidelines and simulation results show its fast tracking performance on estimating knock probabilities for different operating conditions.