Combined Experimental and Numerical Analysis of Knock in Spark Ignition Engines

Abstract

A detailed analysis of knocking event can help improving engine performance and diagnosis strategies. The paper aim is a better understanding of the phenomena involved in knocking combustion through the combination of CFD and signals analysis tools. CFD simulations have been used in order to reproduce knock effect on the in-cylinder pressure trace. In fact, the in-cylinder pressure signal holds information about waves propagation and heat losses: for the sake of the diagnosis it is important to relate knock severity to knock indexes values. For this purpose, a CFD model has been implemented, able to predict the combustion evolution with respect to Spark Advance, from non-knocking up to heavy knocking condition. The CFD model validation phase is crucial for a correct representation of both regular and knocking combustions: the operation has been carried out by means of an accurate statistical analysis of experimental in-cylinder pressure data. The simulation results allow relating the combustion characteristics to their effect on the in-cylinder pressure signal. It is then possible to highlight critical issues regarding typical knock detection methodologies, while disclosing novel approaches. One of the main results is the validation of knock detection strategies based on the low-frequency content of the pressure signal. These strategies can be used together with standard high-frequency based techniques in order to improve the detection robustness.