Effects of engine parameters on ionization current and modeling of excess air coefficient by artificial neural network

Abstract

This study investigates the effects of engine speed, load, ignition timing and excess air coefficient on the ionization current and presents an artificial neural network model to predict the in-cylinder air-fuel ratio by using data of the ionization current. A secondary spark plug was used as an ionization current sensor. Experimental studies were conducted on a spark-ignition engine at variable speed, load, ignition timing, and excess air coefficient. The effects of these parameters on the ionization current were investigated individually. For modeling of the excess air coefficient, an artificial neural network model was developed with the experimental results. The network was trained with Levenberg-Marquardt and Scaled Conjugate Gradient training algorithms. Performance of the network was measured by comparing the predictions with the remaining experimental results. The excess air coefficient can be predicted with the network with a coefficient of determination of 0.99508. This study shows, the ionization current signal can be used to predict the in-cylinder excess air coefficient as a feasible alternative to the production air-fuel ratio sensors.