The role of neuro-fuzzy modelling as a greening technique, in improving the performance of vehicular spark ignition engine

Abstract

The spark ignition engine, by far, is the largest source of motive power in the world. Therefore, continuous endeavours to improve its performance are needed to save in fuel consumption and reduce cost. The main goal of this paper is to develop a neuro-fuzzy model for fuel Injection Time (IT) in order to design a neuro-fuzzy controller for improving the performance of the spark ignition engine. The obtained results showed that the developed neuro-fuzzy model is capable of predicting the fuel IT with a mean squared error less than 0.0072. Furthermore, the power produced by the neuro-fuzzy controller has higher values of about 15-73% than the power produced by the PID controller used in the basic engine. The BSFC is reduced by about 2-5% compared to the PID controller.