부산대학교 도서관

Two very important areas in automotive engine design are low exhaust emissions and better fuel economy. As regulations become increasingly stringent manufacturers of automotive engines are looking for ways to reduce the emission of pollutant gases from the exhaust. This paper describes a technique under development known as Spark Voltage Characterization, which involves capturing the time varying spark voltage waveform and analysing it using a neural network. The neural network associates spark voltage vectors and other engine parameters of interest with specific air-fuel ratio values. The experimental work done so far has proved that the sensor is able to accurately determine the air-fuel ratio. There are factors other than the air-fuel ratio that affect the spark voltage characteristics. The work presented in this paper aimed to investigate the effect of varying engine block temperature on the proposed sensor method. Data was captured at different engine block temperatures and air-fuel ratio settings and used to train and test the neural network. The neural network analysis was conducted to test its ability to determine the air-fuel ratio using data for both trained and untrained engine block temperature settings.