Control-oriented modelling of combustion phasing for a fuel-flexible spark-ignited engine with variable valve timing

Abstract

In an effort to reduce dependence on petroleum-based fuels and increase engine efficiency, fuel-flexible engines with advanced technologies, including variable valve timing, are being developed. Fuel-flexible spark-ignition engines permit the increased use of ethanol–gasoline blends. Ethanol, an alternative to petroleum-based gasoline, is a renewable fuel, which has the added advantage of improving performance in operating regions that are typically knock limited due to the higher octane rating of ethanol. Furthermore, many modern engines are also being equipped with variable valve timing, a technology that can increase engine efficiency by reducing pumping losses. Through control of valve timings, particularly the amount of positive valve overlap, the quantity of burned gas in the engine cylinder can be altered, eliminating the need for intake throttling at many operating points. However, the presence of elevated levels of in-cylinder burned gas and ethanol fuel can have a significant impact on the combustion timing, such that capturing these effects is essential if the combustion phasing is to be properly controlled. This paper outlines a physically based model capable of capturing the impact of the ethanol blend ratio, burned gas fraction, spark timing and operating conditions on combustion timing. Since efficiency is typically tied to an optimal CA50 (crank angle when 50% of fuel is burned), this model is designed to provide accurate estimates of CA50 that can be used for real-time control efforts – allowing the CA50 to be adjusted to its optimal value despite changes in ethanol blend and burned gas fraction, as well as the variations in engine thermodynamic conditions that may occur during transients. The proposed control-oriented model was extensively validated at over 500 points across the engine operating range for four blends of gasoline and ethanol. Furthermore, the model was utilized to determine the impact of ethanol blend and burned gas fraction on the CA50, as well as their impact on the optimal spark timing. This study indicated that the burned gas fraction could change the optimal spark timing by over 20° at some operating conditions and that ethanol content could further affect the optimal spark timing by up to 6°. Leveraging the model in this manner provides direct evidence that accounting for the impact of these two inputs is critical for proper spark-ignition timing control.