Effective torque modeling of a gasoline-ethanol engine for multivariable adaptive control design

An effective torque model has been developed, with application in the domain of flex-fuel engines that operate using gasoline-ethanol fuel mixtures. The study initiates by establishing a mono-fuel model and subsequently investigates the impact of fuel composition on combustion efficiency, friction losses, gaseous pumping, and optimal ignition timing. The resulting bi-fuel effective torque model operates in real time, thereby enabling control design through the utilization of a programmable engine control unit (ECU). To validate the model, simulated and measured effective torque curves have been employed to showcase the model's ability to accurately replicate the behavior exhibited by real systems. The model achieves a mean error of 6.3% and a standard deviation of 6.04% for thermal efficiency estimation, demonstrating its effectiveness despite the inherent uncertainties associated with combustion in automotive engines and the required simplifications for real-time application. The methodology employed in this study involves gathering data from a bench dynamometer equipped with a flex-fuel engine. These data are utilized to develop a mean value effective torque model, intended for integration within the ECU as an estimator, enabling the project of torque controllers. Finally, an adaptive multivariable PI controller is presented, which acts on throttle opening, injection duration, and ignition angle to regulate effective torque. The stability is proven and, as a result, the controlled signal exhibited good reference tracking and an appropriate settling time.