Torque-based Control Strategies Development For Spark Ignition Engines

Torque management system uses torque as the standard interface, benefits infor-mation interchange between di?erent subsystems. Also, the calibration process is sim-plified and module management can be applied in a torque-based control system, whichhas attracted more and more widely attention. Motivated by the 2008 benchmark prob-lem proposed by the Society of Instrument and Control Engineers (SICE), this dis-sertation gives a systemic research focusing on torque-based engine control for SparkIgnition (SI) engines using modern control technology.A torque generation model is first proposed based on the traditional mean value SIengine model. The e?ect on the produced torque and the identification methods of themodel parameters are investigated. The torque generation model is further simplified toreduce the calibration work. Based on the proposed torque generation model, open loopand closed loop torque estimation methods are introduced. The closed loop estimationcan use the simplified torque generation model since on-line engine speed feedbackmodification is applied.Model-based torque control strategies are presented under di?erent cases that thefeedback torque is obtained by open loop estimation or closed loop observer or directmeasurement. The torque controllers use the same architecture which consists of a non-linear model forward term and an on-line PI feedback term. The model forward termis introduced to provide the static input and compensate the nonlinearity of the system.The remaining linear error is modified by the PI feedback term. The given controllersare valid in a large engine operation range and don't require local models and localcontrollers. Since the controller is closed loop, the simplified torque generation modelis enough to be used for feed-forward control.Torque-based concept then is applied to engine (or vehicle) speed control. Thespeed control problem is solved with two parts: torque management decides the en-gine torque demand while torque controller ensures the engine to generate the required output torque. The engine (or vehicle) speed reference is shifted to torque referenceusing optimization theory and the torque reference is realized by torque controller. Itis actually a torque-based optimal engine (or vehicle) speed controller. The simulationor experimental results show that torque-based control can achieve better performancecompared to conventional controllers.Simulation or experimental results have verified the validity of the proposed es-timation methods and controllers. Current research is useful to realize model-basedtorque management and torque control.