Abstract
This thesis deals with advanced modelling and control of energy efficient automotive drivetrains with torque-fill capability during gearshifts. The literature discussing automotive transmission technology, in particular manual transmissions, automated manual transmissions (AMTs) and dual clutch transmissions (DCTs), with the respective gearshift control methodologies, is reviewed in detail. To increase the overall drivetrain efficiency and address the problem of AMT torque interruption during the shift transients, a novel transmission layout, named as ‘6+2’ in the remainder of the thesis, is introduced. The ‘6+2’ is based on the hybridisation of an AMT. In this layout the torque-fill functionality is achieved via an embedded electric motor drive, which is connected to the transmission output shaft. The gearshift performance of the ‘6+2’ is analysed and compared to that of the DCT of a case study high performance passenger car, through a high-fidelity model-based method and objective gearshift performance indicators. The gearshift simulation models are validated in terms of longitudinal acceleration profiles, with gearshift test data provided by the industrial sponsors of the project, Oerlikon Graziano (Italy) and Vocis Driveline Controls (UK). The analysis shows that with the current ‘6+2’ prototype full torque-fill is supported for gearshifts of up to 40% of driver torque demand. This thesis proposes a novel optimal gearshift controller for the clutch reengagement phase of the ‘6+2’. The controller is designed with the purpose of simultaneously reducing clutch energy dissipation and providing smooth clutch engagement. The computationally efficient structure of this controller facilitates its real-vehicle implementation. The performance of the controller is evaluated on the gearshift simulator and compared to that of the conventional gearshift controller of the prototype ‘6+2’. Significant reductions (53% on the average) in clutch dissipation energy are achieved with the proposed control structure. A similar optimal controller and a model predictive controller are designed and successfully assessed for the inertia phase of the case study DCT. The objectives of these advanced formulations are tuneable reductions clutch dissipation losses, smooth clutch engagement, and capability of handling variations in the oncoming clutch torque, determined by the driver input on the accelerator pedal.