Investigation of Novel Transmissions for Electric Vehicles

<p dir="ltr">Electric vehicles (EVs) are regarded as the most promising and sustainable option for personal urban transportation over the coming decades due to their minimal environmental pollution, low noise levels, high efficiency, versatility in energy sources, and capability for energy feedback. However, one significant limitation of EVs is their relatively short driving range. Enhancing the powertrain's efficiency is a crucial and effective strategy to extend the driving distance of EVs. A dual-clutch transmission (DCT) in electric vehicles can substantially reduce torque interruption and enhance shifting comfort. Additionally, a two-speed transmission can significantly improve driving performance, lower energy consumption, reduce the size of the electric motor, and offer a balanced trade-off between efficiency and dynamic performance compared to a central electric motor with a single-speed transmission. This thesis investigates the novel two-speed DCT for EVs.</p><p dir="ltr">This thesis begins by examining existing literature on two-speed transmissions, DCT, magneto-rheological fluid (MRF), MRF-based clutches, modelling of hysteresis, torque control in DCT, and gear shifting mechanisms in DCT. It identifies the relevant research gaps within these areas.</p><p dir="ltr">Secondly, this thesis introduces an innovative two-speed DCT featuring two traditional friction clutches and a two-stage planetary gear set. The powertrain models, encompassing the electric motor, the two-speed DCT and vehicle dynamics, are developed using MATLAB/Simulink®. To enhance both dynamic and economic performance, a fuzzy logic-based gear shift schedule, designed to capture the driver's intentions, is implemented. The effectiveness of this proposed gear shift schedule is validated through comparisons with a conventional two-parameter gear shift schedule. Simulation results demonstrate that the dynamic and economic performance of the novel DCT for EVs is significantly improved with the fuzzy logic gear shift schedule.</p><p dir="ltr">Thirdly, this thesis presents an innovative MRF dual-clutch (MRFDC) design for the two-speed transmission in EVs, combining the advantages of dual-clutch transmissions (DCT) and magneto-rheological fluid (MRF) clutches. The MRFDC consists of an internal clutch and an external clutch, both of which can switch between two gears by adjusting the input current through their respective coils. Finite element analysis of the magnetic field is employed to establish the relationship between the input current and magnetic flux density. The output torque model is formulated using the Herschel-Bulkley model to characterize MRF behaviour, thereby determining the correlation between the transmissible torque and the applied input current. The MRFDC model is experimentally validated on a testbed, including separate transmissible torque and response time tests for the internal and external MRF clutches. The test results are consistent with the simulation results, with differences within 2 N·m, suggesting that the MRFDC can be effectively applied in EVs to enhance vehicle performance.</p><p dir="ltr">Fourthly, this study explores the nonlinear hysteresis phenomena in a prototyped MRFDC used in EV transmission systems. It provides a detailed analysis of three commonly used hysteresis models: the Bouc-Wen model and the algebraic model (parametric), and the NARX model (non-parametric). The models are evaluated based on accuracy, fitting time, and stack size. The results show that the NARX model excels in accuracy but requires significantly more memory. The algebraic model is noted for its computational efficiency due to its simple expression, while the Bouc-Wen model falls in an intermediate position for all three metrics. To enhance the classic Bouc-Wen model (CBW), a fractional-order modified Bouc-Wen model (FOMBW) is introduced, incorporating polynomial input functions and fractional-order derivatives. The FOMBW model demonstrates superior performance in capturing asymmetric and rate-dependent characteristics compared to the CBW model. These insights provide a foundation for selecting an appropriate model to accurately capture nonlinear current hysteresis phenomena in MRFDCs, ensuring precise torque control during gear shifting.</p><p dir="ltr">Finally, this study proposes a Gaussian Process (GP) hysteresis inverse model to compensate for the hysteresis observed in an MRFDC of an EV transmission, which enables the accurate determination of current commands corresponding to desired input torques, thereby facilitating feedforward compensation for the control of the MRFDC. The GP hysteresis inverse model effectively captures the nonlinear and rate-dependent hysteresis characteristics by incorporating both the clutch’s actual torque and its changing rate as inputs. Based on this point, a hybrid controller is designed, which integrates the Proportional-Integral (PI) feedback controller with a GP hysteresis inverse model-based feedforward compensator. The performance of the proposed hybrid controller is evaluated through a series of comparative experiments conducted on a prototyped MRFDC EV transmission platform. The results demonstrate that the proposed hybrid controller consistently outperforms a conventional PI controller after addressing the rate-dependent hysteresis characteristic of the MRFDC, exhibiting superior control performance across all tested scenarios, including single clutch torque tracking and clutch-shifting process.</p>