| Summary: | 碩士 === 國立成功大學 === 工程科學系碩博士班 === 98 === Recently, since people changing their live style in many aspects, the living standard is as higher as possible and the gasoline is greatly consumed. According to statistic from BP Statistical Review, the index of reserved gasoline can only provide about 36~43 years. By the reason, human must develop new energy source which is sustainable and less environmental pollution in following decade. At this time, many researches are working on decreasing the gasoline-consuming efficiently to energy-saving. Since vehicles consumed numerously gasoline, improve these vehicles to energy-saving is important. Electric vehicle (EV) replaced internal combustion engine (ICE) by electric motors is more popular in the world. However, EVs are hard to popularize since the sustainability is lower than it with internal combustion engine. This defect is unacceptable for user to buy or use EVs widely. Therefore, increase the sustainability of EV is a novel research topic.
In generally, the conventional brake applies the friction to decrease the vehicle’s speed, and translate the kinetic energy to heat. The energy is just consumed. In order to recycle the kinetic energy in braking process, regenerative brake method is presented. Regenerative brake translates the kinetic energy to electric energy by utilizing the motor’s back-EMF and internal winding. The method controls the switch sequences of MOSFETs to elevate the back-EMF voltage and recharge the battery. In the meanwhile, the motors operate in braking mode, and produce an inverse torque to reduce the motor’s speed. Depend on the regenerative brake method, the motors drive can accomplish brake function and extending traveling distance without adding any component.
“Extending the traveling distance” and “Control the regenerative current” are the topics of this thesis. Since the brake torque is depended on the recharge current, a current control in EVs with self-tuning fuzzy neural network algorithm is proposed. The gradient descent and back-propagation are used to adjust the parameters of fuzzy neural network, and minimize the current error. Finally, Lyapunov’s stability theorem is applied to prove the system stability.
Finally, the microprocessors dsPIC30f2010 by Microchip, are employed to implement the proposed control algorithms. The driving and braking mode are tested on a motive machine and inertial instrument. Moreover, the implemental results of the proposed system set up on EVs demonstrate the ability of recharge and the feasibility of regenerative brake.
|
|---|
