Dynamic Programming-Based Optimal Torque Control of Induction Machine

• Applicant: Professor Dr.-Ing. Gerd Griepentrog
• Subject Area: Electrical Energy Systems, Power Management, Power Electronics, Electrical Machines and Drives
• Term: from 2016 to 2019
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 323850416

The attempts to control the torque of electrical machines have led to development of different control strategies where Direct Torque Control (DTC) and Direct Self-Control (DSC) as well as strategies based on Pulse Width Modulation (PWM) are known today. The drawbacks of control strategies based on PWM are the unnecessary extra switching operations of the inverter and high Parameter dependencies of the controller. The DTC and DSC control strategies suffer from high torque, flux, and current ripple, but show only low switching losses. With this proposal, a new approach for controlling electric machines is to be developed. Whether the inverter should change its switching state will be decided based upon the present and future state variables along a trajectory which should enable optimal behavior of the drive system. For optimization purposes a cost function will be minimized, which contains weighted parameters such as the deviation of flux and torque from their reference values, time to reach a pre-set speed or position, number of switching events, losses of the inverter and the IM etc. In order to solve the optimization problem within reasonable time and effort, the method of Dynamic Programming will be used. This method solves the optimization problem for a manageable number of future states (the event horizon) and therefore enables an implementation for real-time controller of electric drives. The method is therefore called Dynamic Programing Torque Control (DPTC). With this proposal, the potential and the benefit of DPTC will be investigates by controlling an induction machine (the method, however, can be applied to any type of electric machine). First simulations have shown the huge potential of DPTC compared with methods such as DTC or DSC.

• DFG Programme: Research Grants