Real-time Loss Minimizing Control of Induction Machines for Dynamic Load Profiles under Deadbeat-Direct Torque and Flux Control

This paper develops a real-time loss minimizing control technique on induction machines for dynamic load profiles. Steady-state loss minimization has been well addressed. However, when the steady-state optimal control is applied to a highly dynamic load profile, more losses are induced during transients due to the slow rotor flux dynamics. Several heuristic methods have been developed for real-time loss reduction, but they are not designed for minimum losses. Loss-minimizing flux trajectories have been generated by dynamic programming but requires prior information of load profiles. In this paper, a real-time loss minimizing control technique is developed for applications without prior information of load profiles. It resembles the loss-minimizing stator flux trajectory generated by a modified dynamic programming method, which generates the best practical solution for unknown load profiles. Both simulation and experimental results demonstrate the energy saving capability of the proposed method. Significant energy savings have been demonstrated on traction applications.