Fuzzy logic based online electromagnetic loss minimization of permanent magnet synchronous motor drives

Abstract

The objective of this paper is to give an insight about the development of a fuzzy logic efficiency control system which is incorporated to a standard adjustable speed drive control (i.e. Indirect Vector Control) in order to improve the online efficiency in Permanent Magnet Synchronous Motor (PMSM) drives for both transient and steady state operation. The search criterion in steady state is the minimization of the drive power losses by reducing the stator flux while meeting the speed and load torque demands. The power losses are calculated as the difference between the measured input DC-link power and drive's output power. The power losses are calculated at every sampling step and the change of power losses are calculated over a constant time interval as it is needed by the efficiency control algorithm. Variables in the operating characteristics of the PMSM, such as DC input voltage/current and actual speed are inputs to the fuzzy control system, making the control system independent of any specific load. Our system combines two fuzzy logic designed efficiency controllers which are introduced as Fuzzy(1) and Fuzzy(2) controllers in order to generate the magnetizing current (d-axis component of the stator current). They are separately activated during the steady and transient state respectively. The Fuzzy(1) controller is a search controller (SC) operating in steady state, so to decrease the stator flux for higher efficiency. The Fuzzy(2) controller is a SC operating in transient state, so to increase at minimum the magnetizing current, to achieve fast tracking capability, less steady state error and robust to load disturbance during transient state. The torque component of the stator current is generated by the vector controller in order to achieve higher motor performance. The fuzzy logic controllers (FLCs) prototypes were created and tested using Simulinkreg simulations. Simulations are accomplished in Simulinkreg in order to validate our method (i.e. efficiency improvement, less torque ripple with flux changes and the less speed sensitivity to load disturbances). Experimental results will be available upon conclusion of the laboratory testings. Future experimental deployments will include a DSP board which will be directly programmed in assembler language via Simulinkreg.