An on-line performance optimization of the Switched Reluctance Motor

Abstract

The Switched Reluctance Motor (SRM) has been developed extensively during the past 15 years. Due to the simple motor construction and power converter requirements, SRM drives have been found competitive with traditional AC and DC drives. Optimal performance of the SRM drive may be described by different figures of merit such as drive efficiency, torque per ampere and torque ripple. This thesis addresses the problem of obtaining 'maximum torque per ampere' from the motor. The control of torque and speed in the SRM drive is complex due to the nonlinear relationship between the input variables (phase current magnitude and switching instants) and the motor torque. 'Tuning' of the drive is defined as the process of determining the set of input variables that results in maximum torque per ampere (TPA) at the required speed and load torque. The existence and uniqueness of a solution' to the maximum TPA problem are demonstrated by computer simulation of the machine model, using measured data of the inductance profile from an SRM. Due to manufacturing tolerances, the inductance profiles of the SRM show variations from phase to phase. Furthermore, the profiles may change with time due to effects such as wear on the bearings. A 'self-tuning' control strategy is proposed to compensate for these variations. An algorithm has been developed to maximize TPA on-line, while the SRM is running. It is based on perturbing the input variables and measuring the TPA after each change. This does not involve any complex computations based on the machine model but, just requires measurement of the observable variables of the drive system such as speed and current. The above strategy has been implemented on an SRM test bed with a 16-bit microcontroller. The drive operation is constrained to be in the lowspeed, chopping current control mode. A shaft position sensor is employed for commutation. Experimental results show the increase in TPA after execution of the optimization algorithm, Thus, a real-time self-tuning controller has been demonstrated, that is applicable to all SRM drives with shaft position sensors.