Sensorless speed control of a PM assisted synchronous reluctance motor from zero to rated speed

Abstract

Recently, permanent magnet synchronous motors (PMSMs) with high-energy magnets are proposed to replace the variable speed drives of some induction motors (IMs) to increase power factor and efficiency. However, the high price of high-energy rare earth permanent magnets, especially for applications with upfront cost constraints, calls into question this solution. Permanent magnet assisted synchronous reluctance motor (PMaSynRM), which has a high-efficiency standard and high torque and power density, where rare earth magnets' usage is not preferred, is expected to replace IMs, which is currently the most used motor type. In this study, a PMaSynRM designed for a washing-machine appliance is used to apply a sensorless speed control approach from zero speed to rated speed. The working principle of PMaSynRM is based on the reluctance torque beside the electromagnetic torque. The dominant one is the reluctance torque, and this is the subject where it differs from an interior permanent magnet synchronous motor (IPMSM). In a PMaSynRM, the reluctance torque is provided by rotor air barriers and the electromagnetic torque is provided by permanent magnets placed on the q-axis of the rotor. The mathematical model of PMaSynRM can be derived simply in the rotating reference frame according to rotor. The variables and parameters of stator axes are transformed into the d&q reference frame by using the Clarke and Park reference frame transformations. Amplitude and angle of stator current vector are controlled separately by performing field-oriented control (FOC). Different control strategies are implied to the motor applications in literature taking advantage of FOC. The main one for a PMaSynRM is maximum torque per ampere (MTPA) control because of its simplicity. Another advantage is that the output power is also maximized for a specific speed value besides the output torque. Thus, the motor torque and power density are increased. A two-level inverter is sufficient for a three-phase PMaSynRM driver. The amplitudes and frequencies of phase voltages and currents are adjusted to the desired value by using inverter and FOC technique together. The space vector pulse-width modulation (SVPWM) technique becomes prominent among pulse-width modulation (PWM) techniques for inverters. Sensorless control is based on obtaining the rotor position and speed information not from the position sensor. In this study, the extended electromotive force (EEMF) and high-frequency injection (HFI) models are preferred. The HFI model is applied below a certain speed due to the inadequacy of the EEMF model at lower and zero speeds. In the first step, the PMaSynRM state space model includes EEMF components is obtained on the α&β stationary reference frame. The EEMF components containing rotor position information cannot be obtained directly by measurement. Therefore, EEMF components need to be estimated using an observer structure. In this study, the sliding-mode observer (SMO) structure is used. The estimation errors of α&β-axis currents are used for estimating EEMF components. As a result of simulations, sensorless speed control of PMaSynRM is realized from 300 to 3000 rpm (rated) and also under rated load conditions. For rated speed where efficiency is high and critical, the electrical position estimation error is around 2°, which is acceptable. In the second stage, the HFI technique is applied for the PMaSynRM. High-frequency voltage injection (HFVI) is preferred, since the injected signals are not regulated by the current controllers. As a result of the injected high-frequency voltages, high-frequency signals are formed in the phase currents of the motor. These current signals contain rotor position information. However, the signals must be subjected to some signal processing to obtain this rotor position information. These signals are separated from the motor fundamental current, and position information. A smoother position information is provided with the designed phase-locked loop (PLL) technique. In simulation environment, PMaSynRM sensorless speed control is carried out from zero to 300 rpm speed. In addition, sensorless speed control is also provided for the rated load condition. As a result, the electrical position error is less than 2° for the entire speed range, which is reasonable level for a sensorless operation. In the final stage, the experiments of sensorless control of PMaSynRM are carried out. With the EEMF technique, both rotor position and speed are successfully estimated and the motor is driven from 300 rpm to rated speed. The noise of estimations increases at low speeds because of lower values of EEMF components as expected. For lower speeds, the HFVI technique is tested. The motor is driven under rated load at 30 rpm. Then, the 100 rpm step responses of motor are obtained under rated load. Finally, the motor is tested to be rotated in the opposite direction. With the designed speed control algorithm, PMaSynRM is provided to operate sensorless from zero speed to rated speed. The efficiency value obtained with the sensor is approximately kept, thus avoiding the concern of efficiency loss.