چكيده به لاتين
In modern electric drive control methods, particularly in the control of Permanent Magnet Synchronous Motors (PMSMs), having precise information on the rotor’s speed and position is crucial. The accurate rotor position is essential for implementing the Park and inverse Park transformations, which are necessary for transferring voltage and current parameters to the rotating reference frame for effective control. Historically, and in methods known today as sensor-based motor control, sensors such as resolvers and optical position sensors have been commonly used. However, these sensors introduce drawbacks, including reduced reliability, increased weight, higher costs, and added system complexity. In sensorless methods, rotor speed and position are estimated using motor voltage and current information generated by the control system, with current feedback. Initially, back electromotive forces (EMF) on the axes of the stationary reference frame are estimated. This report proposes the estimation of the mentioned parameters through a smoothed sliding mode observer using soft functions. The applied voltages to the PMSM are fed as inputs to the motor model in the estimator, and the generated currents are compared with the motor's feedback current in the stationary reference frame. Using a sigmoid function, the estimated back EMF forces on the axes are calculated. Subsequently, rotor speed and position information is extracted based on the previous estimations. Traditionally, this information was obtained through the arctangent method, but due to significant noise in the estimated data, a Phase-Locked Loop (PLL) was introduced as a replacement. The feedback in the PLL significantly reduces noise in the estimated data. However, the fixed coefficients in the filter loop of this method prevent it from performing optimally under all operating conditions. In this study, by replacing the Proportional-Integral (PI) controller with a Super-Twisting Algorithm (STA) controller, a new method called the Super-Twisting Algorithm Phase-Locked Loop (STA-PLL) is introduced. Due to the non-linear behavior of the STA controller compared to the previous method, the coefficients change based on input error conditions, showing adaptive behavior that improves rotor position and speed estimation compared to previous methods.
