Modelování a simulace bezsenzorového řízení asynchronního motoru s využitím MRAS pozorovatelů

Abstract

The main objective of this work was to implement sensorless control of an induction motor based on Model Reference Adaptive System (MRAS). After in-depth literature survey of MRAS approaches, MRAS based on fictitious quantity X was selected as a suitable candidate for further implementation. As presented in the literature such relatively novel approach is said to be robust and enables motor to be controlled in all four quadrants. In order to verify parameters of such MRAS, time based model was created in Matlab-simulink which included indirect field oriented control (IFOC) controller, three-phase induction motor and X-MRAS estimation structure. After verification of IFOC controller and induction motor, several implementations of X-MRAS sensorless model were tested in both open and close loop and various parameters were monitored. Results have shown that implementation as presented in state-of-the-art literature does not always guarantee field oriented control which enables induction motors to be controlled like DC counterparts. Moreover, simulations have shown that high speed, long start-up time and large up to nominal load torques are rather problematic due to the loss of oriented coordinates as opposed to low speed, short start-up time and light load torques which performed relatively satisfactory. The results of the simulations also proved that the more the control is from the oriented coordinates, the less torque the motor is capable of delivering. Based on the findings several approaches capable of increasing the range of operation have been proposed and compared. DFOC based control scheme, increase of magnetic excitation, dynamic adjustment of rotor constant within adaptive MRAS model or auxiliary magnetic flux controller are all capable of increasing the range of operation for X-MRAS. Nonetheless, it has been proven that the best approach is to incorporate within the X-MRAS structure the auxiliary magnetic flux controller with current model of induction machine. Such approach is well capable of controlling the motor in oriented coordinates in steady state. The final simulations investigated performance of this adjusted X-MRAS in both low and high speed simulations including nominal load torque applied and all 4 quadrant operation. Last part of the work evaluated sensitivity of different parameters when those parameters differ from nominal ones. It has been demonstrated that rotor time constant plays important role in precise estimation for both IFOC and X-MRAS. Lastly, the non-viable harmonic inverter was replaced by Space Vector Pulse Width Modulator and performance verified.