Research On Sensorless Control Of Induction Motor Based On Full-order Observer

Vector control has been widely used in AC motor control application. The sensor-less vector control, which can cut the cost and improve the reliability of the system, has become research hotspot in recent years. While there are many ways to implement the sensor-less vector control, this dissertation focuses on the way based on the full order observation, which involves several items of technology. In this dissertation, flux observation, speed estimation and parameter identification have been discussed in detail.Firstly, the simulation and experiment platform of vector control with sensor has been established, laying a solid foundation for the research of sensor-less control. Then obtaining the flux and speed through full order observer based on the theory of Model Reference Adaptive System (MRAS):treat the induction motor itself as reference model, and full order observer model as adjustable one, the motor speed could be identified. Thus the sensor-less vector control can be realized.The discretation is an indispensable step for the digital implementation in practical. This dissertation demonstrates the advantages and disadvantages of the commonly used discrete methods, including Euler and bilinear transformation. When high switching frequency is needed, an improved method is taken. That is to discretize two of the four state variables by Euler and two others by bilinear transformation. When switching frequency is lower and time for executing complex calculation is adequate, four state variables are all discretized using bilinear transformation. The validity of both methods has been testified by experiment. This dissertation also illustrates the influence of parameters variation to the full order flux observation. The reliable values of parameters can be acquired through offline and online parameter identification technology. This dissertation pays more attention to the first technology, and sum up a whole set practical method to identify motor parameters offline. Based on the research of the above technology, the sensor-less vector control using full order observation has been achieved in wide speed rang.