Research On Low Speed Performance Improvement Of Speed Sensorless Induction Motor Based On SMO

Induction motors are widely used in metallurgy,textile fields due to their low cost,simple structure and easy maintenance.Variable frequency control can effectively promote energy savings and emission reductions.To achieve high-performance variable frequency control,accurate speed and closed-loop are required.Speed can be obtained by installing encoders,which have the advantages of high accuracy,fast response and good performance.However,the installation of encoders leads to increased costs and is also susceptible to adverse factors such as humidity,heat and dust.Therefore,speed sensorless vector control was born.However,the stability of induction motors decreases under low-speed conditions,and in addition,there are problems such as low speed recognition and sensitivity to internal parameters.Therefore,this paper investigates the above problems at low speeds.When induction motor runs at low speed,SMO has low stability,the reaching rate is not controllable,and the chattering is easy to occur.The reason for these problems is that the switching function adopts the symbolic function.In this paper,the fundamental reason for its dynamic performance is analyzed.SMO is designed by using the idea of reaching law to consider the trajectory of the state point reaching the sliding mode surface.This observer includes stator current error,so that the variable can dynamically follow the changes,adaptively adjust the rate,and improve the dynamic while weakening the chattering.Meanwhile,in order to solve the problem that recognition accuracy decreases at low speed,a compensation method is proposed to further improve the accuracy.Stator resistance mismatch is also an important issue that leads to degradation of the observer performance.In this paper,the effect of stator resistance mismatch on the observer’s estimated rotor flux and stator current is analyzed,and the combination of online stator resistance identification and SMO is considered to eliminate this effect.A new stator resistance online identification method is designed.This method can accurately identify the stator resistance with only the stator current error and feed the identified stator resistance value to the observer.This can realize the simultaneous identification of stator resistance and speed,thus reducing the speed identification error,suppressing the electromagnetic torque and stator current fluctuations,and improving the convergence rate of the estimated speed.