Research On Speed Estimation Methods For Improving The Robust Performance Of The Induction Motor Drive System

For frequency control systems,sensorless control technology has become a significant research direction in the field of AC motor frequency control system since the sensors increase the cost,reduce the reliability and make the system susceptible to the influence of working environment.With appling more high-tech science and technology into the industrial field,the sensorless drive system is required to achieve not only the satisfactory performance under normal working conditions,but also reliable operation state when external or internal dist-urbances occur.In this thesis,the induction motor(IM)is chosen as the research object,and two representative speed identification methods,i.e.adaptive full-order observer(AFO)and sliding mode observer(SMO),are selected to study the key technology of improving their robust performance.By introducing robust control strategy,the robust performance of the observer is improved,and the stable and reliable operation state of the frequency control system under disturbance is guaranteed.AFO has the characteristics of high accuracy and good generality.However,when the gross error disturbances occur,the speed estimation accuracy will drop sharply,which may lead to instability in serious cases.To solve this problem,a speed estimation method based on robust AFO is proposed in this thesis.The influence of external disturbances and internal estimation errors on the speed estimation performance of AFO is analyzed in detail and the robust mechanism is introduced into AFO.When the gross error disturbances occur,the feedback gain matrix coefficients are adjusted adaptively.The influence of external gross errors and internal estimation errors on the estimation performance of the observer is decreased effectively,which improves the robustness of the observer.SMO adopts variable structure control,which is simple in structure and easy to implement.However,in the sliding mode process of the traditional SMO,it is usually only concerned whether the space state point can finally reach the sliding surface,i.e.whether the observer can keep stable.However,the motion trajectory of the space state point to the sliding surface is ignored,which limits the dynamic performance and anti-disturbance ability of SMO.To solve this problem,a speed estimation method based on SMO with improved equal reaching law is proposed in this thesis,the equal reaching law is introduced into SMO,and the motion trajectory of the space state point is controlled.Meanwhile,combined with the application characteristics of SMO,the equal reaching law is optimized.The proposed method improves the dynamic response and robust performance of the system on the bisis of ensuring stability.The traditional SMO has a drawback that the fast response and chattering suppression can not be achieved simultaneously.When the rate of space state point to the sliding surface is accelerated,i.e.the dynamic response of SMO is increased,the inherent chattering problem of SMO will be more serious.While when the chattering phenomenon is suppressed by reducing the sliding mode gain,the dynamic response will be slower,and its robust performance will accordingly decrease.To solve this problem,a speed estimation method based on SMO with improved exponential reaching law is proposed in this thesis.In this method,the exponential reaching law is introduced into SMO and optimized by adjusting exponential reaching law gain in real time.The proposed method can keep the fast dynamic response of SMO and suppress the chattering phenomenon simultaneously.Moreover,the robustness of the observer against motor parameters variation and external load disturbances is enhanced.As a new control method,model predictive control(MPC)has attracted much attention,and has become a research hotspot in the field of power electronics and electric drives in recent years.At present,the sensorless application of MPC of IM is still in its infancy.The inherent problems of MPC,such as high sampling frequency,large computation burden and unfixed switching frequency,make it difficult to achieve ideal control performance,which greatly limits the application and industrialization of MPC.For enhancing the control performance of sensorless MPC system,especially for the low speed performance,a sensorless MPC for IM based on two-parameter identified AFO is proposed in this thesis.The proposed method identifies the stator resistance while estimating the rotor speed,which effectively reduces the adverse effect of stator resistance variation on speed identification performance in low speed range,and it has excellent low speed load capacity.In addition,by appling SMO into the MPC system,a sensorless MPC based on SMO with optimized exponential reaching law is studied.The method achieves good control effect and robust performance by accurately identifying the motor speed and flux linkage,which provides a practical solution for further research and industrialization of MPC.