Research On Parameter Identification Method Of Induction Motor Based On MRAS

With the rapid development of railway passenger transport in China,the railway passenger traffic volume has increased year by year.At the same time,passengers have put forward higher requirements on the punctuality and comfort of trains.As the power source of trains,how to improve the control performance of traction induction motors has become the key to meet passenger's requirements.Accurate motor control parameters play an important role in high performance control of induction motors.On the basis of summarizing the research status of control strategies,flux observation methods and parameter identification technologies of induction motors at home and abroad,this thesis focuses on the rotor flux estimation and the on-line parameter identification method of model reference adaptive system based on vector control system.Verification and analysis are performed by a series of simulations and experiments.First of all,starting from the basic mathematical model of induction motor,the coordinate transformation matrices based on the constant total power and the constant amplitude is obtained.On this basis,the unified description modezonggonglul of the induction motor in two-phase coordinate system is summarized,and the indirect vector control system of induction motor based on rotor flux linkage is derived.Then,the parameter sensitivity of the motor electromagnetic torque is analyzed by drawing the three-dimensional map of parameter changes.Secondly,based on Bode diagrams of flux ratio sensitivity analysis functions,the parameter sensitivity analysis of standard voltage and current rotor flux observer are realized.In order to solve the problem of back EMF integration,two common improved methods are presented,and then the improved voltage rotor flux observer based on DC offset compensation is emphatically deduced.Then,based on the Jury test and pole of the system,the stability of three discrete models of current and voltage rotor flux observers are analyzed.And the current and voltage discrete flux observer models which are easy to be realized by the control system are selected according to the estimation accuracy.Subsequently,based on Popov hyperstability theory and generalized error of rotor flux,a rotor time constant on-line identification method of MRAS is studied,and the mutual inductance is corrected based on the rotor flux magnitude error.Further more,considering the influence of stator resistance on the accuracy of the reference model in the constructed method,a two-parameter online identification method based on mutual correction of rotor flux is studied.On this basis,in order to improve the identification accuracy of actual system parameters,a stator voltage reconstruction method considering the non-ideal factors compensation of inverters is deduced.Finally,the experimental platform based on the traction drive system of CJ5 EMU is introduced in detail,and the above theoretical research is verified and analyzed by the experimental platform.There are 53 figures,1 table and 73 citied references in this thesis.