Research On Motor Parameter Offline And Online Identification In Three-level Inverter

With the development of motor speed control technique, vector control is widely applied on industrial and civil field. In the vector control system, system will have a bad performance if the parameters of control system are not accurate. We must execute the identification of the motor parameter before we control an unknown induction motor.Based on the mathematical model of induction motor, this thesis proposes the off-line method that uses the resources of inverter to get the parameters of the induction motor. This method realizes detecting the parameter automaticallyThe method on off-line parameter identification for induction motor based on three-level inverter and vector control system identify stator resistance, leakage inductance, rotor resistance, leakage inductance, and mutual inductance by DC test, locked-rotor and no-load test without increasing hardware resources. Meanwhile, the thesis explains SVPWM theory and reconstructs voltage. In DC test, making use of the PWM voltage that inverter produces; we measure stator current and stator voltage in order to calculate stator resistance. In locked-rotor test, the induction motor rotor is locked so no electromagnetic torque is produced. At the moment, excited current is very high, and then the reference voltage should not be too high. We calculate rotor resistance and leakage inductance by means of measuring stator current and stator voltage as well as power-factor angle. In no-load test, induction motor is controlled under V/F. In order to calculate mutual inductance, stator current and stator voltage as well as power-factor angle are needed. To guide experiments, the text analyses hardware and software designs, including main circuit design, control circuit design and program flow chart. At the same time, a simulation is set up with MATLAB/Simulink. Compared with the real parameter of induction motor, the results show this method has a good feasible. These methods above belongs off-line identification. In order to further improve the performance of the inverter, the on-line identification is needed. The rotor time constant has the greatest effect on system by analyzing so that a method of rotor time constant identification with model reference adaptation based on rotor flux is proposed. The improved method has a good robust to load torque. For the two kinds of identification above, simulation models are build to illustrate the effectiveness in MATLAB/Simulink.Finally, the off-line identification of induction motor parameters for the proposed theory and simulation results are summarized. Then further researches are prospected.