External mechanical parameters identification of the elastic two-mass system with backlash

AC servo system will encounter outside disturbances and changes of system parameters when it operates. In order to obtain real-time information of the system and complete self-tuning of the controller parameters, online identification method of mechanical parameters is analyzed in the two-mass system with backlash. Firstly, the identification model and implementation process under no-load and loading conditions are studied. On the basis of that, recursive least squares forgetting factor (FFRLS) method is used to identify multiple parameters simultaneously. The effect of load torque and inertia ratio upon the identification precision is analyzed. For the case of no-load and low inertia-ratio, a shaft torque compensator is combined for backlash compensation, further improving the identification precision. Simulation results show that the linear identification method can be effective in the elastic-backlash nonlinear system under any loading or inertia ratio conditions.