Development of friction identification methods for feed drives of CNC machine tools

Friction exists in the feed drive servomechanisms of CNC machine tools and usually affects their motion. Although many experiments have been performed to identify the friction characteristics of a feed drive servomechanism, problems such as unmodeled dynamics and position-dependent perturbations still exist and significantly affect the friction identification results. In this study, in order to obtain reliable friction and velocity data values, a modified velocity control system is developed to reduce the effects induced by unmodeled dynamics, and a friction extraction method is developed to remove the adverse effects of position-dependent perturbations. Furthermore, several experiments and motion tests are carried out on a three-axis CNC milling machine to demonstrate the feasibility of the proposed friction identification method. The experimental results indicate that the friction-compensated motion control system with the friction model obtained in this study reduces the root mean square value of tracking errors by 44.16%. Moreover, as compared to a conventional identification method, the root mean square value of tracking errors is found to further reduce by 9.52% in the experiments.