Identification of the model parameter for milling process simulation with sensor-integrated disturbance observer

Abstract

This study presents a novel parameter identification method using a disturbance observer for computer numerical control (CNC) machine tools with ballscrew-type feed drive systems, which are generally adopted in industrial applications. In the proposed method, the disturbance force acting on the feed drive system was estimated using a multi-inertial disturbance observer using both CNC servo and accelerometer signals. Model-based milling simulation is performed with reference to the estimated disturbance forces, and hence model parameter identification can be systematically achieved. In the milling process model, the structural vibration and cutting forces caused by machine-process interactions are considered. Furthermore, the model is considered the disturbance forces transmitted to the feed drive system via a mechanical structure. To verify the proposed method, a basic end-milling experiment was conducted. It was confirmed that the multi-inertial disturbance observer assisted by the accelerometer is equivalent to the dynamometer in terms of the disturbance force estimation within the frequency range of 600 Hz and less. The model parameters identified using the proposed method were almost identical to those obtained using conventional methods. However, the identification results of the modal parameters and tool eccentricity were slightly different from those estimated using the conventional methods. The milling simulation using the identified parameters agreed with the measurement results obtained using the dynamometer. Hence, inverse analysis of the in-process data can effectively improve the identification accuracy of the model parameters.