Parameter identification for linear model of the milling process using spindle speed variation

Abstract

This study proposes a rapid and accurate identification method for model parameters of the milling process. The proposed method identifies each parameter through an inverse analysis of the in-process data during machining. Furthermore, the structural vibration and cutting force fluctuations caused by machine-process interaction were incorporated into the milling process model. Additionally, the disturbance force transmitted from the tool-workpiece system to the machine structure was modeled. We estimated the cutting force based on the disturbance force that compensated for the inertial force generated in the vibrating machine structure. By separating the cutting force into nominal and chatter components, we were able to identify the parameters of the cutting process and the dynamics using each component type. The parameters were solved using the least-squares method after a linear approximation of the milling process model. In the proposed method, the disturbance force data during spindle speed variation (SSV) was applied to the identification procedure. For a cutting condition with constant spindle speed (CSS), only a quasi-steady vibrational state was generated, which in turn depended on the cutting condition. Conversely, the SSV condition, which forces the state transition, can induce vibrational states that are advantageous for identification. Computational end milling simulations were performed to validate the proposed method. It was found that under CSS conditions, the identification accuracy was not sufficient owing to the chatter-free or chatter conditions, which is highly associated with nonlinear vibration with the multiple regenerative effect. Under the SSV condition, the vibrational states can transition in one cycle of the variation period. Therefore, efficient and accurate identification can be achieved by extracting the disturbance force data in an appropriate time section during the SSV.