OPERATIONAL MODAL ANALYSIS FOR CHATTER PREDICTION IN MILLING

Unstable self-excited vibrations in machining are known as chatter and must be avoided by selecting appropriate machining parameters (e.g. spindle speed and cutting depth). Nonetheless, even when the machining parameters are carefully selected according to chatter models, vibrations may still become unstable due to unmodeled dynamics and process variations. Therefore, it is critical to monitor the process and detect chatter promptly. This paper is an extension of our lab's previous work, where we used Operational Modal Analysis (OMA) to monitor the loss of stability while the process is still stable, unlike current methods that detect chatter only after it occurs. This approach was applied to turning in our previous work, and the present work reports the preliminary results of extending its application to milling. While turning dynamics are time-independent, milling dynamics exhibit periodic variations due to tool rotation. Consequently, it becomes essential to adopt fundamentally different OMA theories and data acquisition methods to address time-periodic characteristic. To address this, we construct a lifted time-independent representation of the periodic dynamics from the measured process vibrations, enabling the application of standard OMA methods to the periodic system. The effectiveness of the presented method in predicting different types of milling chatter is demonstrated by numerical simulations and comparison with the Semi Discretization Method (SDM).