Development of a generalized chatter detection methodology for variable speed machining

•A generalized chatter detection methodology for milling operations was developed.•The method is suitable for non-stationary cutting conditions.•The chatter indicator is based on the theory of cyclostationarity.•The methodology was successfully tested in different cutting conditions. Regenerative chatter is one of the most deleterious phenomena affecting machining operations. It affects the integrity of the tool and the achievement of the targeted performance both for what concerns the material removal rate MRR and the quality of the processed surfaces. The majority of the chatter detection algorithms found in literature were not conceived for machining operations performed in non-stationary conditions although, it was demonstrated, that a continuous modulation of the spindle speed (spindle speed variation SSV) is one of the most profitable chatter suppression methodologies. This limitation represents an obstacle to the development of chatter controller systems that need to rely on effective and robust chatter monitoring procedures. In the present research, a chatter detection algorithm, specifically suitable for dealing with variable speed machining, was thus developed. More in details, the cutting stability assessment, performed in the spindle angular domain, is carried out through the real-time computation of a normalized chatter indicator that refers to the cyclostationary theory. Before computing the chatter indicator, the order tracking and the synchronous averaging methodologies are adopted for pre-processing the vibrational signals and the data coming from the spindle encoder. The devised chatter monitoring methodology was successfully validated executing real milling operations in which both constant and variable speed machining (SSV) were carried out. It was observed that the developed algorithm is capable of fast and robustly detecting chatter in all the tested cutting conditions.