I. Introduction
Chatter is the self-excited high-amplitude vibration in metal cutting. Chatter has been researched for more than a century, and it is still a major problem in machining processes including turning, milling, and drilling [1], [2]. When chatter occurs, it will lead to poor surface quality, excessive noise, accelerating tool wear, and waste of materials. For these reasons, chatter suppression is a topic of enormous interest. Spindle speed variation (SSV) is an active technique for chatter elimination [2]. The regenerative chatter can be suppressed, since after the speed is changed, there is no regenerative feedback loop. This kind of modulating the spindle speed is attracting increasing interest because of its simplicity and effectiveness in chatter suppression. In [3] and [4], the SSV is implemented by online computing the preferred spindle speed. Although these algorithms have different expressions for calculating the preferred spindle speed, the chatter frequency is an essential part, which must be identified before chatter suppression. Al-Regib et al. [5] proposed a programming sinusoidal SSV to suppress chatter with the optimum frequency and amplitude ratio. The spindle speed is varied as a sinusoidal function, which actually disrupts the regenerative chatter. In their algorithm, the chatter frequency and current spindle speed are needed. Therefore, the dominant chatter frequency needs to be estimated from the cutting force or acceleration signal once the onset of chatter has been detected.