A new approach is presented to detect milling chatter in peripheral milling.
The non-stationary chatter signal is decomposed based on the VMD and WPD.
Energy entropy are utilized to demonstrate the condition of milling system.
The method has an excellent performance for chatter identification.