A new modal analysis method applied to changing machine tool using clustering

Abstract

The states of the machine tool, such as the components’ position and the spindle speed, play leading roles in the change of dynamic parameters. However, the traditional modal analysis method that modal parameters manually identified from vibration signal is greatly interfered by harmonics, and the process of eliminating interference is very inefficient and subjective. At present, there is a lack of a standard and efficient method to characterize modal parameter changes in different states of machine tools. This paper proposes a new machine tool modal classification analysis method based on clustering. The characteristics related to the modal parameters are extracted from the response signal in different states, and the clustering results are used to reflect the changes of machine tool modal parameters. After the amplitude of the frequency response function is normalized, the characteristics related to the natural frequency are acquired, and the clustering results further reflect the difference of the natural frequency of the signal. The new method based on clustering can be a standard and efficient method to characterize modal parameter changes in different states of machine tools.