Identification of natural frequencies of machine tools during milling: comparison of the experimental modal analysis and the operational modal analysis

Abstract

Modal parameters (natural frequencies, mode shapes and modal damping) help to understand the dynamic behaviour of complex systems like machine tools. There are several approaches for finding the modal parameters. The Experimental Modal Analysis (EMA) has proven to be effective at standstill of a machine tool. The excitation, realized with impulse hammer or shaker, and excited responses at several locations are measured. Alternatively, the Operational Modal Analysis (OMA) can be deployed for finding the modal parameters during operation. Here, responses to excitation resulting from operation are only measured. The modal parameters are mathematically identified from the measured signals in both cases but with different methods. This paper discusses, to what extent both approaches (EMA and OMA) can lead to plausible identification of natural frequencies of a machine tool during milling. Concerning the EMA, attention is paid to capturing the excitation. Process forces can be assumed to be the most significant excitation. However, there are other excitation sources beside the process forces (e.g. drives, hydraulic and pneumatic aggregates), which are considered by this assumption to be a part of disturbances with consequence for the identification of the modal parameters. Regarding the OMA, attention is paid to the fact that the excitation is assumed to be broadband like the white noise. Unfortunately, this assumption does not match the characteristics of a real excitation. This paper contains the identification of natural frequencies of a machine tool during milling within both approaches. The achieved results are compared and discussed.