Machine tool multibody dynamic model updating using vision-based modal analysis

Abstract

Machine tool dynamic behaviour is influenced by the structural properties of its subsystems assembled at interfaces as well as by the interface characteristics. Interfaces are commonly modelled as spring-damper connections, parameters of which are usually updated by minimizing the difference between model-predicted and measured dynamics characterized by frequency response functions. This model updating approach requires global mode shapes to be measured by roving the hammer and/or the sensor such as to localize the joint parameters to be updated. Such measurements are time consuming and fraught with errors. As a new, alternative, and simpler way to update joint parameters of a machine tool multibody dynamic model, this paper reports on the use of full-field vision-based modal analysis methods. Mode shapes thus identified agree with those estimated with the traditional experimental modal analysis procedures. The updated machine tool multibody dynamic model is a step towards realizing an accurate digital twin.