Integrating tools to improve finite element models of machine tools via experimental modal data

International audienceThis paper presents a specific procedure to improve FE models of multi-axis machine tools that integrates well-tested numerical and experimental techniques for this type of mechanical systems: modal analysis and testing, Design of Experiments (DoE), sensitivity analysis and model updating. First, it is shown that experimental modal analysis must be performed by exciting the machine tool along every main direction and in different geometrical configurations to obtain a complete set of mode shapes and to evaluate the variation of critical parameters like stiffness of joints. In addition, it is demonstrated the interest of using lumped-mass models and DoE techniques to set limit values for parameters that describe joints between structural components and connections to the machine foundation. Results confirm that sensitivity based model updating using natural frequencies and MAC values as responses provides improved FE models that match reasonably well with experimental data