Search

This paper proposes a stiffness sensitivity analysis with principal strain application to decrease the out-of-plane vibration, which is the main source of the sound radiation of mechanical structures with thin plate parts. The sensitivity is evaluated as a differential coefficient of the target response with respect to the design variable, e.g., stiffness or mass. For suppressing the out-of-plane vibration, we pay attention to finding an appropriate location on the structure to add local stiffness. The location is decided according to stiffness sensitivity analysis results. The compliance frequency response function (FRF) is considered as the target response, and the thickness of stiffener is considered as the design variable. The validity of the proposed method is examined through numerical simulation with a finite element method (FEM) model of a thin plate. The modal principal strain distributions, stiffness sensitivity, and FRF changes by local thickening are calculated based on 4 selected natural modes. It is also examined by the experimental approach. The expected reduction of the response is attained by adding the stiffener (a thin stainless plate) to the appropriate location on the plate. [ABSTRACT FROM AUTHOR]

If structures with complex shapes have space limitations, vibration tests using an exciter or impact hammer for the excitation are difficult. Although measuring the auto frequency response function at an unexcited point may not be practical via a vibration test, it can be obtained by assuming that the inertia acting on a dummy mass is an external force on the target structure upon exciting a different excitation point. We propose a method to estimate the auto frequency response functions at unexcited points by attaching a small mass (dummy mass), which is comparable to the accelerometer mass. The validity of the proposed method is demonstrated by comparing the auto frequency response functions estimated at unexcited points in a beam structure to those obtained from numerical simulations. We also consider random measurement errors by finite element analysis and vibration tests, but not bias errors. Additionally, the applicability of the proposed method is demonstrated by applying it to estimate the auto frequency response function of the lower arm in a car suspension. [ABSTRACT FROM AUTHOR]