DETERMINATION OF ELASTIC PROPERTIES BY MEANS OF FINITE ELEMENT MODEL UPDATING

Resumo

The range of available 3D printable materials is constantly expanding and unlocking
advanced applications. Thus, the using of 3D printers to manufacture industrial parts is increasing due
to its adaptable manufacturing of complex geometries and reduction in material waste. Especially the
correct adoption of material elastic properties is central to secure and consistent design. However, the
methodologies that allow the researchers characterize the mechanical properties of the parts after the
impression without damage them are scarce. In this way, this paper presents the calibration of elastic
properties by means of a hybrid numerical-experimental methodology that objective the automatic
updating of a finite element model regarding the experimental natural frequency. Here, natural
frequencies were obtained through acoustic tests of a sample of 30x30x30 mm3 with a square periodic
distribution of circular cylindrical holes manufactured by a Fused Deposition Modeling (FDM) 3D
printer. The sample was positioned on a device which simulates the free-boundary condition and excited
by impact using a standard hammer. 10,000 numerical models were processed in ABAQUS and the
calibration was carried out employing a genetic algorithm implemented in Python. Based on
experimental measurements the elastic constants were determined. This proposed methodology presents
high accuracy, the natural frequencies obtained by the calibrated numerical model differed in less than
3% of the obtained ones by the experimental model and could be applied in a wide range of materials.