Identification by Means of Bayesian Inference of Properties and Parameters of Slender Structures Additively Constructed

Abstract

Nowadays additive construction technologies (or 3D printing) are not a surprise on any scale and are destined to be one of the spearheads of the fourth industrial revolution. When it was still an incipient technology and with some limitations, the interest in analyzing structural properties of the printed components was practically non-existent. In this work, /identification techniques are used to characterize parameters of models for straight and curved beams built additively by the Filament Deposition Modeling (FDM) procedure. The Bayesian inference approach is used with a 1D computational model of a beam with a-priori distribution information and a comparison is performed with various types of tests, and the use of Bayes’ Theorem to achieve a Posteriori distribution and the identification of the parameters. The tests will be static to infer parameters such as flexional and extensional stiffness, or elastic moduli. To carry out the inference procedure, the routines are programmed within the Bayesian Inference package offered in the UQLAB system.