Banca de DEFESA: CRISTIANO PEREIRA DA SILVA

Uma banca de DEFESA de DOUTORADO foi cadastrada pelo programa.
STUDENT : CRISTIANO PEREIRA DA SILVA
DATE: 28/03/2024
TIME: 14:00
LOCAL: Microsoft TEAMS
TITLE:

IMPLEMENTATION OF HYBRID METHOD FOR DETERMINATION OF ISOTROPIC HARDENING PARAMETERS

KEY WORDS:

Parameter Identification, Genetic Algorithm, Levenberg-Marquardt Method, Gurson Model, Additive Manufacturing

PAGES: 151
BIG AREA: Engenharias
AREA: Engenharia Mecânica
SUMMARY:

The process of identifying parameters using various optimization methodologies has been shown
to be an important tool in the determination of parameters of constitutive models characteristic of
materials. The degree of complexity of the constitutive relation of the material is a determining
factor in the choice of the optimization method, since the objective function can present many
local minima or maximums. The present study sought to identify and explore the problems and
difficulties that may arise during the process of identifying parameters. We developed
computational routines that involved an evolutionary algorithm, the Genetic Algorithm (GA) and
a gradient-based algorithm, the Split Levenberg-Marquardt Algorithm (LMD) and proposed a
hybrid model that would unite the search robustness of GA and the speed and accuracy of the
LMD Algorithm. The developed routines were tested and several analyses were performed, both
of the Genetic Algorithm (GA) and of the Split Levenberg-Marquardt Algorithm (LMD) that
received as input the values from the GA. Overall, the performance of both was good, arriving at
the solutions within a desired precision in a few iterations. The critical point verified was the
Finite Element method, represented here by the HYPLAS system. An academic system useful in
the development of the hybrid method, but which added to each iteration of GA and the LMD
Algorithm several seconds of processing. As for the traction test, widely used to obtain
information about a given material and also as a specification control test, some important
characteristics were evaluated and a routine was created to automate it. In conjunction with the
results obtained via the tensile test for the 4340 standardized and annealed steels and the alloy
1524 U2 it was verified the application of the idealized routines in the PYTHON language for the
GA, the LMD and the hybrid method in the determination of crunching parameters according to
the model of Kleinermann and Ponthot (2003) for these materials. Finally, a case study was
carried out in which it was sought to verify the applicability of the porous materials model of
Gurson (1977) in the characterization of the mechanical properties of specimens produced via
additive manufacturing (AM). The hardening curve of the material until fracture and the level of
initial and critical porosity at different points and directions of metal deposition were measured.
It was concluded that the model based on Gurson (1977) can describe the mechanical behavior of
materials supplied by additive manufacturing and can be used as an alternative to predict the
fraction of volumetric void, a defect present in these types of materials. It was also observed that
in the different directions of deposition, the porosity of the material ranged from 5% to 18%,
which influenced the resistance characteristics, such as flow stress, modulus of elasticity and level
of plastic deformation accumulated in the fracture.

COMMITTEE MEMBERS:
Interna - 1125400 - DEBORAH DE OLIVEIRA
Interno - 1249764 - JORGE LUIZ DE ALMEIDA FERREIRA
Presidente - 1816660 - LUCIVAL MALCHER
Externo à Instituição - RAPHAEL ARAÚJO CARDOSO - UFRN
Externo à Instituição - RUI RICARDO LOUREIRO AMARAL - UNIPORTO