Date of Award

2018

Degree Type

Dissertation

Degree Name

Ph.D.

Department

Mechanical, Automotive, and Materials Engineering

First Advisor

Green, Daniel

Second Advisor

Alpas, Ahmet

Keywords

Aluminum Alloys; Artificial Neural Network; Electro-hydraulic Forming; Fracture Mechanism; High Strain Rate Deformation; Numerical Simulation

Rights

CC-BY-NC-ND

Abstract

In this research, formability improvement of AA5182-O aluminium sheet during electrohydraulic forming (EHF) was investigated by means of experimental and finite element analysis. Free and conical die formed EHF was carried out on grid sheet blanks and formability improvement was measured by comparing grid analysis results at each EHF condition with different forming limit curves (FLCs). It is found that AA5182-O shows minor improvement in formability when formed freely into EHF. But a significant rise in effective strain in safe grids is observed when EHF into 34, 40 and 45-degree conical dies provided a critical threshold of input energy has been used. In order to understand the mechanical aspects of formability improvement, related factors such as strain rate, stress triaxiality, and compressive through-thickness stress were studied using finite element simulation with an accurate description of the hardening behaviour of AA5182-O. Another advantage of the numerical simulation carried out in this work is that unlike previously published works, the driving force for EHF deformation was not simplified as uniform pressure and it resembles the actual process of EHF. Ignition and growth model was used in conjunction with Coupled Eulerian-Lagrangian (CEL) approach to simulate the EHF pulse formation. Moreover, 3D solid elements were used instead of shell elements and this facilitated measurement of stress in elements located in the bulk of sheet material. The tensile flow behaviour of AA5182-O sheet was investigated in the strain rate range of 0.001 to 1000s−1 and at different material directions (RD, DD, and TD) by means of both phenomenological models and neural networks (NNs). Genetic algorithm (GA) and linear regression analysis methods were used to calculate the constants in Johnson-Cook (JC), Khan-Huang-Liang (KHL) and modified Voce hardening functions, and user-material subroutines were developed and used in FE software. Moreover, in order to predict the rheological behaviour of AA5182-O without the limitations of a mathematical function, two types of feed-forward back-propagation neural networks were trained and used in the FE model. Simulation results were compared with experimental tensile flow curves and the most accurate method is used to predict the mechanical response of AA5182-O in FE simulations of the EHF process. FE results suggested that a combination of EHF process related parameters including compressive through-thickness stress (negative stress triaxiality) generated during the deformation could postpone the failure, when specimens are formed into a die cavity (EHDF). Also, the increased velocity and significant impact pressure at the final stage of deformation not only prevent strain localization but also help in further suppressing the damage. It is found that very high peak strain rates develop in the sheet as it contacts the die surface which further postpone the failure since AA5182-O exhibits positive strain rate sensitivity at such high-strain rates. Moreover, damage mechanisms of AA5182-O sheets were investigated during EHF tests and are compared with those occur during quasi-static (QS) deformation. The results confirm that void nucleation, growth and coalescence are the main damage mechanisms of AA5182-O at both high and low strain rates. It is found that Mg2Si particles do not significantly influence void formation and the main source of void nucleation is cracking of Al3(Fe-Mn) intermetallic particles. More importantly, it is found that specimens deformed under QS conditions contained more voids in areas away from the sub-fracture surface but EHF specimens exhibit higher rate of void growth close to sub-fracture areas. Optical microscopy results confirmed that void formation is suppressed by increasing the applied energy in EHF. And more importantly, the growth of voids is suppressed due to the high-velocity impact of the sheet against the die which plays an important role in increasing formability of AA5182-O aluminium sheet in EHF. Optical microscopy showed that AA5182-O grains experience significant shearing strain during the EHF deformation in the apex area of conical EHDF specimens. The results of transmission electron microscopy showed that dislocation density increases when specimens are formed using EHF process but the magnitude of this increase is not significantly greater than in quasi-static deformations. Finally, it is concluded that the combination of high strain rate deformation and compressive through-thickness stress during the deformation, leads to formability improvement of AA5182-O in EHDF.

Available for download on Wednesday, January 30, 2019

Share

COinS