Browsing by Author "Nguyen, Trung"

Now showing 1 - 2 of 2

Computational Homogenization: An Approach to Computational Multi-scale Modeling of Mechanical Metamaterials
(2024-04-26) Nguyen, Trung
Insinööritieteiden korkeakoulu | Bachelor's thesis
Metamaterials have played a crucial part in the development of materials science in recent years due to their unnatural novel properties. However, studying these properties using traditional materials modeling is both time-consuming and computationally expensive, and the accuracy is, most of the time, dependent on the scale of the simulation. Computational homogenization is a multi-scale method which aims at addressing this shortcoming: the properties of the metamaterials can be determined by analyzing the known artificial structures of the material on the micro-scale. When applying to elasticity problems of first order, computational homogenization consists of two main parts: solving the localization problem and determining the effective properties. In the first part, a localization problem is defined to determine the deformation of a single unit cell inside the material. Here, periodic boundary condition is implemented in order to preserve the periodicity of the structure and the homogeneity of the deformation. The deformation map at micro-scale is then transferred back to macro-scale by utilizing the averaging quantities: one of which is the Hill-Mandel lemma stating that the average energy at micro-scale is equivalent to that at the macro-scale. Finally, the deformation map at macro-scale can be evaluated to determine the macroscopic properties of the material. In this thesis, an analysis of triangular truss and frame lattices is conducted. In the triangular truss lattice where the bar members are subjected to axial loading, the local strain energy is determined by the direct stiffness method, and the elastic tensor is derived using Hill-Mandel lemma. The Young's modulus of this structure is dependent on the relative density of the cell, whereas the Poisson's value is invariant. In the triangular frame lattice where the beam members are subjected to additional shear and bending effects, the direct stiffness method also takes into account the slope-deflection equations for these flexural effects. Therefore, the elastic tensor of frame lattice comprises an extra term which is dependent on both the relative density of the unit cell and the geometry of the beam members, and this term describes the flexural properties of the structure. Both truss and frame lattices are isotropic, and the null angular displacement shows that these structures are stretching-dominated.
Enhancing early-stage hydrogen damage detection: Simulation studies on different total focusing method algorithms in ultrasonic evaluation
(2024-08-19) Nguyen, Trung
Insinööritieteiden korkeakoulu | Master's thesis
Steel structures play a crucial role in various aspects of hydrogen applications, composed of containers, pipelines, and storage tanks for transporting and storing hydrogen. However, the complex interaction between hydrogen and steel, especially in some specific operational conditions such as high temperature and high pressure, introduces several challenges that can compromise the integrity and safety of these structures. High Temperature Hydrogen Attack (HTHA) is a well-known mechanism affecting carbon steel and low-alloy steel in the petrochemical and refining industries, which turns into a more severe problem for hydrogen applications. With the development and implementation of the Total Focusing Method (TFM) with Full Matrix Capture (FMC), the capabilities of the ultrasonic testing technique have been significantly improved, providing more precise imaging and characterization of defects. This thesis evaluates and compares the performance of four different TFM algorithms, through simulation, in detecting early-stage HTHA. All simulations are conducted with a steel block containing two separate regions of inclusions and voids. The results are reconstructed images using TFM algorithms which include Delay-and-Sum (DAS), Coherence Factor (CF), Sign Coherence Factor (SCF), and Phase Coherence Factor (PCF). The results reveal that a longitudinal wave at 7.5 MHz with DAS or PCF algorithm provides the clearest image of HTHA voids. 5 MHz longitudinal wave generally yields a higher maximum amplitude. However, this makes it more difficult to detect stage 2 damage due to the damage’s low amplitude and noise from inclusions. For shear wave signals post-processed with the same algorithm, the reconstructed images tend to provide more information about smaller voids compared to those using longitudinal waves. In contrast, shear wave mode suffers significantly from inclusion noise, leading to challenges in the distinction between inclusion signals and actual damage signals. This comparative analysis of TFM algorithms demonstrates their potential in detecting early-stage HTHA, each with its strengths and limitations. The choice of algorithm should be selected depending on the specific demands of the inspection task, balancing factors such as resolution, and signal-to-noise ratio.