Microstructural Shielding and Advanced Fatigue Crack Analysis in the Presence of Deformation Twinning

Author

• Seyed Ali Asli, Clemson UniversityFollow

Date of Award

8-2026

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Mechanical Engineering

Committee Chair/Advisor

Garrett J. Pataky

Committee Member

Huijuan Zhao

Committee Member

Enrique Martinez Saez

Committee Member

Qiushi Chen

Abstract

Design methods to mitigate fatigue failures have remained relatively stagnant, as engineers have relied on existing continuum-level models and adhered to the damage tolerance design methodology. Although damage-tolerant design has led to increased safety, extended component in-service life, and better life predictions, it still relies on maintenance routines to identify dangerous cracks. This research aims to test the hypothesis that bundles of deformation twin boundaries will increase the resistance to fatigue crack growth (FCG). Slip and twinning are two dominant mechanisms of plastic deformation in metals. Slip is more common in materials with face-centered cubic (FCC) crystal structures at room temperature. As the temperature decreases, the role of twinning becomes more significant in the deformation process. Several experimental techniques and methods are used to nucleate deformation twinning, including, but not limited to, heat treatment for grain growth, the use of liquid nitrogen (LN₂) to achieve cryogenic temperatures, and pre-strain experiments for plastic deformation. A few different processes are required to identify deformation twinning, consisting of etching to disclose the grain boundary, electron backscatter diffraction (EBSD) to map grain information, transmission electron microscopy (TEM) to support the evidence of the twin, and X-ray diffraction (XRD) to measure dislocation density. The material chosen for this research is a dog-bone-shaped copper with 7 atomic percent Aluminum (Cu7at%Al), due to its very low stacking fault energy, which makes it comparatively easier to nucleate twins. After running monotonic/quasi-static tests on samples with different conditions, including a virgin sample, an annealed virgin sample, an HT and pre-strained sample, an edge notch was created in the middle of the gage section on the samples by electrical discharge machining (EDM). Then, based on the adequate threshold stress intensity factor, the samples underwent high-cycle fatigue (> 10⁴ cycles) to initiate a crack. The crack tip was monitored both optically and using the digital image correlation (DIC) method. The results show a descending and pause da/dN with a crack direction change behavior in stage I FCG for the deformation twin-induced sample.

To gain a deeper understanding of the reason behind this special behavior, further investigation has been conducted. This study also presents a micromechanical investigation of fatigue crack propagation, in which deformation twins (DTs) serve as potent microstructural barriers. By integrating in-situ high-resolution digital image correlation (HRDIC) with EBSD, the full-field evolution of strain, stress, and displacement around a growing fatigue crack was quantified. An anisotropic fracture mechanics framework, incorporating crystallographic orientation and Hill's plasticity criterion, is employed to extract mixed-mode stress intensity factors (SIFs), T-stress, and energy release rate (ERR) values. The results revealed that crack interaction with DT bundles triggered significant deflection of up to ~48°, accompanied by a ~25% drop in the effective crack driving force and a constriction of the plastic zone. Crack direction changes analysis revealed that energy dissipation accounted for over 60% of the total ERR. Furthermore, slip irreversibility analysis identifies the crack tip region as a locus of accumulated cyclic damage. The microstructural shielding behavior of DTs has been analyzed in this study, quantifying how DT bundles promote tortuosity of the crack path and dissipate energy through anisotropic plastic flow, thereby enhancing fatigue resistance. This work establishes a direct quantitative link between discrete microstructural features and macroscopic fracture parameters, providing a blueprint for designing damage-tolerant metallic materials via microstructural engineering.

One of the challenging processes in fatigue analysis is correctly identifying the crack tip location. This research also presented a novel integrated framework combining DIC techniques with multimodal deep learning (DL) algorithms to automate fatigue crack tip (FCT) detection in brittle and ductile materials. The proposed methodology addresses a significant limitation of traditional fatigue crack tip detection methods by developing a dual-stream convolutional neural network (CNN) architecture that simultaneously considers both displacement and strain fields derived from DIC data. This approach also implemented a sophisticated preprocessing pipeline for image texture enhancement and speckle pattern preservation, with automated region-of-interest (ROI) selection based on geometric contour analysis. The core innovation lies in the multimodal approach that learns from complex relationships between displacement discontinuities and strain concentration regions to accurately detect the fatigue crack tip. The proposed framework has been trained and validated on a comprehensive dataset of more than 450 fatigue crack image sets from various materials and then tested on more than 80 samples. The evaluation analysis demonstrated that the model presented in this research achieves a mean absolute error (MAE) of 13 μm, representing a nearly 63% improvement over the traditional geometric-based method. The 5-Fold cross-validation technique confirmed model’s stability with a robust learning curve across all folds. The framework’s modular design facilitates integration with existing experimental workflows, offering a practical solution for automated fatigue crack monitoring in material testing and structural health assessment applications.

Recommended Citation

Asli, Seyed Ali, "Microstructural Shielding and Advanced Fatigue Crack Analysis in the Presence of Deformation Twinning" (2026). All Dissertations. 4378.
https://open.clemson.edu/all_dissertations/4378