Manufacturing near-net-shape hydrogen storage tanks is essential for meeting sustainability goals and achieving carbon-free flights by 2050. Flow-forming is a sustainable process used to produce thin-walled integrated components with strong mechanical properties and minimal material waste. In this study, type-I hydrogen tanks made of AA6061 alloy were manufactured using a cold flow-forming technique, a near-net-shape manufacturing method. The forming was carried out at three different feed-rates (4.5, 6 and 7.5 mm/rev), followed by an investigation of the mechanical properties and surface roughness of the formed components (particularly the dome). It was found that the dome formed at 6 mm/rev showed better strength and dimensional accuracy. Moreover, a variation in hardness was observed, with the highest value measured at the outer surface of the formed dome. The electron backscatter diffraction (EBSD) maps revealed both elongated and refined grains, along with partly dynamic recovered grains and subgrain coalescence. These regions further led to the formation of strain-free grains. Analysis of the crystallographic texture confirmed that the formed domes experienced predominant compression loading with relatively weaker shear forces. Overall, the process of flow-forming improved the structural integrity of AA6061 components, supporting its application in manufacturing sustainable hydrogen tanks.
The effects of tool temperature, geometry, rotation direction, and traverse path on residual stress, surface roughness, and forming forces in AA6061 sheets processed by Induction Tool-Heated Incremental Sheet Forming (ITHISF) were experimentally investigated. X-ray diffraction analysis revealed predominantly compressive residual stresses, which decreased by approximately 75% as the tool temperature increased from 150 °C to 250 °C. Surface roughness is strongly influenced by tool geometry and kinematics; at 150 °C, a flat-ended tool produced ∼128% higher roughness than a hemispherical tool. Tool rotation and traverse path further affect surface finish; the roughness is higher when the spindle rotates counterclockwise and the tool traverses clockwise at a spindle speed of 2000 revolutions per minute. Forming forces are also sensitive to process parameters, including the direction of rotation relative to tool traverse. For example, in one case, the peak axial force increased by ∼50 N by changing the traverse direction. Further, the peak Fz decreased by ∼100 N as the tool temperature increased from 25 °C to 250 °C. These results demonstrate that appropriate control of thermal and kinematic conditions can effectively regulate residual stress, surface roughness, and forming forces during ITHISF.
Fabrication of Functionally Graded Materials (FGM) in Laser Powder Bed Fusion (LPBF) has been achieved by various researchers through sequential deposition of layers, resulting in step gradation rather than a smooth variation. Alternatively, in-plane functional gradation not only provides a smooth material transition but also offers the opportunity to print multiple combination of materials within the same build platform. However, controlling the selective spreading of powder of specific composition within the spreading plane is challenging. In the current study, a partition-based powder dispenser was developed to address the problem. The discrete element method (DEM) is used to simulate the spreading process so as to help in the design of various recoater geometries eventually to mitigate the known limitations of FGM spreading. In this study specific attention is paid to the performance of 4 different recoater geometries, namely, flat, rake, rake with slots and rake with extended slots. The quality of the resulting graded layer is quantified using a gradation index (GI). Our findings demonstrate that while slotted profiles constrain the particle flow, raketype recoaters provide the optimal balance for layer homogeneity, demonstrating the potential of DEM guided design to achieve controlled and continuous gradation in LPBF.
Accurate prediction of sheet metal deformation under complex tool paths remains a fundamental challenge in computational manufacturing. Finite Element Method (FEM) simulations provide the highest fidelity, yet their computational cost makes them impractical for real-time applications—particularly when fine mesh resolution and small time steps are required to capture localized deformation and contact mechanics. This work presents a MeshGraphNet (MGN)-style Graph Neural Network (GNN) surrogate for metal forming process prediction, demonstrated through a case study on the English wheel process. The approach represents sheet and tool surfaces as graphs, where nodes encode 3D coordinates and part identity, and edges capture both mesh topology and dynamic tool–sheet contact interactions. Trained on displacement fields from high-fidelity FEM simulations, the MGN produces node-level deformation predictions up to approximately four orders of magnitude faster than FEM (milliseconds versus minutes per timestep) while maintaining high spatial accuracy. Across ten leave-one-out cross-validation runs, the model achieves a mean nodal displacement root mean squared error (RMSE) of approximately 0.18mm in deforming sheet regions. The surrogate focuses on forming-stage deformation under active tool–sheet contact, capturing localized plastic flow during rolling passes. Although the model relies on simulation-generated FEM training data and exhibits limited extrapolation beyond trained tool paths, its speed–accuracy balance makes it well-suited for digital twin frameworks, reinforcement learning–based adaptive control, and process parameter optimization in advanced manufacturing workflows.
This study reports an interesting phenomenon in high-intensity multi-pulse nanosecond laser ablation with a high pulse frequency, where the confinement by two sidewalls can obviously reduce the ablation depth even when the laser spot boundary-sidewall distance is more than ∼ 14-times the laser spot radius. Through time-resolved imaging and analysis of the laser ablation process, the likely underlying mechanism for the aforementioned phenomenon has been revealed to be the “confinement-induced laser attenuation” effect. The sidewalls confine lateral flow of the gaseous phase (including the target vapor generated during the ablation by one laser pulse), causing more target particles in the laser path and thus increasing the possibility and/or extent of the particles’ attenuation of the next laser pulse. Such attenuation may reduce the laser-target energy coupling efficiency. The laser-particle interaction may generate a secondary plasma and/or scatter laser light, causing other undesirable effect(s). Thus, the “confinement-induced laser attenuation” effect may need to be considered during the design and analysis of related high-frequency multi-pulse nanosecond laser-based manufacturing processes involving geometrical confinement.
