Search for: All records

Understanding the fatigue behavior and accurately predicting the fatigue life of laser powder bed fusion (L-PBF) parts remain a pressing challenge due to complex failure mechanisms, time-consuming tests, and limited fatigue data. This study proposes a physics-informed data-driven framework, a multimodal transfer learning (MMTL) framework, to understand process-defect-fatigue relationships in L-PBF by integrating various modalities of fatigue performance, including process parameters, XCT-inspected defects, and fatigue test conditions. It aims to leverage a pre-trained model with abundant process and defect data in the source task to predict fatigue life nondestructive with limited fatigue test data in the target task. MMTL employs a hierarchical graph convolutional network (HGCN) to classify defects in the source task by representing process parameters and defect features in graphs, thereby enhancing its interpretability. The feature embedding learned from HGCN is then transferred to fatigue life modeling in neural network layers, enabling fatigue life prediction for L-PBF parts with limited data. MMTL validation through a numerical simulation and real-case study demonstrates its effectiveness, achieving an F1-score of 0.9593 in defect classification and a mean absolute percentage log error of 0.0425 in fatigue life prediction. MMTL can be extended to other applications with multiple modalities and limited data. 

ABSTRACT This study utilizes linear elastic fracture mechanics to assess the fatigue criticality of volumetric defects in notched specimens with varying geometries. Contrasting to the existing literature, this study assesses the fatigue criticality of defects, prior to fracture, via a non‐destructive inspection technique, that is, X‐ray computed tomography (XCT). Treating volumetric defects as cracks, based on Murakami's definition, the approach calculates their Mode‐I stress intensity factor (SIF) with their local stresses obtained via linear elastic finite element analysis and utilizes the SIF to represent their criticality. For validation, cylindrical and flat specimens with notch root radii of 5 and 50 mm of AlSi10Mg and 17‐4 precipitation hardened stainless steel were fabricated, XCT scanned, and tested under fatigue loading. All crack initiating defects, observed from fractography, fell within the 99.3 percentile of the defects with the highest stress intensity factor in the respective specimens. 

Abstract Light‐based additive manufacturing methods are widely used to print high‐resolution 3D structures for applications in tissue engineering, soft robotics, photonics, and microfluidics, among others. Despite this progress, multi‐material printing with these methods remains challenging due to constraints associated with hardware modifications, control systems, cross‐contamination, waste, and resin properties. Here, a new printing platform coined Meniscus‐enabled Projection Stereolithography (MAPS) is reported, a vat‐free method that relies on generating and maintaining a resin meniscus between a crosslinked structure and bottom window to print lateral, vertical, discrete, or gradient multi‐material 3D structures with no waste and user‐defined mixing between layers. MAPS is compatible with a wide range of resins shown and can print complex multi‐material 3D structures without requiring specialized hardware, software, or complex washing protocols. MAPS's ability to print structures with microscale variations in mechanical stiffness, opacity, surface energy, cell densities, and magnetic properties provides a generic method to make advanced materials for a broad range of applications.