3d printing

In-situ geometric deformation monitoring during fused deposition modeling (FDM) remains challenging, and conventional multi-view photogrammetry requires numerous cameras for full circumferential coverage. Method: This work introduces an open-source experimental platform featuring a heated build plate integrated with a closed-loop rotational mechanism and minimal-camera photogrammetry, enabling complete 360° in-situ observation of printed parts. The system combines high-precision thermal control (nozzle and bed), environmental parameter sensing, and synchronized multi-camera imaging for dynamic 3D surface reconstruction during printing. Contribution/Results: By co-optimizing rotational substrate motion and photogrammetric acquisition, the platform establishes, for the first time in FDM in-situ monitoring, a quantitative, traceable linkage among process parameters, geometric deformation, and surface defects. Experimental validation demonstrates continuous, sub-millimeter-accurate deformation tracking, providing a reproducible, high spatiotemporal-resolution platform for additive manufacturing process monitoring.

Real-time defect identification and in-situ repair remain challenging in robotic-arm-based additive manufacturing. Method: This paper proposes an adaptive closed-loop system integrating thermal regulation, vision-based inspection, and motion control. Built upon a 6-DOF robotic arm and an IoT-enabled thermal control platform, the system employs ROS2 for synchronized motion–extrusion control; utilizes an Oak-D camera with OpenCV for online interlayer defect detection; and achieves precise re-extrusion via homography transformation and inverse kinematics optimization for visual guidance. Contribution/Results: The system introduces a tightly coupled “perception–decision–execution” architecture enabling millisecond-scale defect response and seamless correction under complex trajectories. Experiments demonstrate >92% accuracy in detecting common defects (e.g., layer misalignment, under-extrusion), with post-repair dimensional error reduced by 67%, significantly enhancing print consistency—particularly beneficial for high-precision applications in aerospace and biomedical manufacturing.

This study addresses the complex relationship between process parameters and part quality in metal alloy additive manufacturing, where traditional trial-and-error approaches are inefficient and costly. The authors propose an AI-driven adaptive experimental design method that integrates domain knowledge, leveraging surrogate modeling and a small-batch active learning strategy to intelligently select a minimal set of candidate parameters for iterative validation. This approach efficiently explores the feasible process window, enabling the first successful fabrication of high-performance Cu-Cr-Nb (GRCop-42) alloy via infrared laser-directed energy deposition. Within three months, multiple defect-free specimens were produced across a wide range of laser powers, significantly accelerating development timelines and reducing reliance on expensive equipment and expert intuition—thereby advancing decentralized manufacturing of critical materials.

This study addresses the challenge in conformal antenna manufacturing where conventional planar fabrication processes fail to simultaneously satisfy aerodynamic shaping, impedance matching, and multimaterial integration. To overcome this, we propose a novel multimaterial additive manufacturing approach leveraging a low-cost, open-source five-axis desktop printer. By co-printing conductive filaments (e.g., silver-doped PLA) with structural polymers and integrating electromagnetic simulation–driven toolpath planning, our method transcends traditional 2.5D printing limitations, enabling direct 3D conformal fabrication of S-band patch and ultra-wideband antennas on curved surfaces. Experimental results demonstrate a 23% improvement in impedance bandwidth, a 65% reduction in prototyping lead time, and ~40% lower per-unit cost. Crucially, this work presents the first experimental validation of five-axis multimaterial 3D printing for GHz-range high-frequency conformal devices, establishing its engineering feasibility and comprehensive advantages.

Operation sequence planning for hybrid additive-subtractive manufacturing (HADD) of arbitrarily shaped parts remains challenging due to geometric complexity and intermediate shape instability. Method: We propose a unified inverse-planning framework that starts from the target model and iteratively applies reversible additive “retraction” and subtractive “undo” operations until reaching an empty shape. The approach employs voxel-based modeling and a scalable search algorithm, with theoretical guarantees of solution existence for any manufacturable geometry. It inherently supports automatic tool switching and scales to large models. Contribution/Results: We validate the algorithm on multiple complex digital models and demonstrate high-precision physical fabrication on a hybrid manufacturing platform. Experimental results confirm both process feasibility and robust structural stability throughout the entire machining sequence—addressing critical challenges in HADD planning and execution.

This work addresses the persistent challenges in fused deposition modeling (FDM) printing—such as poor printability, insufficient mechanical strength, and complex post-processing caused by geometric defects like steep overhangs—by introducing the first end-to-end multi-agent system capable of automatically repairing original CAD models. The proposed framework integrates B-Rep parsing, graph neural network–based semantic recognition, and multimodal large language model reasoning to detect manufacturability issues and generate optimized STEP files along with detailed modification reports. By constructing face adjacency topological graphs, applying GraphSAGE for semantic labeling, leveraging Claude Sonnet for design suggestions, and validating modifications via GPT-4o’s visual reasoning, the system automates the entire pipeline from geometric analysis to natural-language design recommendations. Evaluated on a birdhouse model, it accurately identified overhang regions and effectively proposed corrective strategies such as chamfering, filleting, or part reorientation, substantially overcoming the reliance on manual intervention inherent in traditional design-for-manufacturing approaches.

This work addresses the tight coupling between design intent and printer-specific representations in heterogeneous manufacturing, which hinders cross-platform reuse. The authors propose a novel compiler architecture that models fabrication-aware design as a staged, type-directed lowering process, decoupling source design, attribute translation, and backend compilation to enable manufacturing-agnostic expression. Introducing compiler paradigms to heterogeneous manufacturing for the first time, the approach unifies volumetric information—such as material composition, hardness, and color—through implicit geometry and typed spatial attribute fields, automatically generating voxel stacks, G-code, or slicer projects. Experiments demonstrate successful fabrication of complex objects embedding CT data, Shore hardness fields, and full-color fields on both material jetting and extrusion platforms, validating cross-process reusability. The accompanying Python toolkit is publicly released.

Current two-photon lithography (TPL) systems rely on experience-based maintenance, leading to unplanned downtime or excessive maintenance and severely compromising manufacturing quality and efficiency. To address this, we propose a physics-informed, data-driven health monitoring method. Specifically, we develop a degradation modeling framework based on Physics-Informed Neural Networks (PINNs), jointly trained on experimental data comprising six process parameter combinations and six structural feature dimensions. The framework integrates regression-based prediction with statistical analysis to enable quantitative, multi-scenario health state assessment. Our approach significantly improves accuracy and robustness in health state identification, reliably distinguishing distinct degradation levels. Extensive validation across diverse operational conditions confirms its generalizability and practical applicability. This work establishes an interpretable, deployable paradigm for condition-based intelligent maintenance of TPL systems.

This work addresses the challenge of achieving complex shape transformations and multifunctional integration within a unified technical framework, which has been difficult for conventional shape-changing interfaces. The authors propose a co-design methodology that synergistically combines fused deposition modeling (FDM) 3D/4D printing with heat-sealed pneumatic actuation. By employing an integrated fabrication process, programmable structures and pneumatic components are seamlessly embedded to enable dynamically coupled actuation and constraint. The approach is grounded in a design space defined by four fundamental interaction primitives and supported by a dedicated authoring tool that facilitates end-to-end manufacturing. Multiple application examples demonstrate the method’s effectiveness and feasibility in realizing intricate, controllable deformations alongside functional integration.

To address the challenge of simultaneously achieving programmable deformation, embedded sensing, and low-cost fabrication in soft devices, this paper introduces FluxLab—a unified design and fabrication system for silicone-based 3D-printed shape-changing devices. Methodologically, it integrates actuation, sensing, and structural support within a single physical substrate via a nested multi-layer architecture comprising shape memory alloy (SMA) actuation channels, lattice-based mechanical supports, and helical conductive traces. It further incorporates an interactive deformation editor and inductive sensing for user-defined deformation modeling, alongside a lightweight machine learning classifier enabling real-time, high-accuracy deformation recognition. Fabricated using consumer-grade stereolithography (SLA) printers and elastic silicone resin, FluxLab demonstrates feasibility through functional prototypes—including a steam-cooker clamp, a remote gripper, and an interactive desk lamp—achieving superior manufacturability, sensing accuracy (94.2% mean classification accuracy), and functional versatility.