GraphCompNet: A Position-Aware Model for Predicting and Compensating Shape Deviations in 3D Printing

• DOI: 10.48550/arxiv.2502.09652
• Publication Date: 2025-02-11
• Authors: Lei Lei; Chen Chen; Juheon Lee; Juan Carlos Catana; Tsegai O. Yhdego; Nathan Moroney; Mohammad Amin Nabian; Hui Wang; Jun Zeng
• Keywords: Position (finance)

This paper introduces a data-driven algorithm for modeling and compensating shape deviations in additive manufacturing (AM), addressing challenges geometric accuracy batch production. While traditional methods, such as analytical models metrology, laid the groundwork precision, they are often impractical large-scale Recent advancements machine learning (ML) have improved compensation but issues remain generalizing across complex geometries adapting to position-dependent variations. We present novel approach powder bed fusion (PBF) processes, using GraphCompNet, which is computational framework combining graph-based neural networks with generative adversarial network (GAN)-inspired training process. By leveraging point cloud data dynamic graph convolutional (DGCNNs), GraphCompNet shapes incorporates position-specific thermal mechanical factors. A two-stage procedure iteratively refines compensated designs via compensator-predictor architecture, offering real-time feedback optimization. Experimental validation diverse positions shows significantly improves (35 65 percent) entire print space, work advances development of Digital Twin technology AM, enabling scalable, monitoring compensation, critical gaps AM process control. The proposed method supports high-precision, automated industrial-scale design systems.