A K-Means Clustering Approach for Accelerated Path Planning in GMA-DED: The Fast Advanced-Pixel Strategy

The performance of Gas Metal Arc-Directed Energy Deposition (GMA-DED) strongly depends on efficient path-planning strategies that balance trajectory quality and computational cost. With the purpose of developing a computationally faster and more scalable path-planning approach, this study introduces the Fast Advanced-Pixel strategy by integrating the K-means clustering algorithm into to the Advanced Pixel strategy version to reduce the dimensionality of an optimization problem. Computational validation was conducted on four geometrically distinct parts using different clustering configurations. Statistical analysis (ANOVA) was applied to assess the significance of the results. The findings revealed that by increasing the number of clusters, computational time is substantially reduced, achieving up to a twenty-fold improvement compared with the former strategy, while maintaining consistent trajectory quality. Experimental validation using complex parts, such as a “Jaw Gripper” and a “C-frame” of a resistance spot welding gun, confirmed defect-free deposition and dimensional agreement with the CAD models. Accordingly, within the scope of GMA-DED technology and pixel-based path-planning strategies, the Fast Advanced-Pixel approach demonstrates a significant improvement in computational efficiency while preserving trajectory quality, enabling the accurate and reliable fabrication of geometrically complex metallic parts.