Title: Automated In-Line Monitoring of Fibre Paths in Complex Filament Wound Composite Structures Using Multi-Modal Deep Learning

Authors: Christopher Mimra, Tristan Shelley, Philip Teakle, Peter Schubel

DOI:

Abstract: Filament winding is a highly automated process for manufacturing rotationally symmetric composite components, yet its application to complex geometries remains limited due to fibre path planning challenges and slippage during winding. This research addresses the problem of monitoring fibre placement in real-time for complex parts, where manual inspection is insufficient. Existing systems often fail to detect individual tows due to the low contrast between the different layers of fibres. Furthermore, tracking fibre paths that are partially obscured by subsequent layers is challenging with traditional algorithms. We propose a novel machine learning-based monitoring system that integrates camera imagery and laser triangulation data within a convolutional neural network. Experiments were conducted that investigate different network architectures regarding their ability to fuse the different input modes. The final system produces a segmented output image that illustrates local fibre orientation across the layup with less than 4° error. It enables direct in-line comparison between planned and actual fibre paths. We further outline a direct pipeline to updated simulations to assess the structural impact of deviations to the final part. The system contributes to smart manufacturing by advancing real-time monitoring and data-driven quality assurance in automated composite fabrication.

References: [1] [2] [3] [4] [5] [6] [7] [8] [9] Carosella S, Hügle S, Helber F, Middendorf P. A short review on recent advances in automated fiber placement and filament winding technologies. Composites Part B: Engineering. 2024;287:111843. DOI: 10.1016/j.compositesb.2024.111843 Zhao Y. Defects in filament wound composite products: Experimental and numerical quantification of fibre misalignment and voids: KU Leuven; 2023. URL: https://lirias.kuleuven.be/4087293?limo=0 Sofi T, Neunkirchen S, Schledjewski R. Path calculation, technology and opportunities in dry fiber winding: a review. Advanced Manufacturing: Polymer & Composites Science. 2018;4:57-72. DOI: 10.1080/20550340.2018.1500099 Pooja K, Mikkelsen LP, Dahl AB, Dahl VA. FibreTracker: Open-source 3D Python Tool for Fibre Tracking. 21st European Conference on Composite Materials: European Society for Composite Materials; 2024. p. 1190-6. Ma Q, Rejab M, Azeem M, Idris M, Rani M, Kumar AP. Axial and radial crushing behaviour of thin-walled carbon fiber-reinforced polymer tubes fabricated by the real-time winding angle measurement system. Forces in Mechanics. 2023;10. DOI: 10.1016/j.matpr.2022.09.228 Miene A, Göttinger M, Herrmann AS. Quality assurance by digital image analysis for the preforming and draping process of dry carbon fiber material. SAMPE EUROPE International Conference and Forum2008. Heieck F. Qualitätsbewertung von Faser-Kunststoff-Verbunden mittels optischer Texturanalyse auf 3D-Preformoberflächen: Universität Stuttgart; 2019. DOI: 10.18419/OPUS-10677 Jois KC. Process-Induced Variations and Their Impact on Structural Properties of Fibre Reinforced Pressure Vessels [Doctoral]: RWTH Aachen; 2024. DOI: 10.18154/RWTH2024-10296. Krombholz C, Bock M, Perner M, Röstermundt D, Meyer M. Online bahnkorrektur eines industrieroboters mittels optischer sensoren für den einsatz im fiber-placement-prozess. Deutscher Luft-und Raumfahrtkongress 2011. [10] Schmitt R, Witte A. Control of a thermoplastic tape winding process with optical in-line metrology. Proceedings in Manufacturing Systems. 2012;7. [11] Brabandt D, Lanza G. Data Processing for an Inline Measurement of Preforms in the CFRP-Production. Procedia CIRP. 2015;33:269-74. DOI: 10.1016/j.procir.2015.06.048 [12] Hopmann C, Magura N, Rozo Lopez N, Schneider D, Fischer K. Detection and evaluation of the fibers' deposition parameters during wet filament winding. Polymer Engineering & Science. 2021;61:1353-67. DOI: 10.1002/pen.25649 [13] Schmidt C, Hocke T, Denkena B. Artificial intelligence for non-destructive testing of CFRP prepreg materials. Production Engineering. 2019;13:617-26. DOI: 10.1007/s11740019-00913-3 11 [14] Denkena B, Schmidt C, Völtzer K, Hocke T. Thermographic online monitoring system for Automated Fiber Placement processes. Composites Part B: Engineering. 2016;97:239-43. DOI: 10.1016/j.compositesb.2016.04.076 [15] Scharr H. Optimale Operatoren in der digitalen Bildverarbeitung [Dissertation]: RuprechtKarls-Universitat Heidelberg; 2000. DOI: 10.11588/heidok.00000962 [16] Ronneberger O, Fischer P, Brox T. U-Net: Convolutional Networks for Biomedical Image Segmentation. arXiv [csCV]. 2015. DOI: 10.48550/ARXIV.1505.04597 [17] Mimra C. Data-Driven Quality Assurance of the Dry Fibre Tape Laying Process [Dissertation]. Stuttgart: University of Stuttgart; 2025. DOI: 10.18419/opus-17229

Conference: SAMPE 2026

Publication Date: 2026/04/27

SKU: 26

Pages: 12

Price: $24.00