A deep learning-based model for defect detection in laser-powder bed fusion using in-situ thermographic monitoring
- PDF / 931,864 Bytes
- 9 Pages / 595.276 x 790.866 pts Page_size
- 104 Downloads / 180 Views
FULL RESEARCH ARTICLE
A deep learning‑based model for defect detection in laser‑powder bed fusion using in‑situ thermographic monitoring Hermann Baumgartl1 · Josef Tomas2 · Ricardo Buettner1 · Markus Merkel2 Received: 4 September 2019 / Accepted: 24 December 2019 © The Author(s) 2020
Abstract Additive manufacturing of metal components with laser-powder bed fusion is a very complex process, since powder has to be melted and cooled in each layer to produce a part. Many parameters influence the printing process; however, defects resulting from suboptimal parameter settings are usually detected after the process. To detect these defects during the printing, different process monitoring techniques such as melt pool monitoring or off-axis infrared monitoring have been proposed. In this work, we used a combination of thermographic off-axis imaging as data source and deep learning-based neural network architectures, to detect printing defects. For the network training, a k-fold cross validation and a hold-out cross validation were used. With these techniques, defects such as delamination and splatter can be recognized with an accuracy of 96.80%. In addition, the model was evaluated with computing class activation heatmaps. The architecture is very small and has low computing costs, which means that it is suitable to operate in real time even on less powerful hardware. Keywords Quality assurance · Machine learning · Additive manufacturing · Convolutional neural networks
1 Introduction In recent years, additive manufacturing (AM) as an industry has experienced enormous growth. The layer-by-layer production process makes it easy to manufacture products with complex geometries and different materials. The material segment of the industry reported a record growth in 2018, with metal materials in particular increasing by 41.9%. A continuation of a 5-year growth phase of more than 40% per year was recorded in the latest edition of the annual Wohler’s Report [1]. This growth demonstrates the need to find solutions that are better suited for mass manufacturing rather than rapid prototyping. AM has significant potential in the medical field, particularly for custom designs, aerospace, and automotive for lightweight construction and functionally critical parts manufactured locally at distant locations [2, 3]. However, uncertainties regarding component quality currently delay the full introduction of AM technology in these * Hermann Baumgartl Hermann.Baumgartl@hs‑aalen.de 1
Machine Learning Research Group, Aalen University, Beethovenstr. 1, 73430 Aalen, Germany
Institute for Virtual Product Development, Aalen University, Beethovenstr. 1, 73430 Aalen, Germany
2
areas. The implementation of in-situ and real-time process monitoring is necessary to meet the high-quality requirements of these applications, particularly with metal powder. Laser-powder bed fusion (L-PBF) is an AM process that uses metal powder material, which is dispensed in a layerby-layer fashion, to create geometries from digital files. The digital pa
Data Loading...
