Toward in-situ flaw detection in laser powder bed fusion additive manufacturing through layerwise imagery and machine learning

Abstract Process monitoring in additive manufacturing may allow components to be certified cheaply and rapidly and opens the possibility of healing defects, if detected. Here, neural networks (NNs) and convolutional neural networks (CNNs) are trained to detect flaws in layerwise images of a build, using labeled XCT data as a ground truth. Multiple images were recorded after each layer before and after recoating with various lighting conditions. Classifying networks were given a single image or multiple images of various lighting conditions for training and testing. CNNs demonstrated significantly better performance than NNs across all tasks. Furthermore, CNNs demonstrated improved generalizability, i.e., the ability to generalize to more diverse data than either the training or validation data sets. Specifically, CNNs trained on high-resolution layerwise images from one build showed minimal loss in performance when applied to data from an independent build, whereas the performance of the NNs degraded significantly. CNN accuracy was also demonstrated to be a function of flaw size, suggesting that smaller flaws may be produced by mechanisms that do not alter the surface morphology of the build plate. CNNs demonstrated accuracies of 93.5 % on large (>200 μm) flaws when testing and training on components from the same build and accuracies of 87.3 % when testing on a previously unseen build. Finally, evidence linking the formation of large lack-of-fusion defects to the presence of process ejecta is presented.

[1]  Edward William Reutzel,et al.  Sensing for directed energy deposition and powder bed fusion additive manufacturing at Penn State University , 2016, SPIE LASE.

[2]  S. Beretta,et al.  A comparison of fatigue strength sensitivity to defects for materials manufactured by AM or traditional processes , 2017 .

[3]  Chee How Wong,et al.  Additive manufacturing process monitoring and control by non-destructive testing techniques: challenges and in-process monitoring , 2018 .

[4]  E. Boller,et al.  In-situ Synchrotron imaging of keyhole mode multi-layer laser powder bed fusion additive manufacturing , 2020 .

[5]  Hui Yang,et al.  A hybrid deep learning model of process-build interactions in additive manufacturing , 2020 .

[6]  S. Coeck,et al.  Prediction of lack of fusion porosity in selective laser melting based on melt pool monitoring data , 2019, Additive Manufacturing.

[7]  J. Clayton Optimising metal powders for additive manufacturing , 2014 .

[8]  J. Petrich,et al.  MACHINE LEARNING FOR DEFECT DETECTION FOR PBFAMUSING HIGH RESOLUTION LAYERWISE IMAGINGCOUPLED WITH POST-BUILD CT SCANS , 2017 .

[9]  Jack Beuth,et al.  Anomaly Detection and Classification in a Laser Powder Bed Additive Manufacturing Process using a Trained Computer Vision Algorithm , 2018 .

[10]  T. DebRoy,et al.  Scientific, technological and economic issues in metal printing and their solutions , 2019, Nature Materials.

[11]  Shashi Phoha,et al.  Application of supervised machine learning for defect detection during metallic powder bed fusion additive manufacturing using high resolution imaging. , 2018 .

[12]  Mamidala Ramulu,et al.  Electron Beam Additive Manufacturing of Titanium Components: Properties and Performance , 2013 .

[13]  J. Beuth,et al.  The role of process variables in laser-based direct metal solid freeform fabrication , 2001 .

[14]  Zenghui Wang,et al.  Deep Convolutional Neural Networks for Image Classification: A Comprehensive Review , 2017, Neural Computation.

[15]  S. Shevchik,et al.  Acoustic emission for in situ quality monitoring in additive manufacturing using spectral convolutional neural networks , 2017 .

[16]  Chandrika Kamath,et al.  Observation of keyhole-mode laser melting in laser powder-bed fusion additive manufacturing , 2014 .

[17]  T. Uchida,et al.  Influence of defects, surface roughness and HIP on the fatigue strength of Ti-6Al-4V manufactured by additive manufacturing , 2018, International Journal of Fatigue.

[18]  Jack Beuth,et al.  Using machine learning to identify in-situ melt pool signatures indicative of flaw formation in a laser powder bed fusion additive manufacturing process , 2019, Additive Manufacturing.

[19]  Kamel Fezzaa,et al.  Keyhole threshold and morphology in laser melting revealed by ultrahigh-speed x-ray imaging , 2019, Science.

[20]  A. Beese,et al.  Review of Mechanical Properties of Ti-6Al-4V Made by Laser-Based Additive Manufacturing Using Powder Feedstock , 2016 .

[21]  Linkan Bian,et al.  From in-situ monitoring toward high-throughput process control: cost-driven decision-making framework for laser-based additive manufacturing , 2019, Journal of Manufacturing Systems.

[22]  Robert W. M. Smith,et al.  Selection and installation of high resolution imaging to monitor the PBFam process, and synchronization to post-build 3D computed tomography , 2020 .

[23]  P J Withers,et al.  The Influence of Porosity on Fatigue Crack Initiation in Additively Manufactured Titanium Components , 2017, Scientific Reports.

[24]  Georg Schlick,et al.  Influence of the shielding gas flow on the removal of process by-products in the selective laser melting process , 2016 .

[25]  Richard Leach,et al.  Review of in-situ process monitoring and in-situ metrology for metal additive manufacturing , 2016 .

[26]  M. Leu,et al.  Materials for Additive Manufacturing , 2020, Additive Manufacturing Technologies.

[27]  Jack Beuth,et al.  Synchrotron-Based X-ray Microtomography Characterization of the Effect of Processing Variables on Porosity Formation in Laser Power-Bed Additive Manufacturing of Ti-6Al-4V , 2017 .

[28]  J. Tomas,et al.  A deep learning-based model for defect detection in laser-powder bed fusion using in-situ thermographic monitoring , 2020, Progress in Additive Manufacturing.

[29]  Annamaria Gisario,et al.  Metal additive manufacturing in the commercial aviation industry: A review , 2019, Journal of Manufacturing Systems.

[30]  Iain Todd,et al.  XCT analysis of the influence of melt strategies on defect population in Ti?6Al?4V components manufactured by Selective Electron Beam Melting , 2015 .

[31]  Tony D. James,et al.  Corrigendum: Macrophage Migration Inhibitory Factor is subjected to glucose modification and oxidation in Alzheimer’s Disease , 2017, Scientific Reports.

[32]  M. Marufuzzaman,et al.  Porosity prediction: Supervised-learning of thermal history for direct laser deposition , 2018 .

[33]  M. Javaid,et al.  Additive manufacturing applications in medical cases: A literature based review , 2018, Alexandria Journal of Medicine.

[34]  Robert X. Gao,et al.  Machine learning-based image processing for on-line defect recognition in additive manufacturing , 2019, CIRP Annals.

[35]  Tim Caffrey,et al.  Wohlers report 2013 : additive manufacturing and 3D printing state of the industry : annual worldwide progress report , 2013 .

[36]  Daniel P. Satko,et al.  Defect distribution and microstructure heterogeneity effects on fracture resistance and fatigue behavior of EBM Ti–6Al–4V , 2017 .

[37]  A. Nassar,et al.  Formation processes for large ejecta and interactions with melt pool formation in powder bed fusion additive manufacturing , 2019, Scientific Reports.

[38]  R. Fabbro Depth Dependence and Keyhole Stability at Threshold, for Different Laser Welding Regimes , 2020, Applied Sciences.