PSI - Issue 42

Available online at www.sciencedirect.com Available online at www.sciencedirect.com Available online at www.sciencedirect.com

ScienceDirect

Procedia Structural Integrity 42 (2022) 943–951 Structural Integrity Procedia 00 (2019) 000–000 Structural Integrity Procedia 00 ( 01 ) 000–000

www.elsevier.com / locate / procedia www.elsevier.com / locate / procedia

© 2022 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0) Peer-review under responsibility of the scientific committee of the 23 European Conference on Fracture – ECF23 © 2020 The Authors. Published by Elsevier B.V. T is is an open access article under the CC BY- C-ND license (http: // cr ativec mmons.org / licenses / by-nc-nd / 4.0 / ) P r-re ie under responsibility of 23 European Conference on F acture – ECF23 . Keywords: Artificial Intelligence, Neural Networks, Additive Manufacturing, Manufacturing Defects, UNet, resnet Four rectangular blocks of 316L stainless steel were fabricated using Selective Laser Melting (SLM) and Direct Lased Deposi tion (DED) techniques. The blocks were cut longitudinally, and the surfaces were metallographically polished and imaged using Scanning Electron Microscope (SEM). Deep neural network models were constructed based on three architectures: Fast Fully Convolutional Network (FCN), SegNet and UNet. The models were then trained using a set of micrographs utilizing various sets of hyperparameters. Model testing and validation were conducted using new and unseen images. Performance of the models were assessed based on training run times, testing run times and accuracy of detecting the correct locations, types, and sizes of the defects in the validation datasets. © 2020 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http: // creativecommons.org / licenses / by-nc-nd / 4.0 / ) Peer-review under responsibility of 23 European Conference on Fracture – ECF23 . Keywords: Artificial Intelligence, Neural Networks, Additive Manufacturing, Manufacturing Defects, UNet, resnet Abstract This work is part of a large package of work investigating the structural integrity of additively manufactured components in nuclear applications. Components manufactured using additive manufacturing (AM) often contain high degree of manufacturing defects which may lead to pre-mature failure in service conditions, for example, during fatigue loading. The aim of this work is to automate the detection and analysis of various types of manufacturing defects in AM using image segmentation techniques such that, robust predictive models can be built to correlate the defects with material deformation. The types of defects considered in this work are lack of fusion, micro-crack, spherical porosity, and unmelted particles. Four rectangular blocks of 316L stainless steel were fabricated using Selective Laser Melting (SLM) and Direct Lased Deposi tion (DED) techniques. The blocks were cut longitudinally, and the surfaces were metallographically polished and imaged using Scanning Electron Microscope (SEM). Deep neural network models were constructed based on three architectures: Fast Fully Convolutional Network (FCN), SegNet and UNet. The models were then trained using a set of micrographs utilizing various sets of hyperparameters. Model testing and validation were conducted using new and unseen images. Performance of the models were assessed based on training run times, testing run times and accuracy of detecting the correct locations, types, and sizes of the defects in the validation datasets. Abstract This work is part of a large package of work investigating the structural integrity of additively manufactured components in nuclear applications. Components manufactured using additive manufacturing (AM) often contain high degree of manufacturing defects which may lead to pre-mature failure in service conditions, for example, during fatigue loading. The aim of this work is to automate the detection and analysis of various types of manufacturing defects in AM using image segmentation techniques such that, robust predictive models can be built to correlate the defects with material deformation. The types of defects considered in this work are lack of fusion, micro-crack, spherical porosity, and unmelted particles. 23 European Conference on Fracture – ECF23 Application of Deep Learning models to characterize manufacturing defects in additive manufactured components Satyajit Dey a,b, ∗ , Zhijin Lyu b , Gauri Mahalle b , Anas Achouri c , Abdullah Al Mamun b a EDF Energy, Barnet Way, Gloucester GL4 3RS, United Kingdom 23 European Conference on Fracture – ECF23 Application of Deep Learning models to characterize manufacturing defects in additive manufactured components Satyajit Dey a,b, ∗ , Zhijin Lyu b , Gauri Mahalle b , Anas Achouri c , Abdullah Al Mamun b a EDF Energy, Barnet Way, Gloucester GL4 3RS, United Kingdom b Nuclear Futures Institute, Bangor University, Bangor Gwynedd LL57 2DG, United Kingdom c DONAA limited, 3 The Quadrant, Warwick road, Coventry, CV1 2DY, United Kingdom b Nuclear Futures Institute, Bangor University, Bangor Gwynedd LL57 2DG, United Kingdom c DONAA limited, 3 The Quadrant, Warwick road, Coventry, CV1 2DY, United Kingdom

1. Introduction 1. Introduction

Additive manufacturing (AM) of metallic components o ff ers several advantages over conventional manufacturing techniques, such as reduction of the production time, cost of tooling, material waste and a potential increase in the design flexibility for bespoke engineering parts [1]. Components manufactured using AM often contains high degree Additive manufacturing (AM) of metallic components o ff ers several advantages over conventional manufacturing techniques, such as reduction of the production time, cost of tooling, material waste and a potential increase in the design flexibility for bespoke engineering parts [1]. Components manufactured using AM often contains high degree

∗ Corresponding author. Tel.: + 44 1248382356 ; fax: + 44 1248 382330. E-mail address: s.dey@bangor.ac.uk ∗ Corresponding author. Tel.: + 44 1248382356 ; fax: + 44 1248 382330. E-mail address: s.dey@bangor.ac.uk

2452-3216 © 2022 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0) Peer-review under responsibility of the scientific committee of the 23 European Conference on Fracture – ECF23 10.1016/j.prostr.2022.12.119 2210-7843 © 2020 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http: // creativecommons.org / licenses / by-nc-nd / 4.0 / ) Peer-review under responsibility of 23 European Conference on Fracture – ECF23 . 2210-7843 © 2020 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-N -ND license (http: // creativecommons.org / licenses / by-nc-nd / 4.0 / ) Peer-review under responsibility of 23 European Conference on Fracture – ECF23 .