Deep Representation Learning for Process Variation Management in Laser Powder Bed Fusion

2021 ◽  
pp. 101961
Author(s):  
Sepehr Fathizadan ◽  
Feng Ju ◽  
Yan Lu
Keyword(s):  
Process Variation ◽  
Representation Learning ◽  
Powder Bed Fusion ◽  
Laser Powder Bed Fusion ◽  
Powder Bed ◽  
Variation Management

scholarly journals Deep semi-supervised learning of dynamics for anomaly detection in laser powder bed fusion

2021 ◽  
Author(s):  
Sebastian Larsen ◽  
Paul A. Hooper
Keyword(s):  
High Speed ◽  
State Space Model ◽  
Representation Learning ◽  
Complex Data ◽  
Powder Bed Fusion ◽  
Laser Powder Bed Fusion ◽  
Powder Bed ◽  
Roc Area ◽  
Dynamic Signature ◽  
Data Efficiency

AbstractHighly complex data streams from in-situ additive manufacturing (AM) monitoring systems are becoming increasingly prevalent, yet finding physically actionable patterns remains a key challenge. Recent AM literature utilising machine learning methods tend to make predictions about flaws or porosity without considering the dynamical nature of the process. This leads to increases in false detections as useful information about the signal is lost. This study takes a different approach and investigates learning a physical model of the laser powder bed fusion process dynamics. In addition, deep representation learning enables this to be achieved directly from high speed videos. This representation is combined with a predictive state space model which is learned in a semi-supervised manner, requiring only the optimal laser parameter to be characterised. The model, referred to as FlawNet, was exploited to measure offsets between predicted and observed states resulting in a highly robust metric, known as the dynamic signature. This feature also correlated strongly with a global material quality metric, namely porosity. The model achieved state-of-the-art results with a receiver operating characteristic (ROC) area under curve (AUC) of 0.999 when differentiating between optimal and unstable laser parameters. Furthermore, there was a demonstrated potential to detect changes in ultra-dense, 0.1% porosity, materials with an ROC AUC of 0.944, suggesting an ability to detect anomalous events prior to the onset of significant material degradation. The method has merit for the purposes of detecting out of process distributions, while maintaining data efficiency. Subsequently, the generality of the methodology would suggest the solution is applicable to different laser processing systems and can potentially be adapted to a number of different sensing modalities.

scholarly journals Process variation in Laser Powder Bed Fusion of Ti-6Al-4V

2021 ◽  
pp. 101987
Author(s):  
Zhuoer Chen ◽  
Xinhua Wu ◽  
Chris H.J. Davies
Keyword(s):  
Process Variation ◽  
Powder Bed Fusion ◽  
Laser Powder Bed Fusion ◽  
Powder Bed

Near-Threshold Fatigue Crack Growth Rates of Laser Powder Bed Fusion Produced Ti-6Al-4V

2020 ◽  
Author(s):  
Thorsten Hermann Becker ◽  
Nur Mohamed Dhansay ◽  
Gerrit Matthys Ter Haar ◽  
Kim Vanmeensel
Keyword(s):  
Fatigue Crack ◽  
Fatigue Crack Growth ◽  
Crack Growth ◽  
Growth Rates ◽  
Powder Bed Fusion ◽  
Laser Powder Bed Fusion ◽  
Powder Bed ◽  
Crack Growth Rates ◽  
Near Threshold

scholarly journals On the Relevance of Volumetric Energy Density in the Investigation of Inconel 718 Laser Powder Bed Fusion

Materials ◽  
2020 ◽  
Vol 13 (3) ◽  
pp. 538 ◽  
Author(s):  
Fabrizia Caiazzo ◽  
Vittorio Alfieri ◽  
Giuseppe Casalino
Keyword(s):  
Surface Roughness ◽  
Energy Density ◽  
Process Parameters ◽  
Vickers Hardness ◽  
Powder Bed Fusion ◽  
Processing Conditions ◽  
Laser Powder Bed Fusion ◽  
Powder Bed ◽  
Fractional Density ◽  
Experimental Variation

Laser powder bed fusion (LPBF) can fabricate products with tailored mechanical and surface properties. In fact, surface texture, roughness, pore size, the resulting fractional density, and microhardness highly depend on the processing conditions, which are very difficult to deal with. Therefore, this paper aims at investigating the relevance of the volumetric energy density (VED) that is a concise index of some governing factors with a potential operational use. This paper proves the fact that the observed experimental variation in the surface roughness, number and size of pores, the fractional density, and Vickers hardness can be explained in terms of VED that can help the investigator in dealing with several process parameters at once.

Heat source model calibration for thermal analysis of laser powder-bed fusion

2020 ◽  
Vol 106 (7-8) ◽  
pp. 3367-3379 ◽  
Author(s):  
Shahriar Imani Shahabad ◽  
Zhidong Zhang ◽  
Ali Keshavarzkermani ◽  
Usman Ali ◽  
Yahya Mahmoodkhani ◽  
...  
Keyword(s):  
Thermal Analysis ◽  
Heat Source ◽  
Model Calibration ◽  
Source Model ◽  
Powder Bed Fusion ◽  
Laser Powder Bed Fusion ◽  
Heat Source Model ◽  
Powder Bed

scholarly journals The Effect of Cu Content and Surface Finish on the Metal Dusting Resistance of Additively Manufactured NiCu Alloys

2021 ◽  
Author(s):  
Katrin Jahns ◽  
Anke S. Ulrich ◽  
Clara Schlereth ◽  
Lukas Reiff ◽  
Ulrich Krupp ◽  
...  
Keyword(s):  
Surface Finish ◽  
Powder Bed Fusion ◽  
Metal Dusting ◽  
Laser Powder Bed Fusion ◽  
Roughness Measurement ◽  
Powder Bed ◽  
Cu Content ◽  
Binary Model ◽  
Interesting Candidate ◽  
Nicu Alloys

AbstractDue to the inhibiting behavior of Cu, NiCu alloys represent an interesting candidate in carburizing atmospheres. However, manufacturing by conventional casting is limited. It is important to know whether the corrosion behavior of conventionally and additively manufactured parts differ. Samples of binary NiCu alloys and Monel Alloy 400 were generated by laser powder bed fusion (LPBF) and exposed to a carburizing atmosphere (20 vol% CO–20% H2–1% H2O–8% CO2–51% Ar) at 620 °C and 18 bar for 960 h. Powders and printed samples were investigated using several analytic techniques such as EPMA, SEM, and roughness measurement. Grinding of the material after building (P1200 grit surface finish) generally reduced the metal dusting attack. Comparing the different compositions, a much lower attack was found in the case of the binary model alloys, whereas the technical Monel Alloy 400 showed a four orders of magnitude higher mass loss during exposure despite its Cu content of more than 30 wt%.

Sign in / Sign up

Export Citation Format

Share Document