Linking processing parameters with melt pool properties of multiple nickel-based superalloys via high-dimensional Gaussian process regression

Abstract

A physics-based model is used to predict the melt pool properties in the laser-directed energy deposition of several nickel-based superalloys for different process parameters. The input space is high-dimensional, consisting of a common 19-dimensional composition space for each alloy and the process parameters (laser power and scan velocity). Gaussian Process-based regression frameworks are developed by training surrogates on data generated by a validated analytical model. These surrogates are thereafter used to predict and define relationships between the composition, resultant thermophysical properties, process parameters, and the subsequent melt pool property. The probabilistic predictions are augmented by uncertainty quantification and sensitivity analysis to substantiate the findings further.

Keywords

Laser directed energy deposition / Gaussian process / nickel-based superallloys / melt pool properties / uncertainty quantification / sensitivity analysis

References

[1] Jelvani S,Barekat M.Empirical-statistical modeling and prediction of geometric characteristics for laser-aided direct metal deposition of inconel 718 superalloy.Met Mater Int2020;26:668-81

[2] Dinda G,Mazumder J.Laser aided direct metal deposition of Inconel 625 superalloy: microstructural evolution and thermal stability.Mater Sci Eng A2009;509:98-104

[3] Zhong C,Gasser A,Schleifenbaum J.The influence of the powder stream on high-deposition-rate laser metal deposition with inconel 718.Metals2017;7:443

[4] Sreekanth S,Hurtig K,Andersson J.Effect of direct energy deposition process parameters on single-track deposits of alloy 718.Metals2020;10:96

[5] Gibson I,Stucker B. Additive manufacturing technologies. Cham, Switzerland: Springer, 2021.

[6] Eagar TW.Temperature fields produced by traveling distributed heat sources.Weld J1983;62:346-55

[7] Feenstra D,Birbilis N.Utilisation of artificial neural networks to rationalise processing windows in directed energy deposition applications.Mater Design2021;198:109342

[8] Juhasz M. Machine learning predictions of single clad geometry in directed energy deposition. Available from: https://osf.io/vxdmz/ [Last accessed on 24 Mar 2023]

[9] Akbari P,Kao N.MeltpoolNet: melt pool characteristic prediction in Metal Additive Manufacturing using machine learning.Addit Manuf2022;55:102817

[10] Ye J,Hoseinnezhad R.Predictions of in-situ melt pool geometric signatures via machine learning techniques for laser metal deposition.Int J Comput Integr Manuf2022;Online ahead of print:1-17

[11] Mondal S,Ray A.Investigation of melt pool geometry control in additive manufacturing using hybrid modeling.Metals2020;10:683

[12] Menon N,Basak A.Multi-fidelity surrogate-based process mapping with uncertainty quantification in laser directed energy deposition.Materials2022;15:2902 PMCID:PMC9025529

[13] Lawrence JR. Laser surface engineering: processes and applications. Elsevier; 2014.

[14] Basak A.Additive manufacturing of nickel-base superalloy IN100 through scanning laser epitaxy.JOM2018;70:53-9

[15] Angel NM.On the fabrication of metallic single crystal turbine blades with a commentary on repair via additive manufacturing.JMMP2020;4:101

[16] Graybill B,Malawey D,Alvarado-Orozco JM.Additive manufacturing of nickel-based superalloys.Am Soc Mech Eng2018;51357:V001T01A015

[17] Wang Z,Li W.A high-throughput approach to explore the multi-component alloy space: a case study of nickel-based superalloys.J Alloys Compd2021;858:158100

[18] Qin Z,Wang Y.Phase prediction of Ni-base superalloys via high-throughput experiments and machine learning.Mater Res Lett2021;9:32-40

[19] Xiong W.Additive manufacturing as a tool for high-throughput experimentation.J Mater Inf2022;2:12

[20] Lookman T,Xue D.Active learning in materials science with emphasis on adaptive sampling using uncertainties for targeted design.NPJ Comput Mater2019;5:21

[21] Zhang Y.Lattice misfit predictions via the gaussian process regression for Ni-based single crystal superalloys.Met Mater Int2021;27:235-53

[22] Yu J,Pan S.Machine learning-guided design and development of metallic structural materials.J Mater Inf2021;1:9

[23] Mukherjee T,De A.Dimensionless numbers in additive manufacturing.J Appl Phys2017;121:064904

[24] Wang W,Chai Z.Big data-assisted digital twins for the smart design and manufacturing of advanced materials: from atoms to products.J Mater Inf2022;2:1

[25] Sente Software - JMatPro®. Available from: https://www.sentesoftware.co.uk/jmatpro [Last accessed on 24 Mar 2023]

[26] Pollock TM.Nickel-based superalloys for advanced turbine engines: chemistry, microstructure and properties.J Propuls Power2006;22:361-74

[27] Cannon Muskegon. Available from: https://cannonmuskegon.com/ [Last accessed on 24 Mar 2023]

[28] Rosenthal D.The theory of moving sources of heat and its application to metal treatments.Trans ASME1946;68:849-65

[29] Lee YS,Babu SS. Influence of fluid convection on weld pool formation in laser cladding. Available from: https://www.researchgate.net/profile/Yousub-Lee-2/publication/268278864_Influence_of_Fluid_Convection_on_Weld_Pool_Formation_in_Laser_Cladding/links/549af80b0cf2b80371371766/Influence-of-Fluid-Convection-on-Weld-Pool-Formation-in-Laser-Cladding.pdf [Last accessed on 31 Mar 2023]

[30] Rasmussen CE. Gaussian processes in machine learning. In: Bousquet O, von Luxburg U, Rätsch G, editors. Advanced Lectures on Machine Learning. Berlin: Springer Berlin Heidelberg; 2004. pp. 63-71.

[31] Fletcher R. Practical methods of optimization. John Wiley & Sons; 2013.

[32] Duvenaud D. Automatic model construction with Gaussian processes. University of Cambridge; 2014.

[33] Le Q,Smola A. Fastfood-approximating kernel expansions in loglinear time. Available from: http://proceedings.mlr.press/v28/le13-supp.pdf [Last accessed on 24 Mar 2023]

[34] GPy by SheffieldML. Available from: https://sheffieldml.github.io/GPy/ [Last accessed on 24 Mar 2023]

[35] Pedregosa F,Gramfort A. Scikit-learn: machine learning in python. Available from: https://www.jmlr.org/papers/volume12/pedregosa11a/pedregosa11a.pdf?ref=https:/ [Last accessed on 31 Mar 2023]

[36] Breusch TS.A simple test for heteroscedasticity and random coefficient variation.Econometrica1979;47:1287

[37] Wang C.Gaussian process regression with heteroscedastic or non-Gaussian residuals.arXiv preprint arXiv:1212.62462012

[38] Sobol′ I.Global sensitivity indices for nonlinear mathematical models and their Monte Carlo estimates.Math Comput Simul2001;55:271-80

[39] Herman J.SALib: an open-source python library for sensitivity analysis.J Open Source Softw2017;2:97

[40] Saltelli A,Andres T. Global sensitivity analysis: the primer. John Wiley & Sons; 2008.

[41] Zieliñska M,Porêba M. Thermal properties of cast nickel based superalloys. Available from: https://www.researchgate.net/profile/Jan-Sieniawski/publication/50302351_Thermal_properties_of_cast_nickel_based_superalloys/links/543fade30cf21227a11a8218/Thermal-properties-of-cast-nickel-based-superalloys.pdf [Last accessed on 31 Mar 2023]

[42] Chen X,Fox P.Direct laser remelting of iron with addition of boron.Mater Sci Technol2004;20:715-9

[43] Ahmed SH.Influence of material property variation on computationally calculated melt pool temperature during laser melting process.Metals2019;9:456