A Supervised Machine Learning Model for Regression to Predict Melt Pool Formation and Morphology in Laser Powder Bed Fusion

Aug 4, 2026

Upload by Eliana

Laser powder bed fusion (L-PBF) requires careful optimization of laser power and scanning speed to achieve target quality, productivity, and build-volume goals while controlling conduction zones. A supervised machine learning workflow is proposed to regress melt pool dimensions as a function of the P/V parameter combination, accelerating construction of printability maps and selection of promising P-V configurations. Experimental validation uses Inconel 718 samples, incorporating effects of layer thickness (30–90 µm) and substrate preheating temperature. Supervised regression models are trained in KNIME with AutoML selection via R2 and MAE, and gradient boosted trees outperform Rosenthal’s analytical model.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages17

Reading Time43 min

AccessFree to download

Introduction

Additive manufacturing evolution toward L-PBF

Materials and Methods

Supervised machine learning for melt pool regression

Training setup and AutoML model selection

Experimental campaign and parameter study

Results and Discussion

Predictive performance versus Rosenthal analytical model

Conclusion

applied sciences Article A Supervised Machine Learning Model for Regression to Predict Melt Pool Formation and Morphology in Laser Powder Bed Fusion Niccolò Baldi 1, Alessandro Giorgetti 2, *, Alessandro Polidoro 1, Marco Palladino 3, Iacopo Giovannetti 3, Gabriele Arcidiacono 1 and Paolo Citti 1 Citation: Baldi, N.; Giorgetti, A.; Polidoro, A.; Palladino, M.; Giovannetti, I.; Arcidiacono, G.; Citti, P. A Supervised Machine Learning Model for Regression to Predict Melt Pool Formation and Morphology in Laser Powder Bed Fusion. Appl. Sci. 2024, 14, 328. [https://doi.org/](https://doi.org/)[ ](https://doi.org/)[10.3390/app14010328](10.3390/app14010328) Academic Editor: Soshu Kirihara Received: 2 November 2023 Revised: 22 December 2023 Accepted: 27 December 2023 Published: 29 December 2023 Copyright: © 2023 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ([https://](https://)[ ](https://)[creativecommons.org/licenses/by/](creativecommons.org/licenses/by/)[ ](creativecommons.org/licenses/by/)[4.0/](4.0/)) . 1 Department of Engineering Science, Guglielmo Marconi University, 00193 Rome, Italy; [n.baldi@unimarconi.it](n.baldi@unimarconi.it) (N.B.); [a.polidoro@unimarconi.it](a.polidoro@unimarconi.it) (A.P.); [g.arcidiacono@unimarconi.it](g.arcidiacono@unimarconi.it) (G.A.); p.citti@unimarconi.it (P.C.) 2 Department of Industrial, Electronic and Mechanical Engineering, Roma Tre University, 00146 Rome, Italy 3 Baker Hughes, Nuovo Pignone, 50127 Florence, Italy; [marco.palladino@bakerhughes.com](marco.palladino@bakerhughes.com) (M.P.); [iacopo.giovannetti@bakerhughes.com](iacopo.giovannetti@bakerhughes.com) (I.G.) * Correspondence: [alessandro.giorgetti@uniroma3.it](alessandro.giorgetti@uniroma3.it) Abstract: In the additive manufacturing laser powder bed fusion (L-PBF) process, the optimization of the print process parameters and the development of conduction zones in the laser power (P) and scanning speed (V) parameter spaces are critical to meeting production quality, productivity, and volume goals. In this paper, we propose the use of a machine learning approach during the process parameter development to predict the melt pool dimensions as a function of the P/V combination. This approach turns out to be useful in speeding up the identification of the printability map of the material and defining the conduction zone during the development phase. Moreover, a machine learning method allows for an accurate investigation of the most promising configurations in the P-V space, facilitating the optimization and identification of the P-V set with the highest productivity. This approach is validated by an experimental campaign carried out on samples of Inconel 718, and the effects of some additional parameters, such as the layer thickness (in the range of 30 to 90 microns) and the preheating temperature of the building platform, are evaluated. More specifically, the experimental data have been used to train supervised machine learning models for regression using the KNIME Analytics Platform (version 4.7.7) . An AutoML (node for regression) tool is used to identify the most appropriate model based on the evaluation of R2 and MAE scores. The gradient boosted tree model also performs best compared to Rosenthal’s analytical model. Keywords: laser powder bed fusion; melt pool morphology; powder bed fusion–laser melting; PBF–LM; Inconel 718; design for additive manufacturing; single track; nickel-based alloy; machine learning 1. Introduction The introduction of direct metal laser sintering (DMLS) technology by EOS in 1994 was a consistent technological development that progressively evolved into laser powder bed fusion (L-PBF) technology referred to as powder bed fusion–laser melting (PBF-LM) in ISO/ASTM 52900 [1] as OEMs successfully adapted their selective laser sintering (SLS

FAQs

What melt pool property does the supervised learning model predict in L-PBF?

The model predicts melt pool dimensions by regressing them as a function of the laser power and scanning speed (P/V) combination.

How does the proposed approach help during process parameter development?

It speeds up identifying the material printability map and defining conduction zones, while enabling exploration of promising P-V configurations to maximize productivity.

What tools and evaluation metrics are used to build and choose the regression model?

The models are trained in the KNIME Analytics Platform, and AutoML selects the most appropriate approach using R2 and MAE scores. Gradient boosted trees show the best performance compared with Rosenthal’s analytical model.

Aug 4, 2026

Upload by Eliana