Prediction of porosity, hardness and surface roughness in additive manufactured AlSi10Mg samples

Fatma Alamri, Imad Barsoum, Shrinivas Bojanampati and Maher Maalouf

PLOS ONE, 2025, vol. 20, issue 3, 1-29

Abstract: Despite the advantages of additive manufacturing, its widespread adoption is still hindered by the poor quality of the fabricated parts. Advanced machine learning techniques to predict part quality can improve repeatability and open additive manufacturing to various industries. This study aims to accurately predict the relative density, surface roughness and hardness of AlSi10Mg samples produced by selective laser melting regarding process parameters such as scan speed, layer thickness, laser power, and hatch distance. For this purpose, data including porosity, surface hardness, and roughness were extracted from the literature, and additional measurements were performed on additive manufactured samples in the current work. This work compares five supervised machine learning algorithms, including artificial neural networks, support vector regression, kernel ridge regression, random forest, and Lasso regression. These models are evaluated based on the coefficient of determination and the mean squared error. On the basis of the computational results, the artificial neural network outperformed in predicting relative density, surface roughness, and hardness. Feature importance analysis on the compiled dataset using ANN revealed that laser power and scan speed are the most important features affecting relative density (e.g., porosity) and hardness, while scan speed and layer thickness significantly impact the surface roughness of the parts. The study identified an optimal laser power and scan speed region that achieves a relative density > 99%, surface roughness 120 HV. The results presented in this study provide significant advantages for additive manufacturing, potentially reducing experimentation costs by identifying the process parameters that optimize the quality of the fabricated parts.

Date: 2025
References: Add references at CitEc
Citations:

Downloads: (external link)
https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0316600 (text/html)
https://journals.plos.org/plosone/article/file?id= ... 16600&type=printable (application/pdf)

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:plo:pone00:0316600

DOI: 10.1371/journal.pone.0316600

Access Statistics for this article

More articles in PLOS ONE from Public Library of Science
Bibliographic data for series maintained by plosone ().