Close menu

SURE

Sunderland Repository records the research produced by the University of Sunderland including practice-based research and theses.

Comparative analysis of metaheuristic algorithms for optimising wear behaviour of LPBF printed AlSi10Mg at elevated temperatures

Jatti, Vijaykumar S., Saiyathibrahim, A., Vijayan, S. N., Krishnan, R. Murali and Mohan, Dhanesh G. (2025) Comparative analysis of metaheuristic algorithms for optimising wear behaviour of LPBF printed AlSi10Mg at elevated temperatures. Canadian Metallurgical Quarterly. pp. 1-29. ISSN 1879-1395

Item Type: Article

Abstract

This research presents a comparative analysis of six metaheuristic algorithms such as Genetic Algorithm (GA), Simulated Annealing (SA), Particle Swarm Optimisation (PSO), Teaching-Learning-Based Optimisation (TLBO), Cohort Intelligence (CI), and JAYA Algorithm focused on optimising the wear behaviour of LPBF-printed AlSi10Mg parts under elevated temperatures. The potential of each algorithm to optimise the process factors and reduce wear loss and Specific Wear Rate (SWR) was used to assess its performance. SEM analysis reveals that the LPBF-processed AlSi10Mg alloy has a refined microstructure, with a predominantly α-Al matrix and finely dispersed silicon particles. It attains a high degree of compaction and minimal porosity with density close to the theoretical value. EDS mapping shows the presence of aluminium, silicon, magnesium, and copper, with silicon predominantly in eutectic form. X-ray diffraction analysis reveals that there is an aluminium phase arranged in a face-centered cubic structure and magnesium silicide. The results also indicated that PSO, TLBO, and JAYA outperformed other algorithms and that TLBO had the best result in terms of accuracy. The wear loss obtained by TLBO was 0.044924 grams, the least among other methods, with the minimal experimental deviation of 6.21%, which shows that the TLBO technique is capable of optimising the wear resistance of LPBF-printed AlSi10Mg parts. PSO, TLBO, and JAYA were able to obtain the lowest SWR values, setting them apart from the other algorithms in terms of their performance. The minimum experimental deviation of 0.91% was obtained by TLBO, which was the most accurate predictor. It was found that the use of CI to achieve the wear resistance was more effective than through TLBO. Abrasion and adhesion wear mechanisms were minimised by TLBO and resulted in significantly lower wear loss and SWR. On the other hand, CI gave rise to more aggressive wear. The use of TLBO to identify optimal factor combinations successfully reduced material removal while enhancing the tribological performance of the AlSi10Mg components, especially under high-temperature conditions. This indicates that TLBO effectively enhances the wear-resisting properties of AlSi10Mg prints.

Cette recherche présente une analyse comparative de six algorithmes métaheuristiques tels que l'algorithme génétique (GA), le recuit simulé (SA), l'optimisation par essaim de particules (PSO), l'optimisation basée sur l'enseignement et l'apprentissage (TLBO), l'intelligence de cohorte (CI) et l'algorithme JAYA centrés sur l'optimisation du comportement à l'usure des pièces d'AlSi10Mg imprimées par fusion laser sur lit de poudre (LPBF) sous des températures élevées. On a utilisé le potentiel de chaque algorithme à optimiser les facteurs du procédé et à réduire les pertes par usure et le taux d'usure spécifique (SWR) pour évaluer ses performances. L'analyse microstructurale présente matrice dominée par Al-a avec des particules de silicium finement dispersées, ayant un haut degré de compaction et une porosité minimale. La cartographie EDS montre la présence d'aluminium, de silicium, de magnésium et de cuivre, le silicium étant principalement sous forme eutectique. L'analyse par diffraction des rayons X révèle la présence d'une phase d'aluminium arrangée en structure cubique à faces centrées et de siliciure de magnésium. Les résultats ont indiqué également que PSO, TLBO et JAYA surpassaient les autres algorithmes et que TLBO a obtenu le meilleur résultat en termes de précision. La perte d'usure obtenue par TLBO était de 0.044924 gramme, la plus faible parmi les autres méthodes, avec un écart expérimental minimal de 6.21%. Le TLBO a obtenu l'écart expérimental minimal de 0.91%, ce qui était le prédicteur le plus précis. TLBO a minimisé les mécanismes d'usure par abrasion et adhérence, ce qui a entraîné une réduction importante des pertes par usure et du SWR. Par contre, CI a entraîné une usure plus agressive. Globalement, ces résultats indiquent que TLBO améliore efficacement les propriétés de résistance à l'usure à haute température de l'AlSi10Mg imprimé.

Full text not available from this repository.

More Information

Additional Information: ** From Crossref journal articles via Jisc Publications Router ** History: received 08-04-2025; accepted 05-09-2025; epub 13-10-2025; issued 13-10-2025; published 13-10-2025.
SWORD Depositor: Publication Router
Depositing User: Publication Router

Identifiers

Item ID: 19602
Identification Number: 10.1080/00084433.2025.2560208
ISSN: 1879-1395
URI: https://sure.sunderland.ac.uk/id/eprint/19602
Official URL: https://www.tandfonline.com/doi/abs/10.1080/000844...

Users with ORCIDS

ORCID for Dhanesh G. Mohan: ORCID iD orcid.org/0000-0002-4652-4198

Catalogue record

Date Deposited: 04 Dec 2025 15:13
Last Modified: 04 Dec 2025 15:13

Contributors

Author: Dhanesh G. Mohan ORCID iD
Author: Vijaykumar S. Jatti
Author: A. Saiyathibrahim
Author: S. N. Vijayan
Author: R. Murali Krishnan

University Divisions

Faculty of Business and Technology > School of Computer Science and Engineering

Subjects

Engineering

Actions (login required)

View Item (Repository Staff Only) View Item (Repository Staff Only)