Page 48 - IJAMD-2-3

International Journal of AI for
Materials and Design Biomimetic ML for AFSD aluminum properties

process. This may be due to logarithmic strain’s heightened Acknowledgments
sensitivity to local fluctuations in material qualities or
process parameters or to influences from factors not fully None.
captured by the current set of input variables. Despite this, Funding
both GA-DT and GA-RF models demonstrate the ability
to capture underlying patterns in the data, with GA-RF None.
consistently outperforming GA-DT across both prediction
tasks. These findings underscore the promise of ML Conflict of interest
techniques – particularly those optimized through GAs – The author declares no competing interests.
for predicting complicated material behaviors in advanced
manufacturing processes such as AFSD. Author contributions
The proposed biomimetic ML approach effectively This is a single-authored article.
integrates finite element simulations with GA-optimized
predictive models to estimate von Mises stress and Ethics approval and consent to participate
logarithmic strain in AFSD aluminum alloy structures. Not applicable.
This approach is consistent with previous studies, such
as that conducted by Shi et al. , who developed AFSD- Consent for publication
40
Nets to accurately predict temperature evolution during Not applicable.
AFSD using physics-informed ML. Similarly, Qiao et al.
18
demonstrated the utility of ML techniques, such as RF Availability of data
models, in predicting the mechanical properties of AFSD-
fabricated AA6061. Collectively, these findings support The data will be provided to the readers upon reasonable
the viability of ML in enhancing process knowledge and request.
optimization within solid-state additive manufacturing. Further disclosure

4. Conclusion The paper has been uploaded to or deposited in arXiv
preprint server: https://arxiv.org/abs/2408.05237.
In this presented work, a biomimetic ML-based
approach was developed for predicting two important References
mechanical properties – von Mises stress and logarithmic
strain – in AFSD aluminum alloy-walled structures. 1. Yu HZ, Mishra RS. Additive friction stir deposition:
The proposed method integrates finite element analysis A deformation processing route to metal additive
with GA-optimized DT and RF models. This combined manufacturing. Mater Res Lett. 2021;9(2):71-83.
approach demonstrates strong predictive capability, with doi: 10.1080/21663831.2020.1847211
R² values of 0.9676 and 0.7201 for von Mises stress and 2. Patil SM, Sharma S, Joshi SS, Jin Y, Radhakrishnan M,
logarithmic strain, respectively, utilizing the GA-RF model. Dahotre NB. Additive friction stir deposition of Al 6061-
These results indicate a high level of correlation within the B4C composites: Process parameters, microstructure and
simulated dataset; however, full experimental validation property correlation. Mater Sci Eng A. 2024;910:146840.
is needed to confirm the model’s accuracy in real-world doi: 10.1016/j.msea.2024.146840
applications.
3. Korganci M, Bozkurt Y. Recent developments in additive
Future research will focus on validating the model friction stir deposition (AFSD). J Mater Res Technol.
against experimental data to strengthen confidence in 2024;30:4572-4583.
its applicability. Expanding the model to encompass a doi: 10.1016/j.jmrt.2024.04.179
broader range of alloy types and other relevant materials
could further enhance its industrial utility. Incorporating 4. Liu H, Xu M, Li X. Achievement of high-reliability and high-
real-time monitoring data during the AFSD process efficient deposit of PA66 by additive friction stir deposition.
may improve predictive accuracy and facilitate adaptive Compos Part B Eng. 2024;284:111682.
control. In addition, the development of a user-friendly doi: 10.1016/j.compositesb.2024.111682
interface could accelerate adoption in industrial settings 5. Chen L, Lu L, Zhu L, et al. Microstructure evolution and
and contribute to transformative improvements in mechanical properties of multilayer AA6061 alloy fabricated
process optimization and quality control for the additive by additive friction stir deposition. Metall Mater Trans A.
manufacturing of high-performance aluminum structures. 2024;55(4):1049-1064.

Volume 2 Issue 3 (2025) 42 doi: 10.36922/ijamd.5014