International Journal of Advanced Engineering Application

Machine Learning-Optimised Lattice Architectures for Lightweight Aerospace Brackets

Author(s):Aditi R. Krishnan

Affiliation: Department of Aerospace and Mechanical Engineering

Page No: 60-65

Volume issue & Publishing Year: Volume 3, Issue 7, July 2026

Journal: International Journal of Advanced Engineering Application (IJAEA)

ISSN NO: 3048-6807

DOI:

Download PDF Cite this article

Abstract:
Mass reduction in aerospace structural brackets directly improves fuel efficiency, payload capacity, and lifecycle emissions, yet conventional topology optimisation and manually designed lattice infill strategies struggle to navigate the vast design space spanned by unit-cell topology, strut diameter, relative density gradients, and laser powder bed fusion (LPBF) process parameters simultaneously. This study develops a machine learning (ML) surrogate-assisted optimisation framework that couples a gradient-boosted regression surrogate model, trained on 4,200 finite element analysis (FEA) simulations, with a generative design search to identify hybrid lattice architectures that maximise specific stiffness subject to manufacturability and fatigue-life constraints.
Gyroid triply periodic minimal surface (TPMS), octet-truss, and ML-optimised hybrid lattices were evaluated across relative densities of 0.10-0.40, fabricated in Ti-6Al-4V via LPBF, and characterised through quasi-static compression testing, micro-computed tomography (CT) porosity quantification, and high-cycle axial fatigue testing (R = 0.1) against wrought Ti-6Al-4V baselines. Process-parameter sensitivity was mapped across laser power (150-300 W) and scan speed (800-1200 mm/s) to identify a low-porosity build window, and the ML surrogate's feature importance was extracted via SHAP analysis to identify the dominant design drivers.
The ML-optimised hybrid lattice achieved 41.8% mass reduction relative to a solid topology-optimised baseline while retaining 87% of the baseline's fatigue strength at 10^6 cycles, outperforming octet-truss (35.2% mass reduction, 68% fatigue retention) and gyroid TPMS (31.8% mass reduction, 79% fatigue retention) alternatives. The ML surrogate model achieved R² = 0.97 against held-out FEA validation data, with strut diameter, unit-cell type, and relative density identified as the three dominant predictors of stiffness. The recommended LPBF process window (210-240 W laser power, 800 mm/s scan speed) reduced CT-measured porosity to below 0.4%. The ML-optimised hybrid lattice also reduced embodied CO₂ per bracket by 41.4% relative to the solid baseline, combining structural and environmental performance gains.

Keywords: lattice structures, topology optimisation, machine learning, additive manufacturing, laser powder bed fusion, fatigue, Ti-6Al-4V, generative design, aerospace lightweighting

Reference:

• [1] Ashby, M. F. (2006). The properties of foams and lattices. Philosophical Transactions of the Royal Society A, 364(1838), 15-30.
• [2] Bates, S. R. G., Farrow, I. R., & Trask, R. S. (2019). Compressive behaviour of 3D printed thermoplastic polyurethane honeycombs with graded densities. Materials & Design, 162, 130-142.
• [3] Beyer, C., & Figueroa, D. (2016). Design and analysis of lattice structures for additive manufacturing. Journal of Manufacturing Science and Engineering, 138(12), 121014.
• [4] Du Plessis, A., Yadroitsava, I., & Yadroitsev, I. (2020). Effects of defects on mechanical properties in metal additive manufacturing. Additive Manufacturing, 35, 101424.
• [5] Gibson, L. J., & Ashby, M. F. (1997). Cellular Solids: Structure and Properties (2nd ed.). Cambridge University Press.
• [6] Liu, F., Mao, Z., Zhang, P., et al. (2018). Functionally graded porous scaffolds in multiple patterns. Materials & Design, 160, 849-860.
• [7] Lundberg, S. M., & Lee, S. I. (2017). A unified approach to interpreting model predictions. Advances in Neural Information Processing Systems, 30, 4765-4774.
• [8] Maskery, I., Aboulkhair, N. T., Aremu, A. O., et al. (2017). A mechanical property evaluation of graded density Al-Si10-Mg lattice structures. Materials Science and Engineering A, 670, 264-274.
• [9] Plocher, J., & Panesar, A. (2019). Review on design and structural optimisation in additive manufacturing. Materials & Design, 183, 108164.
• [10] Tao, W., & Leu, M. C. (2016). Design of lattice structure for additive manufacturing. International Symposium on Flexible Automation, 325-332.
• [11] Yan, C., Hao, L., Hussein, A., & Raymont, D. (2012). Evaluations of cellular lattice structures manufactured using selective laser melting. International Journal of Machine Tools and Manufacture, 62, 32-38.
• [12] Yang, L., Mertens, R., Ferrucci, M., et al. (2019). Continuous graded Gyroid cellular structures fabricated by selective laser melting. Materials & Design, 162, 394-404.
• [13] Zhang, P., Toman, J., Yu, Y., et al. (2015). Efficient design-optimization of variable-density hexagonal cellular structure by additive manufacturing. Journal of Manufacturing Science and Engineering, 137(2), 021004.