STRUCTURAL WEIGHT MINIMIZATION AND OPTIMIZATION USING MACHINE-LEARNING-DRIVEN SURROGATE FINITE ELEMENT MODELS

This research paper demonstrates an integrated machine-learning-driven surrogate FEM framework for structural weight minimization. A parametric beam problem was used to generate 1500 physics-based design cases containing seven input variables and four structural response quantities. The trained surrogate models successfully learned the nonlinear relationship between geometry, material properties, load and response. Among the tested methods, Random Forest provided the best balanced prediction performance, while Gradient Boosting showed competitive stress and weight accuracy. The feasibility analysis showed that 83% of the sampled design space satisfied the imposed constraints, confirming that optimization must balance weight reduction with strength and serviceability. The validated optimum maintained positive stress, deflection and factor-of-safety margins, proving that the lightest selected design remained structurally acceptable. The study confirms that surrogate modeling can significantly accelerate design-space exploration and reduce dependence on repeated FEM runs. However, final validation remains essential because surrogate outputs are approximations. The proposed workflow is therefore useful as a decision-support tool that combines speed, interpretability and engineering verification for sustainable structural optimization.

[1] Vurtur Badarinath, P., Chierichetti, M., and Davoudi Kakhki, F. (2021). A Machine Learning Approach as a Surrogate for a Finite Element Analysis: Status of Research and Application to One Dimensional Systems. Sensors, 21(5), 1654. [2] Kneifl, J., Grunert, D., and Fehr, J. (2021). A Nonintrusive Nonlinear Model Reduction Method for Structural Dynamical Problems Based on Machine Learning. International Journal for Numerical Methods in Engineering, 122(17), 4774-4786. [3] Hashemi, A., Jang, J., and Beheshti, J. (2023). A Machine Learning-Based Surrogate Finite Element Model for Estimating Dynamic Response of Mechanical Systems. IEEE Access, 11, 54509-54525. [4] Wu, H., Wu, Y., Zhi, P., Wu, X., and Zhu, T. (2023). Structural Optimization of Single-Layer Domes Using Surrogate-Based Physics-Informed Neural Networks. Heliyon, 9(10), e20867. [5] Etim, B., Al-Ghosoun, A., Renno, J., Seaid, M., and Mohamed, M. S. (2024). Machine Learning-Based Modeling for Structural Engineering: A Comprehensive Survey and Applications Overview. Buildings, 14(11), 3515. [6] Pei, Y., Han, B., Kumar, D., et al. (2025). Machine Learning as a Surrogate for FEM: Predicting Mechanical Properties of Tyres. Advanced Industrial and Engineering Polymer Research, 8(4), 499-515. [7] Huang, L., Li, G., Guan, Y., Jiao, W., Gong, S., and Xu, S. (2025). Machine Learning and Finite Element Integration-Driven Surrogate Model for Fluid-Structure Interaction Seismic Response Analysis of Aqueduct Structures. Results in Engineering, 27, 106176. [8] Zhao, Z., Wang, J., and Wu, N. (2025). Machine Learning-Based Surrogate Ensemble for Frame Displacement Prediction Using Jackknife Averaging. Buildings, 15(16), 2872. [9] Zhou, H., Shen, Y., Chen, S., and Jing, X. (2025). Intelligent Reliability-Based Design Optimization of Complex Engineering Structures Using Manifold Learning-Enhanced Surrogate Modeling. Advances in Mechanical Engineering, 17(12). [10] Bishop, C. M. (2006). Pattern Recognition and Machine Learning. Springer.