Generative AI-Driven Design Optimization of Lightweight Polymer Composites for Electric Vehicles

Notice

This is an unedited manuscript accepted for publication and provided as an Article in Press for early access at the author’s request. The article will undergo copyediting, typesetting, and galley proof review before final publication. Please be aware that errors may be identified during production that could affect the content. All legal disclaimers of the journal apply.

Year : 2026 | Volume : 14 | 04 | Page :
By

S. Maheswari,

Manna Sheela Rani Chetty,

Rashid Hashmi,

Mehak Jonjua,

Amit Bindal,

Narne Sravanthi,

  1. Assistant Professor, Department of Artificial Intelligence and Data Science, Vel Tech Rangarajan Dr. Sagunthala R&D Institute of Science and Technology, Chennai, Tamil Nadu, India
  2. Professor, Department of Computer Science and Engineering, K L University, Vaddeswaram, Guntur District, Andhra Pradesh, India
  3. Professor of Practice, Sharda School of Media, Film & Entertainment, Sharda University, Uttar Pradesh, India
  4. Professor, Sharda School of Media, Film & Entertainment, Sharda University, Uttar Pradesh, India
  5. Professor, Department of Computer Science & Engineering, MM Engineering College, Maharishi Markandeshwar (Deemed to be University), Mullana, Ambala, Haryana, India
  6. Assistant Professor, Department of Computer Science and Engineering, Koneru Lakshmaiah Education Foundation, Green Fields, Vaddeswaram, Guntur, Andhra Pradesh, India

Abstract

Lightweight polymer composites are increasingly important for electric vehicles, where mass reduction must be achieved without compromising structural performance, thermal stability, manufacturability, or material reliability. This study develops a generative AI-driven inverse-design framework for identifying experimentally credible lightweight polymer-composite configurations under coupled EV-oriented constraints. Public experimental polymer-composite datasets were integrated through leakage-controlled preprocessing and group-aware validation. A multi-task neural surrogate predicted mechanical response, while a conditional variational autoencoder explored feasible composition–processing combinations. Generated candidates were screened using constituent balance, experimental-domain limits, predictive uncertainty, and Pareto-based multi-objective optimization. No generated records were used for model validation or testing. Under leave-one-composite-type-out validation, the proposed surrogate achieved (R^2) values of 0.966 for flexural strength and 0.933 for flexural modulus. Descriptor analysis identified consolidation pressure and glass-transition temperature as the dominant transferable variables. From 12,000 candidate designs, 3,889 satisfied the complete feasibility criteria, with the highest-ranked solutions converging toward a narrow high-performance processing region. The framework provides a reproducible route for reducing polymer-composite design space before fabrication, supporting data-guided development of lightweight EV structures while maintaining physical and experimental credibility.

Keywords: Generative AI, Polymer Composites, Lightweight Materials, Electric Vehicles, Multi-Objective Optimization.

How to cite this article: S. Maheswari, Manna Sheela Rani Chetty, Rashid Hashmi, Mehak Jonjua, Amit Bindal, Narne Sravanthi. Generative AI-Driven Design Optimization of Lightweight Polymer Composites for Electric Vehicles. Journal of Polymer & Composites. 2026; 14(04):-.
How to cite this URL: S. Maheswari, Manna Sheela Rani Chetty, Rashid Hashmi, Mehak Jonjua, Amit Bindal, Narne Sravanthi. Generative AI-Driven Design Optimization of Lightweight Polymer Composites for Electric Vehicles. Journal of Polymer & Composites. 2026; 14(04):-. Available from: https://journals.stmjournals.com/jopc/article=2026/view=252985

References

  1. Park D, Park M, Ryu S. Expanding design spaces in digital composite materials: a multi-input deep learning approach enhanced by transfer learning and multi-kernel network. Adv Theory Simul. 2023;6:2300465. doi:10.1002/adts.202300465.
  2. Xu X, Zhao W, Wang L, Lin J, Du L. Efficient exploration of compositional space for high-performance copolymers via Bayesian optimization. Chem Sci. 2023;14(37):10203-10211. doi:10.1039/D3SC03174H.
  3. Deng H, Zhao Q, Gao X, Peng HX, Zhou H. Exploring the design space of discontinuous metal matrix composites through domain-knowledge enhanced machine learning. Extreme Mech Lett. 2024;70:102176. doi:10.1016/j.eml.2024.102176.
  4. Sheikh N, Mamodiya U, Somwanshi DK. A lightweight agentic AI framework for real-time instructional adaptation in portable VR learning platforms. AIP Conf Proc. 2026;3426(1):040023. doi:10.1063/5.0327307.
  5. Semitekolos D, Araújo A, Santos RM, Pernechele C, Panozzo F, Vescovi L, et al. Design, manufacturing, and evaluation of race and automotive prototypal components fabricated with modified carbon fibres and resin. Polymers (Basel). 2024;16:2062. doi:10.3390/polym16142062.
  6. Lv T, Chen Z, Wang D, Du X. Parametric design and modeling method of carbon fiber reinforcement plastic-laminated components applicable for multi-material vehicle body development. J Comput Des Eng. 2024;11(1):261-287. doi:10.1093/jcde/qwae007.
  7. Qiu H, Sun ZY. On-demand reverse design of polymers with PolyTAO. NPJ Comput Mater. 2024;10:273. doi:10.1038/s41524-024-01466-5.
  8. Maurya S, Kishor I, Sheikh N, Somwanshi DK. A diffusion model-based framework for multi-agent agent-based generative planning. AIP Conf Proc. 2026;3426(1):040024. doi:10.1063/5.0327308.
  9. Lee KH, Lim HJ, Yun GJ. A data-driven framework for designing microstructure of multifunctional composites with deep-learned diffusion-based generative models. Eng Appl Artif Intell. 2024;129:107590. doi:10.1016/j.engappai.2023.107590.
  10. Kumar G, Sharma R. Innovative AI techniques for solar energy systems: predictive analytics and real-time optimization. AIP Conf Proc. 2026;3426(1):040022. doi:10.1063/5.0327557.
  11. Cebe E, Irez AB. Novel hybrid polymer composites for optimized thermal performance in electric vehicle battery cases. Polym Compos. 2025;46(7):6609-6622. doi:10.1002/pc.29381.
  12. Dobrotă D, Sava GA, Bărbușiu AM, Dobrescu GT. Sustainable polymer composites for thermal insulation in automotive applications: a systematic literature review. Polymers (Basel). 2025;17:2200. doi:10.3390/polym17162200.
  13. Kopac T. Advancing polymer science and energy storage solutions through the integration of artificial intelligence and machine learning: a transformative approach. Polymers (Basel). 2025;17:3267. doi:10.3390/polym17243267.
  14. Liu Y, Zheng W, Ai H, Cheng L, Guo R, Song X. Machine learning for performance prediction and optimization of polymer composites: unveiling the dominant role of thermally conductive pathways. ACS Appl Mater Interfaces. 2025;17(51):70143-70155. doi:10.1021/acsami.5c21877.
  15. Azzopardi B, Hapid A, Kaleg S, Onggo D, Budiman AC. Recent advances in battery pack polymer composites. Energies. 2023;16(17):6223. doi:10.3390/en16176223.
  16. Vogel G, Weber JM. Inverse design of copolymers including stoichiometry and chain architecture. Chem Sci. 2025;16(3):1161-1178. doi:10.1039/D4SC05900J.
  17. Jigyasu R, Ghai M, Kakarla NML. Explainable quantum-neuromorphic intelligence: a self-evolving hybrid framework for cognitive computing and autonomous decision systems. In: Emerging Hybrid Models for Neuromorphic AI and Quantum Computing. Hershey (PA): IGI Global; 2026. Chapter 11. doi:10.4018/979-8-3373-7779-7.ch011.
  18. Lu D, Zhang S, Li Z, Chen J, Xiong F, Zhang Y, et al. Carbon fiber reinforced composite material layup design optimization method and its application in automobile battery box. PLoS One. 2025;20(12):e0322729. doi:10.1371/journal.pone.0322729.
  19. Wang B, Zhan L, Guan C. Advanced polymer composites and applications. Polymers (Basel). 2025;17(15):2062. doi:10.3390/polym17152062.
  20. García-Hernández Z, Molina-Ramírez O, Rivera-Salinas JE, Sifuentes-Nieves I, González-Morones P, Hernández-Hernández E. Boron nitride: the key material in polymer composites for electromobility. Polym Compos. 2025;46(3):1976-2029. doi:10.1002/pc.29106.
  21. Zhang S, Lu D, Huang M, Li R, Li Z, Xu W. Research on the lightweight design and multilevel optimization method of B-pillar of hybrid material based on side-impact safety and multicriteria decision-making. Polym Compos. 2024;45(5):4051-4077. doi:10.1002/pc.28042.
  22. Kishor I, Mamodiya U, Dutta PK. AI-driven solar energy optimization for autonomous 3D printing units in smart manufacturing ecosystems. In: Durakbasa NM, Cetinkaya K, Demircioglu P, Bogrekci I, editors. Additive Manufacturing Technologies. Cham: Springer; 2026. Chapter 10. doi:10.1007/978-3-032-21865-0_10.
  23. Zhang X, Lin Q, Xiao Y, Jia L, Yang T, Wang L, et al. Top-down design approach of lightweight composite battery pack enclosure for electric vehicles based on numerical modeling and topology optimization. Polymers (Basel). 2025;17:2897. doi:10.3390/polym17212897.
  24. He D, Kim HC, Sommacal S, Stojcevski F, Soo VK, Lipiński W, et al. Improving mechanical and life cycle environmental performances of recycled CFRP automotive component by fibre architecture preservation. Compos Part A Appl Sci Manuf. 2023;175:107749. doi:10.1016/j.compositesa.2023.107749.
  25. Liu X, Zhou XY, Liu B, Gao C. Multiscale modeling of woven composites by deep learning neural networks and its application in design optimization. Compos Struct. 2023;324:117553. doi:10.1016/j.compstruct.2023.117553.
  26. Wang Z, Zheng C, Wang L, Gu J, Jing L, Lai X. Optimal design of composite grid/skin structures based on deep learning and Double-Double layup strategy. Aerosp Sci Technol. 2024;147:109030. doi:10.1016/j.ast.2024.109030.
  27. Park D, Lee J, Park K, Ryu S. Hierarchical generative network: a hierarchical multitask learning approach for accelerated composite material design and discovery. Adv Eng Mater. 2023;25:2300867. doi:10.1002/adem.202300867.
  28. Zheng T, Cai C, Yang F, Wang R, Liu W. Physics-informed neural-network-based generation of composite representative volume elements with non-uniform distribution and high-volume fractions. Polymers (Basel). 2026;18:97. doi:10.3390/polym18010097.
  29. Seo M, Min S. DL-MSTO+: a deep learning-based multi-scale topology optimization framework via positive definiteness ensured material representation network. Comput Methods Appl Mech Eng. 2023;415:116276. doi:10.1016/j.cma.2023.116276.
  30. Fonseca JH, Jang W, Han D, Kim N, Lee H. Strength and manufacturability enhancement of a composite automotive component via an integrated finite element/artificial neural network multi-objective optimization approach. Compos Struct. 2024;327:117694. doi:10.1016/j.compstruct.2023.117694.
  31. Zhang H, Song Z, Zhang Y, Zhang L, Zhu P. A concurrent design optimization framework for IMSFRP composite structures considering material and structural parameters simultaneously. Thin-Walled Struct. 2024;196:111449. doi:10.1016/j.tws.2023.111449.
  32. Liu T, Zhang T, Chen Y, Wang W, Jiang Y, Huang Y, et al. Neural co-optimization of structural topology, manufacturable layers, and path orientations for fiber-reinforced composites. ACM Trans Graph. 2025;44(4):128. doi:10.1145/3730922.
  33. Xu Y, Zhang J, Xu T, et al. AI-powered inverse design of high-performance unidirectional CFRP composites: breaking the compressive-tensile trade-off via machine learning. Macromol Rapid Commun. 2026;47(12):e00482. doi:10.1002/marc.202500482.
  34. Yossef M, Noureldin M, Alqabbany A. Explainable artificial intelligence framework for FRP composites design. Compos Struct. 2024;341:118190. doi:10.1016/j.compstruct.2024.118190.
  35. Malashin I, Masich I, Tynchenko V, Gantimurov A, Nelyub V, Borodulin A, et al. Machine learning in 3D and 4D printing of polymer composites: a review. Polymers (Basel). 2024;16(22):3125. doi:10.3390/polym16223125.
  36. Shimakawa H, Kumada A, Sato M. Prevention of leakage in machine learning prediction for polymer composite properties. J Chem Inf Model. 2024;64(9):3621-3629. doi:10.1021/acs.jcim.3c01894.
  37. Alsheghri A, Drakonakis V. Predicting mechanical properties of CFRP composites using data-driven models with comparative analysis. Mendeley Data. 2024;Version 1. doi:10.17632/fspdwb4mst.1.
  38. Orlov M, Grishin S, Aksenov K, Aksenova L. PCM mechanical properties. Mendeley Data. 2024;Version 1. doi:10.17632/p63pgr2hb9.1.
Ahead of Print Subscription Original Research
Volume 14
04
Received 08/08/2026
Accepted 17/08/2026
Published 20/08/2026
Publication Time 12 Days

Login

My IP

PlumX Metrics