@article{WANG2026, 
author = {Zihao WANG and Chunhua WAN and Xiaohua NIE and Liang CHANG and Yile HU},
title = {A CGAN-based rapid optimization method for lightweight design of multi-rivet joint structures in aircraft},
year = {2026},
journal = {Acta Aeronautica et Astronautica Sinica},
volume = {47},
number = {20},
keywords = {multi-rivet joint structure, conditional generative adversarial network, structural optimization design, lightweight, Wasserstein distance},
url = {https://www.sciopen.com/article/10.7527/S1000-6893.2026.33252},
doi = {10.7527/S1000-6893.2026.33252},
abstract = {Multi-rivet fastening is a key joining process for aircraft structures, directly affecting the overall load-bearing capacity of aircraft. Traditional design optimization methods have significant limitations in balancing high-dimensional data processing, computational cost control, and engineering practicality. This paper proposes a lightweight optimization method for multi-rivet joint structures based on Conditional Generative Adversarial Networks (CGAN), with the core lying in deeply integrating physical laws, lightweight objectives, and engineering constraints into the deep learning model to achieve efficient collaborative optimization. Aiming at the defect that pure data-driven models tend to violate mechanical principles, the multi-rivet load distribution law and shear failure criteria are embedded into the loss function of the CGAN generator; meanwhile, the objective of minimizing structural mass and engineering manufacturing constraints are incorporated to ensure that the design schemes are lightweight, process-feasible, and compliant with mechanical principles. A finite element model is established based on the Johnson-Cook constitutive model and failure criterion, and after calibration with test results, a “test-simulation” integrated dataset is formed to expand the parameter coverage; the Wasserstein distance is introduced to avoid gradient vanishing or explosion, improving the stability of model training. The training results show that the generator and discriminator achieve stable adversarial equilibrium, and both the strength and weight loss functions effectively converge, verifying the model's multi-objective collaborative optimization capability of “strength-lightweight”. Under the tensile condition of the multi-rivet joint structure, by automatically optimizing structural parameters, a weight reduction of more than 5% is achieved while meeting the strength requirements, and the generation time of the optimized scheme is only 300 ms. This method can quickly produce optimized design schemes that meet the actual engineering needs, greatly shorten the R&D cycle, reduce costs, and provide a new technical path for the high-performance and lightweight design of aircraft multi-rivet joint structures.}
}