Abstract: The lightweight design of the wheel hub was carried out, and the original aluminum alloy rim was replaced by carbon fiber composite material, and the optimization method using neural network as the surrogate model was adopted. Based on the isostiffness theory, the initial thickness of the carbon fiber composite rim was determined. Considering the influence of aluminum alloy spokes and carbon fiber rim thickness on their performance, Latin hypercube sampling was used to generate multiple groups of test samples. Based on the experimental samples, the neural network was used as a surrogate model to optimize the thickness of the spokes and the thickness of the carbon fiber layup at various angles of the rim. In order to obtain the best carbon fiber layup sequence, the carbon fiber rim layup sequence was further optimized in Optistruct. The resulting CFRP/aluminum hub is 18.43% lighter in weight and meets the requirements for stiffness and strength.

Key words: aluminum/carbon fiber, wheels, lightweighting, neural networks, layup optimization, composites