Emotion-driven optimization method shapes new energy SUV wheel hub styling

The study, published in Scientific Reports, combines expert screening, imagery analysis, machine learning and multi-objective optimization into a single workflow.

The team built three imagery dimensions through scenario-based expert screening, refinement of imagery words, principal component analysis and user ratings. Morphological analysis and the analytic hierarchy process then identified eight design inputs describing wheel hub form. These inputs fed a set of predictive models, including a particle swarm optimization–backpropagation neural network (PSO-BP), a plain BP network, support vector regression, a sparrow search algorithm–optimized SVR, and ordinary least squares regression.

Across 20 matched random data splits, PSO-BP achieved the lowest equally weighted three-target test root mean square error, at 0.1896. The statistical comparisons were cautious: after correlation adjustment and Holm correction, only the difference between PSO-BP and OLS for modern dynamic imagery was significant, with p = 0.0315. The authors note that surrogate model changes and prediction errors can affect rankings, even though equal weighting and six local weight perturbations preserved the baseline top-ranked candidate within a fixed 354-candidate set.

To generate and select designs, the researchers used the third-generation nondominated sorting genetic algorithm (NSGA-III) together with entropy-weighted TOPSIS. Ten optimization searches produced a final hypervolume of 1.10825 ± 0.00360, indicating consistent convergence of the candidate set across repeated runs.

The final implemented design was checked for structural safety. Its maximum radial static stress reached 175 MPa, below the 252 MPa yield strength of the material, suggesting the styling choice remains within engineering limits. The authors describe the overall decoupling–mapping–optimization–verification approach as a support for early styling decisions, while emphasizing that its usefulness depends on sample coverage, predictive uncertainty and the specific engineering conditions of each project.

Subject of Research: A styling optimization method for new energy SUV wheel hubs based on positive emotional granularity

Article Title: A styling optimization design method for new energy SUV wheel hubs based on positive emotional granularity

Article References: Zhou, Q., Li, G., Zhang, W., Li, X., Zhang, C., & Jie, M. (2026). A styling optimization design method for new energy SUV wheel hubs based on positive emotional granularity. Scientific Reports. https://doi.org/10.1038/s41598-026-77508-3

Image Credits: AI Generated

DOI: 10.1038/s41598-026-77508-3

Keywords: wheel hub design, new energy SUV, positive emotional granularity, PSO-BP neural network, NSGA-III, TOPSIS, morphological analysis, analytic hierarchy process, styling optimization, Scientific Reports, styling, optimization