EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

A Span-Prior-Guided Explainable Multimodal Neural Network Method for Final-State Quality Inspection of Hairpin Windings

Xiaopeng Chang iD, Bangcheng Zhang iD, Zhi Gao iD, Siyu Chen iD, Jingru Liu

Sensors · 01.08.2026 · 26 (15) · 4857

DOI: 10.3390/s26154857 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Sensors, online erschienen am 01.08.2026. Verfasst von Xiaopeng Chang und 4 weiteren Autor:innen. Beteiligte Einrichtung: School of Mechatronic Engineering, Changchun University of Technology, Changchun 130012, China. Die Arbeit nennt 35 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Jilin Province Science and Technology Department, Jilin Provincial Scientific and Technological Development Program.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

For final-state quality inspection of three-dimensional stamped hairpin windings, existing studies still lack multimodal methods that integrate mechanical geometric constraints, prior-guided fusion, and decision interpretability. This study proposes a span-prior-guided explainable multimodal neural network method and develops SPIMA-Net. The final-state images were acquired at a fixed inspection station with a fixed camera position and imaging angle under a CCD vision light source. Final-state images are used as visual inputs, while geometric priors are constructed from span measurements and model-type information. A visual branch and a span branch extract image and prior features, and a span-prior-assisted gating mechanism modulates visual features to enable collaborative fusion. Experimental results show that SPIMA-Net achieves an accuracy of 98.14%, an F1-score of 96.55%, and an AUC of 0.9983 on the test set. Its nonconforming-class F1-score is improved by 10.44, 3.04, 6.27, 1.38, and 0.66 percentage points over the image-only, span-only, direct-fusion, SE-fusion, and CBAM-fusion models, respectively, while the total number of misclassifications decreases to five. Interpretability analysis shows that the model mainly focuses on span openings, end profiles, and local abnormal regions. Relative and absolute span deviations are identified as the main mechanical geometric factors affecting final-state quality classification. The proposed method provides a neuro-mechanical fusion approach that demonstrates high discriminative performance and engineering interpretability for hairpin winding quality inspection on the investigated industrial dataset.

Schlagwörter

Stichwörter der Autor:innen

explainable multimodal neural networkfinal-state quality inspectionhairpin windingsmultimodal sensing fusionspan prior

Autor:innen und Institutionen

  • School of Mechatronic Engineering, Changchun University of Technology, Changchun 130012, China.
  • School of Mechanical and Electrical Engineering, Changchun Institute of Technology, Changchun 130103, China.
  • FAW Tool & Die Manufacturing Co., Ltd., No. 1999, Jieda Road, Changchun 130011, China.

ORCID-Profile: Xiaopeng Chang, Bangcheng Zhang, Zhi Gao, Siyu Chen

Förderung

  • Jilin Province Science and Technology Department
  • Jilin Provincial Scientific and Technological Development Program

Zitieren

APA

Chang, X., Zhang, B., Gao, Z., Chen, S., & Liu, J. (2026). A Span-Prior-Guided Explainable Multimodal Neural Network Method for Final-State Quality Inspection of Hairpin Windings. Sensors, 26(15), 4857. https://doi.org/10.3390/s26154857

Vancouver

Chang X, Zhang B, Gao Z, Chen S, Liu J. A Span-Prior-Guided Explainable Multimodal Neural Network Method for Final-State Quality Inspection of Hairpin Windings. Sensors. 2026;26(15):4857. doi:10.3390/s26154857

BibTeX
@article{Chang2026,
  author = {Chang, Xiaopeng and Zhang, Bangcheng and Gao, Zhi and Chen, Siyu and Liu, Jingru},
  title = {A Span-Prior-Guided Explainable Multimodal Neural Network Method for Final-State Quality Inspection of Hairpin Windings},
  journal = {Sensors},
  year = {2026},
  volume = {26},
  number = {15},
  pages = {4857},
  doi = {10.3390/s26154857}
}
Kostenlos starten →