EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Thermodynamics-Informed Residual-Disentangled Network for Manufacturing Defect Diagnosis of Reciprocating Compressors Under Extreme Class Imbalance

Jiangqing Wang, Xinqiao Jin, Mengke Zhao, Yuling Gong, Lin Chen

Processes · 19.11.2026 · 14 (22) · 3723

DOI: 10.3390/pr14223723 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Processes, online erschienen am 19.11.2026. Verfasst von Jiangqing Wang und 4 weiteren Autor:innen. Die Arbeit nennt 27 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Scientific Research Start-Up Foundation for High-Level Talents of Taizhou University, Wuxi University Research Start-Up Fund for Introduced Talents, 2025 Jiangsu Provincial Industry-University-Research Cooperation Project.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Manufacturing defect diagnosis of reciprocating compressors is limited by two coupled problems: extreme class imbalance and the physical opacity of data-driven models. A physics-informed residual-disentangled network (PIRD-Net) is proposed in this paper, which incorporates thermodynamic knowledge through four mechanisms: a differentiable thermodynamic residual layer that converts deviations from thermodynamic consistency relations into trainable diagnostic evidence; a dual-branch encoder with physics-gated fusion to arbitrate between physical and data-driven models at the sample level; label-free physics prior injection; and a composite physics-informed loss function. A physics-constrained SMOTE strategy restricts the synthetic minority samples to a range that does not violate thermodynamics. Validation of 921 end-of-line units with eight defect types and an imbalance ratio of 43.4:1 showed that PIRD-Net achieved 95.90% accuracy, 94.31% Macro-F1, and 92.99% Matthews correlation coefficient after 5 × 5 repeated stratified cross-validation, and performed better than the traditional classification methods and the label-free physics prior in imbalance-sensitive indicators. Accuracy is still 91.44% after adding 30% test noise to the natural process. Out-of-fold analysis shows that the errors are concentrated in mechanism-overlapping modes; the two rarest modes are correctly identified, and physics-dominated decisions are free from errors, so a low data-branch weight can be used as an internal reliability indicator.

Förderung

  • Scientific Research Start-Up Foundation for High-Level Talents of Taizhou University
  • Wuxi University Research Start-Up Fund for Introduced Talents
  • 2025 Jiangsu Provincial Industry-University-Research Cooperation Project

Zitieren

APA

Wang, J., Jin, X., Zhao, M., Gong, Y., & Chen, L. (2026). Thermodynamics-Informed Residual-Disentangled Network for Manufacturing Defect Diagnosis of Reciprocating Compressors Under Extreme Class Imbalance. Processes, 14(22), 3723. https://doi.org/10.3390/pr14223723

Vancouver

Wang J, Jin X, Zhao M, Gong Y, Chen L. Thermodynamics-Informed Residual-Disentangled Network for Manufacturing Defect Diagnosis of Reciprocating Compressors Under Extreme Class Imbalance. Processes. 2026;14(22):3723. doi:10.3390/pr14223723

BibTeX
@article{Wang2026,
  author = {Wang, Jiangqing and Jin, Xinqiao and Zhao, Mengke and Gong, Yuling and Chen, Lin},
  title = {Thermodynamics-Informed Residual-Disentangled Network for Manufacturing Defect Diagnosis of Reciprocating Compressors Under Extreme Class Imbalance},
  journal = {Processes},
  year = {2026},
  volume = {14},
  number = {22},
  pages = {3723},
  doi = {10.3390/pr14223723}
}
Kostenlos starten →