EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

SELF-SUPERVISED ADAPTIVE STATISTICAL PROCESS CONTROL USING SPECTRAL CONTRASTIVE LEARNING AND CAUSAL GRAPH INTELLIGENCE FOR AUTONOMOUS QUALITY MONITORING IN SMART MANUFACTURING PROCESSES

V. Anusha, Lilly George, K. S. Harish

Universal Journal of Mathematics and Mathematical Sciences · 23.06.2026 · 22 (2) · 119-156

DOI: 10.17654/2277141726006 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Universal Journal of Mathematics and Mathematical Sciences, online erschienen am 23.06.2026. Verfasst von V. Anusha und 2 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Quality control using conventional Statistical Process Control (SPC) is limited to stationary, label-based approaches owing to the high cost of labeling defects and the inability of these approaches to accommodate non-stationarity in manufacturing processes. In practice, this limitation results from the scarcity of available labeled data and the changing nature of manufacturing processes. The proposed contribution is an adaptive Self-Supervised Statistical Process Control (SS-SPC) framework that enables label-free quality monitoring. The framework includes the following: (1) A spectral contrastive process representation network that learns latent representations of the process; (2) Entropy-regularized clustering method to identify regimes; (3) A Neural Statistical Field (NSF) that produces adaptive control limits; (4) A Causal Graph Network (CGN) to identify the root cause of failures and (5) Reinforcement-guided controller to optimize process parameters. The experimental evaluation shows better regime separations and fewer false alarms, as well as an increased ability to detect defects and improve process stability when the manufacturing process is changing.

Zitieren

APA

Anusha, V., George, L., & Harish, K. S. (2026). SELF-SUPERVISED ADAPTIVE STATISTICAL PROCESS CONTROL USING SPECTRAL CONTRASTIVE LEARNING AND CAUSAL GRAPH INTELLIGENCE FOR AUTONOMOUS QUALITY MONITORING IN SMART MANUFACTURING PROCESSES. Universal Journal of Mathematics and Mathematical Sciences, 22(2), 119-156. https://doi.org/10.17654/2277141726006

Vancouver

Anusha V, George L, Harish KS. SELF-SUPERVISED ADAPTIVE STATISTICAL PROCESS CONTROL USING SPECTRAL CONTRASTIVE LEARNING AND CAUSAL GRAPH INTELLIGENCE FOR AUTONOMOUS QUALITY MONITORING IN SMART MANUFACTURING PROCESSES. Universal Journal of Mathematics and Mathematical Sciences. 2026;22(2):119-156. doi:10.17654/2277141726006

BibTeX
@article{Anusha2026,
  author = {Anusha, V. and George, Lilly and Harish, K. S.},
  title = {SELF-SUPERVISED ADAPTIVE STATISTICAL PROCESS CONTROL USING SPECTRAL CONTRASTIVE LEARNING AND CAUSAL GRAPH INTELLIGENCE FOR AUTONOMOUS QUALITY MONITORING IN SMART MANUFACTURING PROCESSES},
  journal = {Universal Journal of Mathematics and Mathematical Sciences},
  year = {2026},
  volume = {22},
  number = {2},
  pages = {119--156},
  doi = {10.17654/2277141726006}
}
Kostenlos starten →