EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

MHA-PINN: A Novel Physics-Informed Neural Network for Predicting Fiber Dyeability

Feier Zhou, Yuxiang Liu, Shuo Yang, Fan Guo, Xiaofeng Yuan iD, Ruimin Xie

Sensors · 24.03.2026 · 26 (7) · 2018

DOI: 10.3390/s26072018 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Sensors, online erschienen am 24.03.2026. Verfasst von Feier Zhou und 5 weiteren Autor:innen. Beteiligte Einrichtung: School of Information and Intelligent Science, Donghua University, Shanghai 201620, China. Die Arbeit nennt 27 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Cultivation Project of Discipline Innovation, the Cultivation Project of Discipline Innovation, National Natural Science Foundation of China.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Fiber dyeability is a core indicator of textile quality and added value. Pre-experiment accurate prediction of fiber dyeability reduces the waste and inefficiency of trial-and-error methods. However, due to the limited data volume and complex mechanisms of fiber dyeability, there are no relevant studies to date. Thus, this paper proposes a novel prediction model integrating domain knowledge and process data called multi-head attention–physics-informed neural network (MHA-PINN). Within the MHA-PINN framework, limited experimental data is first augmented by using variational autoencoders, and subjected to ensemble feature selection on the augmented samples. Subsequently, a multi-head attention layer is introduced to capture the interdependencies among sample variables, thereby outputting a new feature matrix that represents the weighted fusion of these variables. Finally, a physics-informed neural network module embeds the dyeing kinetic equations into the loss function, guiding the model to converge towards accurate solutions for sample predictions. The effectiveness and superiority of the proposed MHA-PINN have been validated on a fiber dyeability experimental dataset.

Schlagwörter

Stichwörter der Autor:innen

data augmentationdomain knowledgefiber dyeabilityphysics-informed neural networksvariational autoencoder

Autor:innen und Institutionen

  • School of Information and Intelligent Science, Donghua University, Shanghai 201620, China.
  • College of Materials Science and Engineering, Donghua University, Shanghai 201620, China.
  • School of Automation, Nanjing Institute of Technology, Nanjing 211167, China.
  • School of Automation, Central South University, Changsha 410083, China.

ORCID-Profile: Xiaofeng Yuan

Förderung

  • Cultivation Project of Discipline Innovation
  • the Cultivation Project of Discipline Innovation
  • National Natural Science Foundation of China
  • Advanced Materials-National Science and Technology Major Project
  • Scientific and Technological Project in the Inner Mongolia Autonomous Region
  • cientific Research Foundation of Nanjing Institute of Technology
  • Natural Science Foundation of Jiangsu Fundamental Research Program
  • Scientific Research Foundation of Nanjing Institute of Technology

Zitieren

APA

Zhou, F., Liu, Y., Yang, S., Guo, F., Yuan, X., & Xie, R. (2026). MHA-PINN: A Novel Physics-Informed Neural Network for Predicting Fiber Dyeability. Sensors, 26(7), 2018. https://doi.org/10.3390/s26072018

Vancouver

Zhou F, Liu Y, Yang S, Guo F, Yuan X, Xie R. MHA-PINN: A Novel Physics-Informed Neural Network for Predicting Fiber Dyeability. Sensors. 2026;26(7):2018. doi:10.3390/s26072018

BibTeX
@article{Zhou2026,
  author = {Zhou, Feier and Liu, Yuxiang and Yang, Shuo and Guo, Fan and Yuan, Xiaofeng and Xie, Ruimin},
  title = {MHA-PINN: A Novel Physics-Informed Neural Network for Predicting Fiber Dyeability},
  journal = {Sensors},
  year = {2026},
  volume = {26},
  number = {7},
  pages = {2018},
  doi = {10.3390/s26072018}
}
Kostenlos starten →