EUVIMEDEUVIMED entdecken
Crossref

Physics-Informed Ensemble Feature Transfer for Energy Prediction in Generalized Milling Process

Yang Xie, Yiqun Dai, Fei Wang, Chaoyong Zhang

IEEE Transactions on Automation Science and Engineering · 2026 · 23 · 13302-13315

DOI: 10.1109/tase.2026.3714746 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Transactions on Automation Science and Engineering, 2026. Verfasst von Yang Xie und 3 weiteren Autor:innen. Die Arbeit nennt 41 Literaturangaben. Gefördert von: National Natural Science Foundation of China, Shandong Key Laboratory of Computer Numerical Control (CNC) Machine Tool Functional Components.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

The efficient lean management of energy consumption offers technical support for energy-saving manufacturing of CNC equipment. However, intricate processing scenarios result in differences in the distribution of energy characteristics, which seriously reduces prediction accuracy. Additionally, existing machine learning (ML) techniques depend on numerous labeled samples exhibiting identical probability distributions, resulting a dramatic deterioration in the performance of conventional prediction methods. Therefore, this paper builds a metamodel based on physics-informed feature transfer learning, which is used to predict energy consumption in general milling scenarios. Firstly, combining different scenarios of CNC machining, Gaussian process (GP) model is employed to construct the variable mapping relationship of energy consumption in source domain and target domain. Next, feature transfer and energy prediction are carried out based on the shared feature processor of XGBoost. Meanwhile, Wasserstein distance (WD) and Shapley additive explanations (SHAP) are applied to measure and reveal the distribution distance and feature correlation between source domain and target domain. The overall prediction accuracy of the proposed model has been verified to be over 98.5% through experiments in target domain, demonstrating that the metamodel can be applied to general CNC milling process, thereby further promoting the development of green manufacturing. Note to Practitioners—This article was motivated by the problem of low accuracy when the prediction model applies to different CNC machines. Previous research mainly focused on the energy consumption prediction problem of specific CNC equipments, and the models established did not have good universality and feasibility. This article suggests a physics and data-driven metamodel that can have equally excellent energy consumption prediction capabilities on different machine tools with the same process. In this paper, we mathematically adopt power feature transfer to achieve energy consumption feature mapping for different equipment processes, and utilize it as input for a hybrid prediction model to accurately predict energy consumption in the target domain. This method can utilize energy consumption prediction in different scenarios and can be extended to the field of process planning for multi process processing in the future, providing technical support for process path optimization.

Förderung

  • National Natural Science Foundation of China
  • Shandong Key Laboratory of Computer Numerical Control (CNC) Machine Tool Functional Components

Zitieren

APA

Xie, Y., Dai, Y., Wang, F., & Zhang, C. (2026). Physics-Informed Ensemble Feature Transfer for Energy Prediction in Generalized Milling Process. IEEE Transactions on Automation Science and Engineering, 23, 13302-13315. https://doi.org/10.1109/tase.2026.3714746

Vancouver

Xie Y, Dai Y, Wang F, Zhang C. Physics-Informed Ensemble Feature Transfer for Energy Prediction in Generalized Milling Process. IEEE Transactions on Automation Science and Engineering. 2026;23:13302-13315. doi:10.1109/tase.2026.3714746

BibTeX
@article{Xie2026,
  author = {Xie, Yang and Dai, Yiqun and Wang, Fei and Zhang, Chaoyong},
  title = {Physics-Informed Ensemble Feature Transfer for Energy Prediction in Generalized Milling Process},
  journal = {IEEE Transactions on Automation Science and Engineering},
  year = {2026},
  volume = {23},
  pages = {13302--13315},
  doi = {10.1109/tase.2026.3714746}
}
Kostenlos starten →