A multi-dimensional data integration approach for machining system multi-level energy consumption modelling and prediction

Chen, Yuxuan, Zhang, Hua, Yan, Wei, Zhang, Xumei, Jiang, Zhigang and Liu, Ying ORCID: https://orcid.org/0000-0001-9319-5940 2026. A multi-dimensional data integration approach for machining system multi-level energy consumption modelling and prediction. International Journal of Computer Integrated Manufacturing 39 (2) , pp. 245-266. 10.1080/0951192X.2025.2478011

PDF - Accepted Post-Print Version Available under License Creative Commons Attribution. Download (2MB)

Abstract

Multi-level energy consumption modeling and prediction (MECMP) of machining systems is crucial for refined energy management (REM). Due to the complexity of the coupling relationship between system levels (equipment, part and process) and corresponding energy influencing factors, it is difficult to select appropriate data for MECMP. Moreover, the multi-dimensional nature of the influencing factors also poses a challenge to integrate them for energy consumption modeling. To fill these gaps, this paper proposes a multi-dimensional data integration (MDDI) approach for MECMP. First, the associated influencing factors are identified by analyzing the energy consumption nature of machining systems, and a data cube model is established to facilitate efficient MDDI. Second, energy consumption models at three levels are developed to establish the correlation between energy consumption and multi-dimensional data. Then, a convolutional neural network with long short-term memory (CNN-LSTM) model is designed for energy consumption prediction. Finally, a milling case is exploited to verify the effectiveness of the proposed method. The results show that the proposed model achieves 99% accuracy in multi-level energy consumption prediction, and MDDI improves the accuracy by 7–8% compared to single dimension data. Furthermore, CNN-LSTM model outperforms the benchmark model in prediction accuracy and response time, thereby supporting REM implementation and promoting sustainable manufacturing.

Item Type:	Article
Date Type:	Publication
Status:	Published
Schools:	Schools > Engineering
Additional Information:	RRS policy applied
Publisher:	Taylor and Francis Group
ISSN:	0951-192X
Date of First Compliant Deposit:	13 March 2025
Date of Acceptance:	4 March 2025
Last Modified:	12 Feb 2026 14:15
URI:	https://orca.cardiff.ac.uk/id/eprint/176848

Actions (repository staff only)

Edit Item