Submit

Login Login

Register Register

Home / Journals / CMC / Online First / doi:10.32604/cmc.2026.084014
Journal Menu
Special Issues
Table of Content

All issues

Online First

2026

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

Open Access

ARTICLE

A Data-Driven Fault Prediction Method for Bearing Ring CNC Grinding Machines

Yanan Wang, Xiaoying Yang*, Zhijie Pei, Xin Yang, Bo Li
School of Mechatronics Engineering, Henan University of Science and Technology, Luoyang, China
* Corresponding Author: Xiaoying Yang. Email: email

Computers, Materials & Continua https://doi.org/10.32604/cmc.2026.084014

Received 15 April 2026; Accepted 29 June 2026; Published online 28 July 2026

Abstract

Sudden faults in bearing ring computer numerical control (CNC) grinding machines significantly impact product processing quality and production efficiency, making precise state prediction urgent to avoid downtime risks. However, the numerous operational parameters collected on-site and the focus of existing methods on outputting fault labels without analyzing the evolution trends of the equipment’s operational state lead to unclear fault discrimination criteria and weak traceability, making it difficult to provide effective early-warning support during the incipient stages of a fault. To address these issues, this paper constructs a data-driven integrated algorithm adopting a “predict-then-classify” approach. First, the Pearson-ReliefF algorithm is utilized to eliminate redundant features and retain sensitive parameters. Second, the BiGRU-Attention algorithm is employed to capture bidirectional dependencies in time-series data, realizing the prediction of the equipment’s operational state trends. Finally, the Sparrow Search Algorithm (SSA) is introduced to optimize core Support Vector Machine (SVM) parameters, achieving precise identification of fault types in bearing ring CNC grinding machines. Experimental results indicate that the proposed algorithm exhibits robust performance in both the prediction and classification stages. The prediction metrics MAE, RMSE, and R2 are 0.0124, 0.0152, and 0.975, respectively, and the average multi-fault identification accuracy based on 10 repeated experiments with random seeds reaches 98.25%. This verifies the effectiveness of the method, which is of significant importance for ensuring the processing quality of bearing rings, reducing operation and maintenance costs, achieving intelligent online predictive diagnosis, and supporting the preventive maintenance of equipment.

Keywords

Bearing; CNC grinding machine; data-driven; fault prediction

  • 98

    View

  • 14

    Download

  • 0

    Like

Related articles
Copyright© 2026 Tech Science Press
© 1997-2026 TSP (Henderson, USA) unless otherwise stated

Share Link