Application of Machine Tool Thermal Error Compensation in Digital Twin-based System

Thermal error significantly affects machining accuracy, demanding careful control. A robust system integrating a highly accurate thermal error model is the key to this control. This can be achieved through deep learning-based models. However, even such methods are sensitive to problems like collinea...

Full description

Saved in:
Bibliographic Details
Main Authors: MA Chi, LI Minging, LIU Jialan, HE Jialong, HUA Chunlei, WANG Liang
Format: Article
Language:zho
Published: Harbin University of Science and Technology Publications 2025-02-01
Series:Journal of Harbin University of Science and Technology
Subjects:
Online Access:https://hlgxb.hrbust.edu.cn/#/digest?ArticleID=2405
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Thermal error significantly affects machining accuracy, demanding careful control. A robust system integrating a highly accurate thermal error model is the key to this control. This can be achieved through deep learning-based models. However, even such methods are sensitive to problems like collinearities among temperature variables, manual parameter tuning, and limited real-time compensation capabilities. To address these issues, the application of thermal error compensation in the digital twin-based system is studied. The digital twin-based thermal error compensation system (DTTECS) is designed to guarantee the real-time performance of the thermal error compensation process. Moreover, a novel collinearity exclusion-based thermal error method is proposed in this work based on the modified tolerance value. A novel prediction model utilizing a strong-convergence chimp optimization algorithm combined with a minimal gated unit-attention mechanism is proposed to characterize the dependence of the current thermal error on the historical thermal informant data, forming strong-convergence chimp optimization algorithm-minimal gated unit-attention ( SC-ChOA-MGU-A) model,ensuring data integrity. A nonlinear convergence factor is proposed for chimp optimization algorithm to enhance computational speed and optimize hyper-parameters. The key advantages over existing models include enhanced data integrity, faster computation, and higher prediction accuracy. Comparative analysis reveals the superior performance of the SC-ChOA-MGU-A model in terms of fitting accuracy, convergence rate, and predictive accuracy. The implementation of this digital twin system was crucial to reducing the geometric errors affecting the most critical sizes of the considered benchmark by about 75. 00%.
ISSN:1007-2683