Scholarlink

<p>Thermal errors have a significant impact on the precision of computer numerical control (CNC) machine tools, particularly in high-speed spindle systems such as the BL20-HSY. This study presents a novel and data-efficient framework for thermal error modeling and compensation, designed to enhance machining accuracy under thermal load conditions. An experimental platform was developed to collect comprehensive thermal data from critical locations on the spindle structure. To improve the selection for temperature-sensitive points, fuzzy C-means (FCM) clustering, optimized through ant colony optimization (ACO), was employed. Additionally, a multivariable grey correlation analysis (MGCA) was conducted to further refine sensor selection by identifying critical points that exhibited a high correlation with thermal deviations. To overcome the challenges associated with small datasets, a robust model-agnostic meta-learning (MAML) approach was integrated with the grey wolf optimizer (GWO), resulting in an adaptive model capable of generalizing across various operating conditions. Experimental validation conducted on the T65 CNC lathe demonstrated that the proposed MAML-GWO model achieved a reduction of over 90% in thermal deviation along the <i>X</i> and <i>Y</i> axes, with a predictive accuracy of <i>R</i><sup>2</sup> values reaching 0.97. These findings indicate that the model is suitable for real-time thermal error compensation in structurally similar CNC machine tools.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A novel robust thermal error modeling and compensation for slant bed CNC machines tool using small dataset based feature selection

  • 期刊论文

  • 发表日期：

    2026年6月3日

查看全文
  • Musab Alataiqeh,
  • Hu Shi,
  • Hai-Tao Wang

摘要

Thermal errors have a significant impact on the precision of computer numerical control (CNC) machine tools, particularly in high-speed spindle systems such as the BL20-HSY. This study presents a novel and data-efficient framework for thermal error modeling and compensation, designed to enhance machining accuracy under thermal load conditions. An experimental platform was developed to collect comprehensive thermal data from critical locations on the spindle structure. To improve the selection for temperature-sensitive points, fuzzy C-means (FCM) clustering, optimized through ant colony optimization (ACO), was employed. Additionally, a multivariable grey correlation analysis (MGCA) was conducted to further refine sensor selection by identifying critical points that exhibited a high correlation with thermal deviations. To overcome the challenges associated with small datasets, a robust model-agnostic meta-learning (MAML) approach was integrated with the grey wolf optimizer (GWO), resulting in an adaptive model capable of generalizing across various operating conditions. Experimental validation conducted on the T65 CNC lathe demonstrated that the proposed MAML-GWO model achieved a reduction of over 90% in thermal deviation along the X and Y axes, with a predictive accuracy of R2 values reaching 0.97. These findings indicate that the model is suitable for real-time thermal error compensation in structurally similar CNC machine tools.