Scholarlink

<p>The establishment of a feed drive system model capable of high-precision position prediction facilitates the development of high-fidelity digital twins for CNC machine tools and provides a solid foundation for model predictive control. To address the limitations of existing models in position prediction accuracy and structural complexity, this paper proposes a novel data-driven model named the Hybrid Response Model (HRM). Inspired by finite impulse response (FIR) model, the HRM parameterizes the feed drive system using impulse response sequences. Unlike conventional FIR models, this study effectively suppresses cumulative errors in position prediction through an error substitution method. Furthermore, the nonlinear modeling capability is significantly enhanced via friction reconstruction. Experimental results on both the single-axis test bench and the 3-axis machine tool demonstrate that the HRM achieves position prediction accuracies of 0.25 <InlineEquation ID="IEq1"> <EquationSource Format="TEX">\(\mu\)</EquationSource> </InlineEquation>m and 1.91 <InlineEquation ID="IEq2"> <EquationSource Format="TEX">\(\mu\)</EquationSource> </InlineEquation>m respectively, reaching levels comparable to the dynamic repeatability of their respective systems. Compared with existing data-driven models, the proposed model not only features a more elegant structure but also exhibits superior position prediction accuracy and robustness capability.</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

Hybrid response model: an elegant data-driven model for feed drive systems with high precision in position prediction

  • 期刊论文

  • 发表日期：

    2026年2月11日

查看全文
  • Donghao Li,
  • Huicheng Zhou,
  • Guangda Xu,
  • Jihong Chen

摘要

The establishment of a feed drive system model capable of high-precision position prediction facilitates the development of high-fidelity digital twins for CNC machine tools and provides a solid foundation for model predictive control. To address the limitations of existing models in position prediction accuracy and structural complexity, this paper proposes a novel data-driven model named the Hybrid Response Model (HRM). Inspired by finite impulse response (FIR) model, the HRM parameterizes the feed drive system using impulse response sequences. Unlike conventional FIR models, this study effectively suppresses cumulative errors in position prediction through an error substitution method. Furthermore, the nonlinear modeling capability is significantly enhanced via friction reconstruction. Experimental results on both the single-axis test bench and the 3-axis machine tool demonstrate that the HRM achieves position prediction accuracies of 0.25 \(\mu\) m and 1.91 \(\mu\) m respectively, reaching levels comparable to the dynamic repeatability of their respective systems. Compared with existing data-driven models, the proposed model not only features a more elegant structure but also exhibits superior position prediction accuracy and robustness capability.