Scholarlink

<p>To address the challenge of time-varying characteristics of the tool influence function (TIF) and insufficient surface form convergence induced by dynamic factors such as polishing pad wear and polishing slurry concentration fluctuation in the ultra-precision polishing process, this paper proposes a process framework-level closed-loop optimization method based on online iterative correction of the TIF. Different from the traditional empirical trial-and-error mode that solely relies on the superposition of surface form residuals, the proposed method establishes a self-learning control framework with the capability of adaptive updating of underlying parameters. Its core lies in the introduction of an online dynamic correction algorithm based on residual minimization. Through least-squares inversion of the measured surface form error and dwell time data from the previous processing round, the comprehensive removal efficiency coefficient k of the Preston model is quantitatively solved and calibrated online. To verify the universality and robustness of the proposed method, multiple rounds of continuous single-run processing verification were conducted on circular fused silica elements (polished with a polyurethane small tool) and square glass-ceramic elements (processed via bonnet polishing), respectively. The experimental results show that, affected by dynamic time-varying disturbances, the prediction accuracy of the first-round surface form Peak-to-Valley (PV) value using the traditional static model is only 61.4% and 63.4%, respectively. In contrast, after dynamic parameter calibration is implemented with the proposed closed-loop framework, the form correction prediction accuracy of subsequent rounds jumps to 91.1% and 99.9%, respectively. Meanwhile, the convergence trajectory of the parameter k accurately reflects the life cycle characteristics of dynamic wear of the polishing tool. This method effectively counteracts time-varying disturbances under complex working conditions, greatly improves the predictability of form correction, and provides a highly universal and robust process solution for the automated and deterministic manufacturing of high-precision optical elements.</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

Optimization of Optical Component Shaping and Polishing Process Based on Iterative Correction of Material Removal Function

  • 期刊论文

  • 发表日期：

    2026年5月27日

查看全文
  • Ri Pan,
  • Chuqi Chen,
  • Zhigang Yuan,
  • Yang Li,
  • Dongju Chen,
  • Bin Liu

摘要

To address the challenge of time-varying characteristics of the tool influence function (TIF) and insufficient surface form convergence induced by dynamic factors such as polishing pad wear and polishing slurry concentration fluctuation in the ultra-precision polishing process, this paper proposes a process framework-level closed-loop optimization method based on online iterative correction of the TIF. Different from the traditional empirical trial-and-error mode that solely relies on the superposition of surface form residuals, the proposed method establishes a self-learning control framework with the capability of adaptive updating of underlying parameters. Its core lies in the introduction of an online dynamic correction algorithm based on residual minimization. Through least-squares inversion of the measured surface form error and dwell time data from the previous processing round, the comprehensive removal efficiency coefficient k of the Preston model is quantitatively solved and calibrated online. To verify the universality and robustness of the proposed method, multiple rounds of continuous single-run processing verification were conducted on circular fused silica elements (polished with a polyurethane small tool) and square glass-ceramic elements (processed via bonnet polishing), respectively. The experimental results show that, affected by dynamic time-varying disturbances, the prediction accuracy of the first-round surface form Peak-to-Valley (PV) value using the traditional static model is only 61.4% and 63.4%, respectively. In contrast, after dynamic parameter calibration is implemented with the proposed closed-loop framework, the form correction prediction accuracy of subsequent rounds jumps to 91.1% and 99.9%, respectively. Meanwhile, the convergence trajectory of the parameter k accurately reflects the life cycle characteristics of dynamic wear of the polishing tool. This method effectively counteracts time-varying disturbances under complex working conditions, greatly improves the predictability of form correction, and provides a highly universal and robust process solution for the automated and deterministic manufacturing of high-precision optical elements.