Scholarlink

<p>The on-machine measurement (OMM) system is subject to various error sources, making it challenging to identify and compensate for them. Existing error compensation methods necessitate extensive calibration experiments and the development of complex error identification and compensation techniques, making it difficult to ensure measurement accuracy for complex structures like curved surfaces. Therefore, this paper proposes a measurement data-driven method for OMM error compensation. The method is designed for a specific surface with complex geometry, and firstly, probe error compensation is realized by the method of standard sphere calibration, which compensates the probe error to a certain extent. Then, the surface is measured by coordinate measuring machine (CMM) and OMM system respectively. The error identification and compensation of the measurement results of specific geometrical structures are realized by designing a convolutional neural network (CNN). Simulation and experimental results demonstrate that the proposed method effectively enhances the compensation of measurement errors for specific geometries.</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

Data-driven compensation method for on-machine measurement errors of freeform surfaces

  • 期刊论文

  • 发表日期：

    2025年8月16日

查看全文
  • Zhen Sun,
  • Tao Wu,
  • Xiang Zuo,
  • Jiankun Liu,
  • Qiulin Hou,
  • Bofeng Fu,
  • Honggen Zhou,
  • Guochao Li

摘要

The on-machine measurement (OMM) system is subject to various error sources, making it challenging to identify and compensate for them. Existing error compensation methods necessitate extensive calibration experiments and the development of complex error identification and compensation techniques, making it difficult to ensure measurement accuracy for complex structures like curved surfaces. Therefore, this paper proposes a measurement data-driven method for OMM error compensation. The method is designed for a specific surface with complex geometry, and firstly, probe error compensation is realized by the method of standard sphere calibration, which compensates the probe error to a certain extent. Then, the surface is measured by coordinate measuring machine (CMM) and OMM system respectively. The error identification and compensation of the measurement results of specific geometrical structures are realized by designing a convolutional neural network (CNN). Simulation and experimental results demonstrate that the proposed method effectively enhances the compensation of measurement errors for specific geometries.