Scholarlink

<p>Effective tool condition monitoring (TCM) is essential for ensuring machining efficiency, product quality, and reliable identification of tool state in modern manufacturing. However, many existing TCM approaches rely on manual extraction of tool-wear features, which introduces subjectivity, limits scalability, and hinders real-time implementation. To address these limitations, this study proposes a novel vision-based tool wear indicator, termed Total Pixel Area (TPA), for real-time tool wear evaluation. A dedicated vision system is developed to capture projected tool images during rotation, in which white-pixel distributions are extracted to identify and classify wear states, overcoming the constraints of conventional microscopic image analysis. In parallel, sensor signals are acquired to monitor tool-life progression, thereby enabling multimodal condition assessment. The feasibility and robustness of the proposed approach are experimentally validated by correlating TPA with flank wear (VB) measurements and with signal features extracted from the time, frequency, and time–frequency domains across various milling parameters and tool geometries. Experiments are conducted using HSCO, tungsten carbide, and TiAlN-coated tools. The results demonstrate that the proposed TPA indicator exhibits high reliability and consistency, achieving correlation coefficients ranging from 96.63% to 98.58% with VB across different conditions. Moreover, strong correlations are observed between the TPA wear index and cutting-signal features. Unlike conventional methods, the proposed approach enables simultaneous detection of radial and axial wear, providing a more comprehensive and accurate real-time assessment of tool condition.</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 vision-based total pixel area wear indicator with sensor signal correlation for real-time tool condition monitoring

  • 期刊论文

  • 发表日期：

    2026年7月14日

查看全文
  • Ahmed Abdeltawab,
  • Mohamed T. Eraky,
  • Zhang Xi,
  • Zhang Longjia,
  • Chuanjun LI,
  • Ossama B. Abouelatta,
  • Abdelkhalik Eladl

摘要

Effective tool condition monitoring (TCM) is essential for ensuring machining efficiency, product quality, and reliable identification of tool state in modern manufacturing. However, many existing TCM approaches rely on manual extraction of tool-wear features, which introduces subjectivity, limits scalability, and hinders real-time implementation. To address these limitations, this study proposes a novel vision-based tool wear indicator, termed Total Pixel Area (TPA), for real-time tool wear evaluation. A dedicated vision system is developed to capture projected tool images during rotation, in which white-pixel distributions are extracted to identify and classify wear states, overcoming the constraints of conventional microscopic image analysis. In parallel, sensor signals are acquired to monitor tool-life progression, thereby enabling multimodal condition assessment. The feasibility and robustness of the proposed approach are experimentally validated by correlating TPA with flank wear (VB) measurements and with signal features extracted from the time, frequency, and time–frequency domains across various milling parameters and tool geometries. Experiments are conducted using HSCO, tungsten carbide, and TiAlN-coated tools. The results demonstrate that the proposed TPA indicator exhibits high reliability and consistency, achieving correlation coefficients ranging from 96.63% to 98.58% with VB across different conditions. Moreover, strong correlations are observed between the TPA wear index and cutting-signal features. Unlike conventional methods, the proposed approach enables simultaneous detection of radial and axial wear, providing a more comprehensive and accurate real-time assessment of tool condition.