Scholarlink

<p>A compact clamping fixture integrating multi-axis force sensing and a tri-axial MEMS accelerometer was developed to stabilize workpieces and enable in-process monitoring during micro-drilling with tungsten carbide drills. Sensor streams were synchronized and transmitted to an edge-computing unit for on-site preprocessing and feature extraction. Tool condition was characterized using low-complexity statistical indicators derived from both time- and frequency-domain signals, including FFT-based spectral features and correlation-based measures between orthogonal axes tracked over drilling cycles. The combined force–vibration measurements captured progressive changes associated with wear evolution and provided complementary sensitivity to abrupt fracture events. The proposed architecture provides a practical approach for real-time tool condition monitoring without reliance on machine-learning classifiers, supporting predictive maintenance and process optimization in precision micro-drilling.</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

Edge-computing-enabled statistical monitoring of micro tungsten carbide drill wear in CNC machining

  • 期刊论文

  • 发表日期：

    2026年3月2日

查看全文
  • Meng-Hsun Tsai,
  • Syuan-Cheng Chang,
  • Fu-Tun Chang

摘要

A compact clamping fixture integrating multi-axis force sensing and a tri-axial MEMS accelerometer was developed to stabilize workpieces and enable in-process monitoring during micro-drilling with tungsten carbide drills. Sensor streams were synchronized and transmitted to an edge-computing unit for on-site preprocessing and feature extraction. Tool condition was characterized using low-complexity statistical indicators derived from both time- and frequency-domain signals, including FFT-based spectral features and correlation-based measures between orthogonal axes tracked over drilling cycles. The combined force–vibration measurements captured progressive changes associated with wear evolution and provided complementary sensitivity to abrupt fracture events. The proposed architecture provides a practical approach for real-time tool condition monitoring without reliance on machine-learning classifiers, supporting predictive maintenance and process optimization in precision micro-drilling.