Scholarlink

<p>We propose a data-driven approach for monitoring the milling process quality condition. Audio-based frequencies and related wave plots are used for inference in two conditions: (i) machining process in a suitable state; and (ii) machining process in a non-suitable state (anomaly detection). Audio signals are first gathered from the computer numeric control (CNC) machine, and noise signals are processed to analyze in the wave, spectrogram, and frequency formats. These features are analyzed considering machining process parameters such as cutting speed, depth of cut, and feed rate. It is shown that the audio signals provide sufficient information about the machining capabilities of the CNC machine tool to produce new information about the machining quality without additional experimental measurements. The root mean square (RMS) values ranged from 0.05 to 0.24 after normalization, with the 90th percentile threshold computed as 0.18. Recordings associated with idle operations showed low RMS values (0.05–0.10), indicating stable, defect-free machining states. In contrast, representing active drilling with a 5&#xa0;mm tool at 2500&#xa0;rpm, exhibited the highest RMS value (0.24), exceeding the 90th percentile. This suggests a notable increase in mechanical load, vibration, or tool–workpiece interaction. We propose an attempt to use an audio signal-based machining process improvement study based on the controllable machining parameters adjustment, which is one step forward in integration with an adaptive control system development model that illustrates the main contribution of the paper to the literature.</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

Using the audio signal frequencies synthesized from the process to improve the computer-numeric control machining quality in precision manufacturing

  • 期刊论文

  • 发表日期：

    2025年10月3日

查看全文
  • Yusuf Tansel Ic,
  • Mustafa Sert,
  • Barış Kececi

摘要

We propose a data-driven approach for monitoring the milling process quality condition. Audio-based frequencies and related wave plots are used for inference in two conditions: (i) machining process in a suitable state; and (ii) machining process in a non-suitable state (anomaly detection). Audio signals are first gathered from the computer numeric control (CNC) machine, and noise signals are processed to analyze in the wave, spectrogram, and frequency formats. These features are analyzed considering machining process parameters such as cutting speed, depth of cut, and feed rate. It is shown that the audio signals provide sufficient information about the machining capabilities of the CNC machine tool to produce new information about the machining quality without additional experimental measurements. The root mean square (RMS) values ranged from 0.05 to 0.24 after normalization, with the 90th percentile threshold computed as 0.18. Recordings associated with idle operations showed low RMS values (0.05–0.10), indicating stable, defect-free machining states. In contrast, representing active drilling with a 5 mm tool at 2500 rpm, exhibited the highest RMS value (0.24), exceeding the 90th percentile. This suggests a notable increase in mechanical load, vibration, or tool–workpiece interaction. We propose an attempt to use an audio signal-based machining process improvement study based on the controllable machining parameters adjustment, which is one step forward in integration with an adaptive control system development model that illustrates the main contribution of the paper to the literature.