Scholarlink

<p>This study presents a cost-effective laser-based system for real-time geometric measurement in material extrusion (MEX) printing. By integrating affordable sensors, including a laser positioning module and photoresistors, onto a commercial MEX machine, the system enables in situ geometric monitoring. As the print platform moves parts through the laser light, signal changes are captured to reconstruct the part’s 2D geometry in real-time. A reference database links sensor signals to actual dimensions, allowing height and depth measurements with reasonable accuracy. To improve measurement reliability, the system incorporates advanced signal processing techniques, including an unsupervised learning method, to enhance robustness and precision under varying geometric and environmental conditions. The system was validated using test specimens with diverse geometries and demonstrated effective detection of common MEX process defects, such as layer shifting and internal voids. A reference standard and image-based geometric comparison technique support automatic error detection within a closed-loop process control framework. If deviations exceed predefined thresholds, a custom removal mechanism eliminates defective parts, allowing immediate reprints. While the system shows promise in improving process monitoring, there are limitations in measurement precision and broader applicability to complex geometries. Nevertheless, this approach offers a practical step toward real-time in situ geometry monitoring, with potential benefits in process efficiency, material savings, and quality control. Key contributions include the development of a low-cost laser scanning system, an automated workflow for height and depth measurements geometric measurement, and an error detection framework an integrated error detection and removal mechanism to support further advancements in additive manufacturing process control.</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

In situ optical monitoring for geometric accuracy and error detection in extrusion-based additive manufacturing

  • 期刊论文

  • 发表日期：

    2025年6月26日

查看全文
  • Mei-Jyun Lin,
  • Dian-Ru Li

摘要

This study presents a cost-effective laser-based system for real-time geometric measurement in material extrusion (MEX) printing. By integrating affordable sensors, including a laser positioning module and photoresistors, onto a commercial MEX machine, the system enables in situ geometric monitoring. As the print platform moves parts through the laser light, signal changes are captured to reconstruct the part’s 2D geometry in real-time. A reference database links sensor signals to actual dimensions, allowing height and depth measurements with reasonable accuracy. To improve measurement reliability, the system incorporates advanced signal processing techniques, including an unsupervised learning method, to enhance robustness and precision under varying geometric and environmental conditions. The system was validated using test specimens with diverse geometries and demonstrated effective detection of common MEX process defects, such as layer shifting and internal voids. A reference standard and image-based geometric comparison technique support automatic error detection within a closed-loop process control framework. If deviations exceed predefined thresholds, a custom removal mechanism eliminates defective parts, allowing immediate reprints. While the system shows promise in improving process monitoring, there are limitations in measurement precision and broader applicability to complex geometries. Nevertheless, this approach offers a practical step toward real-time in situ geometry monitoring, with potential benefits in process efficiency, material savings, and quality control. Key contributions include the development of a low-cost laser scanning system, an automated workflow for height and depth measurements geometric measurement, and an error detection framework an integrated error detection and removal mechanism to support further advancements in additive manufacturing process control.