Scholarlink

<p>This study investigates the effectiveness of ultrasonic-assisted micro-milling (UAMM) in cortical bone machining by integrating controlled experimental procedures with machine learning-based predictive modeling. Sixty-four milling tests were conducted on bovine bone samples, systematically varying spindle speed, feed rate, tool diameter, depth of cut, vibration amplitude, and machining axis. Cutting forces and surface temperatures were accurately recorded using piezoelectric dynamometers and infrared thermometry. Four regression models Decision Tree, Random Forest, XGBoost, and Gradient Boosting were developed to predict force and temperature, with Gradient Boosting yielding the highest accuracy in force prediction (R = 0.92) and XGBoost outperforming others in temperature prediction (R = 0.98). Sensitivity analysis revealed spindle speed (43.03%) and machining axis (25.99%) as dominant factors influencing force, while machining axis (56.00%) and tool diameter (35.92%) were most influential on temperature. The results demonstrate that UAMM can significantly reduce cutting forces and control thermal buildup, thereby enhancing surgical safety and minimizing the risk of thermal necrosis in bone tissue.</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

Ultrasonic-Assisted Micro-Milling in Cortical Bone Cutting: Force and Temperature Prediction by Machine Learning Models on Experimental Results

  • 期刊论文

  • 发表日期：

    2025年8月25日

查看全文
  • Vahid Tahmasbi,
  • Amir Hossein Rabiee,
  • Mohammad Baraheni

摘要

This study investigates the effectiveness of ultrasonic-assisted micro-milling (UAMM) in cortical bone machining by integrating controlled experimental procedures with machine learning-based predictive modeling. Sixty-four milling tests were conducted on bovine bone samples, systematically varying spindle speed, feed rate, tool diameter, depth of cut, vibration amplitude, and machining axis. Cutting forces and surface temperatures were accurately recorded using piezoelectric dynamometers and infrared thermometry. Four regression models Decision Tree, Random Forest, XGBoost, and Gradient Boosting were developed to predict force and temperature, with Gradient Boosting yielding the highest accuracy in force prediction (R = 0.92) and XGBoost outperforming others in temperature prediction (R = 0.98). Sensitivity analysis revealed spindle speed (43.03%) and machining axis (25.99%) as dominant factors influencing force, while machining axis (56.00%) and tool diameter (35.92%) were most influential on temperature. The results demonstrate that UAMM can significantly reduce cutting forces and control thermal buildup, thereby enhancing surgical safety and minimizing the risk of thermal necrosis in bone tissue.