Scholarlink

<p>In ultra-precision machining of micro-structured surfaces, fast tool servo (FTS) is used for vibration-assisted turning; however, its dynamic response is limited because its resonant frequency remains low. A high-frequency bandwidth can enhance the machining efficiency in vibration-assisted turning for difficult-to-cut materials. To enhance the dynamic response, this research proposes a hybrid framework that combines topology–kinematic synthesis with physics-informed modeling to design a high-frequency FTS mechanism for vibration-assisted turning (VAT). The proposed methodology couples topology optimization with kinematic synthesis to establish a systematic design paradigm for flexure-based FTS mechanisms. To precisely model the relationship between frequency and stress, a Physics-Informed Deep Neural Network (PIDNN) integrated with Lagrangian analytical modeling is employed. The PIDNN embeds physical constraints, including stress and safety factors, into the learning process, while the Hunger Games Search algorithm is utilized to optimize the network parameters. A multi-objective constrained optimization procedure is subsequently applied to maximize the natural frequency while ensuring that stress and stroke remain within permissible limits. The finalized design achieves a resonant frequency of 3583&#xa0;Hz, a stroke of 29.95&#xa0;µm, and a maximum stress of 36.208&#xa0;MPa, as verified through both finite element analysis and experimental validation. The experimental results show excellent agreement with the simulation results, with deviations of less than 0.13%. Moreover, VAT experiments on SKD11 hardened steel confirm that the developed FTS significantly enhances surface quality, decreasing surface roughness from Ra = 1.9&#xa0;µm to Ra = 0.68&#xa0;µm.</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

Topology–kinematic synthesis and physics-informed dynamic modeling of a high-frequency fast tool servo for vibration-assisted turning

  • 期刊论文

  • 发表日期：

    2026年7月18日

查看全文
  • Tan Thang Nguyen,
  • Nhat Linh Ho,
  • Hieu Giang Le,
  • Thanh-Phong Dao

摘要

In ultra-precision machining of micro-structured surfaces, fast tool servo (FTS) is used for vibration-assisted turning; however, its dynamic response is limited because its resonant frequency remains low. A high-frequency bandwidth can enhance the machining efficiency in vibration-assisted turning for difficult-to-cut materials. To enhance the dynamic response, this research proposes a hybrid framework that combines topology–kinematic synthesis with physics-informed modeling to design a high-frequency FTS mechanism for vibration-assisted turning (VAT). The proposed methodology couples topology optimization with kinematic synthesis to establish a systematic design paradigm for flexure-based FTS mechanisms. To precisely model the relationship between frequency and stress, a Physics-Informed Deep Neural Network (PIDNN) integrated with Lagrangian analytical modeling is employed. The PIDNN embeds physical constraints, including stress and safety factors, into the learning process, while the Hunger Games Search algorithm is utilized to optimize the network parameters. A multi-objective constrained optimization procedure is subsequently applied to maximize the natural frequency while ensuring that stress and stroke remain within permissible limits. The finalized design achieves a resonant frequency of 3583 Hz, a stroke of 29.95 µm, and a maximum stress of 36.208 MPa, as verified through both finite element analysis and experimental validation. The experimental results show excellent agreement with the simulation results, with deviations of less than 0.13%. Moreover, VAT experiments on SKD11 hardened steel confirm that the developed FTS significantly enhances surface quality, decreasing surface roughness from Ra = 1.9 µm to Ra = 0.68 µm.