Scholarlink

<p>Robotized ultrasonic multi-needle peening (UMNP) provides a promising approach for surface enhancement but suffers from roughness accumulation and progressive deformation due to direct tool-workpiece contact. To address these issues, this work proposes a curvature-adaptive oscillatory path strategy to effectively mitigate directional striations and adapt to geometric evolution. Furthermore, to resolve the trade-off between forming precision and surface integrity, an integrated constrained optimization framework is established. This framework leverages dual backpropagation neural networks (BPNNs) as surrogate models and utilizes particle swarm optimization (PSO) to identify optimal path parameters that minimize roughness under strict geometric constraints. Validated on an experimental platform, the proposed method significantly reduces surface roughness while ensuring high forming accuracy. These results demonstrate significant engineering potential for extending the fatigue life of thin-walled components through automated and precise surface integrity management.</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

Curvature-adaptive oscillatory path planning for robotized peen forming with surface roughness control via dual-surrogate modeling

  • 期刊论文

  • 发表日期：

    2026年5月30日

查看全文
  • Haoyu Guo,
  • Yafeng Tian,
  • Adnan Saifan,
  • Yanbo Shen,
  • Junjie Li,
  • Chi Zhang,
  • Guilin Yang,
  • Silu Chen

摘要

Robotized ultrasonic multi-needle peening (UMNP) provides a promising approach for surface enhancement but suffers from roughness accumulation and progressive deformation due to direct tool-workpiece contact. To address these issues, this work proposes a curvature-adaptive oscillatory path strategy to effectively mitigate directional striations and adapt to geometric evolution. Furthermore, to resolve the trade-off between forming precision and surface integrity, an integrated constrained optimization framework is established. This framework leverages dual backpropagation neural networks (BPNNs) as surrogate models and utilizes particle swarm optimization (PSO) to identify optimal path parameters that minimize roughness under strict geometric constraints. Validated on an experimental platform, the proposed method significantly reduces surface roughness while ensuring high forming accuracy. These results demonstrate significant engineering potential for extending the fatigue life of thin-walled components through automated and precise surface integrity management.