Scholarlink

<p>Shape memory alloys (SMAs) are a specific category of smart materials established for their superior mechanical, physical, and biomedical characteristics. SMAs demonstrate two significant behaviours- pseudoplasticity and shape memory effect which produce them highly beneficial for advanced operational uses. To use these materials in real components, they must be accurately machined into precise shapes. In the present study, WEDM was used to machine a Ti₅₀Ni₄₀Co₁₀ shape memory alloy, and key machining performance measures namely surface roughness (SR) and material removal rate (MRR) were evaluated. In addition, the machined surfaces were thoroughly characterized in terms of morphology and topography. While ANN is employed to optimize process parameters such as voltage and for the best possible surface roughness, and MRR. ANN findings showed 0.96 for validation, 0.93 for testing, and 0.97 overall, efficient parameter tuning of Wire EDM variables for the SMA, achieving strong correlation coefficients (R-values) of 0.99 for training. The most suitable verification efficiency is obtained with a MSE (Mean Squared Error) of 0.729 at phase 5, after training completion within 11phase. These findings suggest that the ANN model is reliable and steady for estimating the operating parameters.</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

Data-driven machine learning modelling in wire EDM of TiNiCo shape memory alloy

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年3月3日

查看全文
  • Rashi Tyagi,
  • Hargovind Soni,
  • Ashutosh Tripathi,
  • S M Nabil Rahman,
  • Charles Mbohwa,
  • C. Durga Prasad,
  • Sunil Kumar Tiwari

摘要

Shape memory alloys (SMAs) are a specific category of smart materials established for their superior mechanical, physical, and biomedical characteristics. SMAs demonstrate two significant behaviours- pseudoplasticity and shape memory effect which produce them highly beneficial for advanced operational uses. To use these materials in real components, they must be accurately machined into precise shapes. In the present study, WEDM was used to machine a Ti₅₀Ni₄₀Co₁₀ shape memory alloy, and key machining performance measures namely surface roughness (SR) and material removal rate (MRR) were evaluated. In addition, the machined surfaces were thoroughly characterized in terms of morphology and topography. While ANN is employed to optimize process parameters such as voltage and for the best possible surface roughness, and MRR. ANN findings showed 0.96 for validation, 0.93 for testing, and 0.97 overall, efficient parameter tuning of Wire EDM variables for the SMA, achieving strong correlation coefficients (R-values) of 0.99 for training. The most suitable verification efficiency is obtained with a MSE (Mean Squared Error) of 0.729 at phase 5, after training completion within 11phase. These findings suggest that the ANN model is reliable and steady for estimating the operating parameters.