Scholarlink

<p>This study presents an artificial intelligence-driven approach to predict the corrosion behavior of Electrical Discharge Machining (EDM)-treated Co–Cr biomedical alloys. By integrating experimental data from 18 samples with Gaussian Process Regression (GPR), a machine learning model was developed to forecast corrosion rates across all 162 possible combinations of EDM parameters (discharge current, pulse-on/off time, electrode material, and dielectric medium). The model demonstrated high predictive accuracy (R<sup>2</sup> = 0.96, RMSE = 0.0018&#xa0;mm/year) through leave-one-out cross-validation. Key findings reveal non-linear effects of parameters, with optimal corrosion resistance achieved using tungsten-copper electrodes in deionized water at intermediate currents. This AI-based methodology significantly reduces experimental effort while providing actionable insights for optimizing EDM processing to enhance the long-term biocompatibility and durability of Co–Cr implants. SEM–EDS analysis of the best-performing sample confirmed uniform morphology and protective oxide layers, supporting the AI-driven insights. This approach accelerates biomaterial optimization, enhancing biocompatibility and durability for clinical applications.</p> Graphical abstract <p></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

Artificial intelligence–driven prediction of corrosion behavior of EDM-treated Co–Cr biomedical alloys

  • 期刊论文

  • 发表日期：

    2026年4月30日

查看全文
  • Amit Mahajan,
  • Parneet Kaur,
  • Sandeep Devgan,
  • Gurcharan Singh,
  • Sarabjeet Singh Sidhu

摘要

This study presents an artificial intelligence-driven approach to predict the corrosion behavior of Electrical Discharge Machining (EDM)-treated Co–Cr biomedical alloys. By integrating experimental data from 18 samples with Gaussian Process Regression (GPR), a machine learning model was developed to forecast corrosion rates across all 162 possible combinations of EDM parameters (discharge current, pulse-on/off time, electrode material, and dielectric medium). The model demonstrated high predictive accuracy (R2 = 0.96, RMSE = 0.0018 mm/year) through leave-one-out cross-validation. Key findings reveal non-linear effects of parameters, with optimal corrosion resistance achieved using tungsten-copper electrodes in deionized water at intermediate currents. This AI-based methodology significantly reduces experimental effort while providing actionable insights for optimizing EDM processing to enhance the long-term biocompatibility and durability of Co–Cr implants. SEM–EDS analysis of the best-performing sample confirmed uniform morphology and protective oxide layers, supporting the AI-driven insights. This approach accelerates biomaterial optimization, enhancing biocompatibility and durability for clinical applications.

Graphical abstract