Scholarlink

The Inconel 718 nickel-based super alloy has been in-demand material for aerospace industries because of their high toughness and high temperature withstanding properties. But the conventional machining of Inconel 718 is challenging for industries due to catastrophic failure of cutting tool during machining, thereby resulting in poor machining efficiency and greater tool wear rate. So to overcome these poor results, current research is focused on adaptation of non-traditional machining—Wire Electrical Discharge Machining (Wire EDM) for machining of Inconel 718. An efficient Machine Learning (ML) technique is developed to predict optimal range of machining parameters (Current, Pulse ON, Pulse OFF) for an efficient dimensional accuracy with reduced Tool Wear Rate (TWR). A total of 40 number of holes with varying diameter (ranging from 2 mm to 15 mm) are fine drilled by Wire EDM on a block of Inconel 718 with dimensions of (153 × 20 × 20) mm. The developed ML technique performs Principal Component Analysis (PCA) to identify Independent variables or Features (Current, Pulse ON, Pulse OFF) that are majorly effecting Targets (TWR, Material Removal Rate (MRR), machining TIME). Next, the Feature Selection by Lasso Regression indicates Pulse ON as most effective feature for TWR & MRR variations. Due to complex and non-linear relationship between Feature & Targets, the non-linear regression model - Random Forest (RF) approach is adapted to predict optimal range of machining parameters. RF results illustrate optimal range of Pulse ON (27–37 μs) to obtain efficient MRR of 14240.777558 & reduced TWR of 0.000599 for Wire EDM process on Inconel 718.

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

Generative Machine Learning Technique for Wire Electrical Discharge Machining Optimization of Inconel 718 – A Predictive Maintenance Approach

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Pramodkumar S. Kataraki,
  • Vishwanath Koti,
  • Harsha Kusnoorkar,
  • Sha Muzammil Quadri,
  • Ahmad Baharuddin Abdullah

摘要

The Inconel 718 nickel-based super alloy has been in-demand material for aerospace industries because of their high toughness and high temperature withstanding properties. But the conventional machining of Inconel 718 is challenging for industries due to catastrophic failure of cutting tool during machining, thereby resulting in poor machining efficiency and greater tool wear rate. So to overcome these poor results, current research is focused on adaptation of non-traditional machining—Wire Electrical Discharge Machining (Wire EDM) for machining of Inconel 718. An efficient Machine Learning (ML) technique is developed to predict optimal range of machining parameters (Current, Pulse ON, Pulse OFF) for an efficient dimensional accuracy with reduced Tool Wear Rate (TWR). A total of 40 number of holes with varying diameter (ranging from 2 mm to 15 mm) are fine drilled by Wire EDM on a block of Inconel 718 with dimensions of (153 × 20 × 20) mm. The developed ML technique performs Principal Component Analysis (PCA) to identify Independent variables or Features (Current, Pulse ON, Pulse OFF) that are majorly effecting Targets (TWR, Material Removal Rate (MRR), machining TIME). Next, the Feature Selection by Lasso Regression indicates Pulse ON as most effective feature for TWR & MRR variations. Due to complex and non-linear relationship between Feature & Targets, the non-linear regression model - Random Forest (RF) approach is adapted to predict optimal range of machining parameters. RF results illustrate optimal range of Pulse ON (27–37 μs) to obtain efficient MRR of 14240.777558 & reduced TWR of 0.000599 for Wire EDM process on Inconel 718.