High-performance turning of WCu alloy using cryogenically treated tool with predictive machine learning approach and...

Area of Science:

• Materials Science and Engineering
• Manufacturing Processes
• Tribology

Background:

• Tungsten-copper (WCu) alloys present unique challenges in machining due to their combined properties.
• Optimizing cutting parameters and tool conditions is crucial for efficient WCu alloy processing.
• Cryogenic treatment is a potential method to enhance tool performance in machining.

Purpose of the Study:

• To evaluate the machinability of WCu alloy using cryogenically treated and conventional tungsten carbide tools.
• To assess the impact of cutting parameters (speed, feed, depth of cut) on surface roughness, tool wear, and cutting force.
• To develop predictive models for machining responses using machine learning algorithms.

Main Methods:

• Taguchi L27 orthogonal array design for experimentation.
• Cryogenic treatment of tools at -196°C for 24 hours.
• Application of regression analysis, machine learning (Linear Regression, Random Forest, SVR, ANN), and Grey Relational Analysis (GRA).

Main Results:

• Cryogenic treatment enhanced tool hardness and edge stability, leading to reduced wear, lower cutting forces, and improved surface finish.
• Random Forest model achieved high prediction accuracy for cutting force (R²=0.96) under cryogenic conditions.
• Surface roughness prediction accuracy was limited, suggesting the influence of unmeasured dynamic factors.
• Grey Relational Analysis identified optimal parameters: 1200 rpm cutting speed, 0.04 mm/rev feed rate, and 0.5 mm depth of cut.

Conclusions:

• Cryogenic treatment offers substantial benefits for WCu alloy machinability.
• Data-driven predictive modeling, particularly Random Forest, aids in optimizing cutting force prediction.
• Integrated multi-response optimization using GRA provides robust parameter settings for precision turning.