Hybrid ABC Optimized MARS-Based Modeling of the Milling Tool Wear from Milling Run Experimental Data

Area of Science:

• Manufacturing Engineering
• Materials Science
• Computational Intelligence

Background:

• Milling cutters are critical in machining but susceptible to wear and failure.
• Predicting tool wear is essential for optimizing milling operations and tool lifespan.
• Existing models may not fully capture the complex interactions influencing tool wear.

Purpose of the Study:

• To develop a novel hybrid model for predicting milling tool flank wear.
• To integrate artificial bee colony (ABC) optimization with multivariate adaptive regression splines (MARS) for enhanced prediction accuracy.
• To identify the most influential parameters affecting milling tool wear.

Main Methods:

• A hybrid model combining ABC optimization and MARS regression was developed.
• The ABC algorithm was used to optimize MARS hyperparameters for improved regression accuracy.
• The model predicted milling tool flank wear based on experimental parameters like time, depth of cut, feed, and material type.

Main Results:

• The ABC-MARS model achieved a high determination coefficient (R²) of 0.94, indicating excellent goodness of fit.
• The model accurately predicted tool wear for regular, entry, and exit cuts.
• Key parameters influencing milling tool flank wear were identified.

Conclusions:

• The proposed ABC-MARS model offers a robust and accurate method for predicting milling tool wear.
• The model's ability to identify influential parameters can guide improvements in milling machine design and operation.
• This approach enhances the understanding and management of tool wear in milling processes.