Surface Roughness Prediction in Ultra-Precision Milling: An Extreme Learning Machine Method with Data Fusion.

Area of Science:

• Manufacturing Engineering
• Machine Learning Applications
• Surface Metrology

Background:

• Accurate surface roughness prediction is crucial in ultra-precision manufacturing.
• Traditional methods often struggle with complex machining dynamics and limited data.
• Extreme Learning Machines (ELM) offer potential due to their rapid learning and fitting capabilities.

Purpose of the Study:

• To pioneer the application of Extreme Learning Machines (ELM) for surface roughness prediction in ultra-precision milling.
• To enhance ELM prediction accuracy by fusing machining parameters and force signal data at the feature level.
• To validate the proposed data-fusion-based ELM method through experimental verification.

Main Methods:

• Utilizing the Extreme Learning Machine (ELM) algorithm for predictive modeling.
• Implementing feature-level data fusion by combining machining parameters and force signal data.
• Conducting ultra-precision milling experiments to generate comprehensive datasets.
• Evaluating prediction performance using metrics such as mean absolute percentage error (MAPE).

Main Results:

• The proposed data-fusion-based ELM method significantly improves surface roughness prediction accuracy.
• ELM with data fusion achieved a low mean absolute percentage error of 1.6%.
• The model training time was remarkably short, requiring only 18 seconds.
• The ELM approach demonstrated superior performance compared to other state-of-the-art methods.

Conclusions:

• Extreme Learning Machines, enhanced by feature-level data fusion, are highly effective for surface roughness prediction in ultra-precision milling.
• This approach offers a computationally efficient and accurate solution for real-time manufacturing quality control.
• The findings pave the way for integrating advanced machine learning techniques in precision engineering.