EXPLAINABLE MACHINE LEARNING FOR SURFACE ROUGHNESS PREDICTION AND CUTTING PARAMETER OPTIMIZATION

Abstract

Surface roughness is a critical quality indicator in precision machining because it directly influences product performance, dimensional accuracy, wear resistance, fatigue strength, and manufacturing reliability. Conventional machine learning models can accurately predict surface roughness but often operate as black-box systems, limiting their adoption in industrial decision-making where transparent reasoning is essential. This paper proposes an Explainable Machine Learning (XML) framework integrating intelligent sensors, predictive analytics, Industrial Internet of Things (IIoT), explainable artificial intelligence, cloud manufacturing, and adaptive machining optimization for surface roughness prediction and cutting parameter optimization. The proposed methodology combines predictive learning with explainability techniques to identify the influence of machining parameters on surface quality while recommending optimal cutting conditions. Experimental evaluation demonstrates improvements in prediction accuracy, machining quality, interpretability, tool utilization, production efficiency, and manufacturing sustainability. The proposed framework provides an intelligent, transparent, and Industry 4.0-ready solution for explainable machining optimization in advanced manufacturing environments.