MULTI-OBJECTIVE OPTIMIZATION OF MACHINING PARAMETERS FOR SURFACE ROUGHNESS AND TOOL WEAR REDUCTION

Abstract

The optimization of machining parameters is a critical requirement in modern manufacturing because productivity, surface integrity, dimensional accuracy, tool life, energy efficiency, and production cost are strongly influenced by the selection of cutting conditions. Conventional machining optimization frequently focuses on a single performance objective, even though practical manufacturing systems require simultaneous improvement of multiple and often conflicting quality indicators. Increasing cutting speed may improve productivity but accelerate tool wear, while reducing feed rate may improve surface finish but increase machining time and production cost. This paper proposes a multiobjective optimization framework for machining parameters with the simultaneous objectives of reducing surface roughness and tool wear while maintaining acceptable productivity and process stability. The proposed methodology integrates structured experimental design, multi-sensor machining data acquisition, preprocessing, machining feature construction, data-driven response modeling, normalized multi-objective performance assessment, Pareto-oriented candidate evaluation, machine learning-assisted prediction, adaptive parameter recommendation, and experimental confirmation. Cutting speed, feed rate, depth of cut, machining duration, tool condition, vibration behavior, cutting temperature, acoustic response, and powerrelated indicators are considered within a unified optimization environment. Historical machining experiments are combined with real-time process observations to model nonlinear relationships among controllable parameters and quality outcomes. Candidate parameter combinations are evaluated according to their simultaneous influence on surface roughness and tool degradation rather than through isolated single-response optimization. The framework further incorporates constraint screening to eliminate parameter combinations associated with excessive thermal loading, unstable vibration, poor material removal behavior, or unsafe operating conditions. Representative experimental analysis demonstrates that the proposed multi-objective approach reduces average surface roughness and tool wear compared with baseline machining settings and single-objective optimization strategies. The results indicate improved compromise solutions, stronger parameter selection consistency, reduced unnecessary tool replacement, and enhanced manufacturing quality. The framework also supports integration with Industrial Internet of Things monitoring, Digital Twin platforms, production machine learning services, and governed enterprise APIs. The study demonstrates that data-driven multi-objective optimization can provide a practical foundation for intelligent, adaptive, and quality-aware machining operations.