Evaluating and optimizing surface roughness using genetic algorithm and artificial neural networks during turning of AISI 52100 steel

• Published in: International journal on interactive design and manufacturing 2024-10, Vol.18 (8), p.6151-6160
• Main Authors: Rao, G. Srinivasa, Mukkamala, Usha, Hanumanthappa, Harish, Prasad, C. Durga, Vasudev, Hitesh, Shanmugam, Bharath, KishoreKumar, K. Ch
• Format: Article
• Language: English
• Subjects: Artificial neural networks; Bearing steels; CAE) and Design; Chromium steels; Computer-Aided Engineering (CAD; Cutting tools; Design analysis; Design factors; Design of experiments; Design optimization; Electronics and Microelectronics; Engineering; Engineering Design; Genetic algorithms; Industrial Design; Instrumentation; Mathematical analysis; Mechanical Engineering; Metal cutting; Neural networks; Optimization; Original Paper; Orthogonal arrays; Prediction models; Process parameters; Productivity; Rake angle; Surface roughness; Surface roughness effects; Turning (machining)
• ISSN: 1955-2513; 1955-2505
• DOI: 10.1007/s12008-023-01549-5

The major activity in research of metal cutting is to derive the mathematical models for surface roughness in turning. The present work aims at experimental investigation of hard turning of AISI 52100 bearing steel in order to determine the combined effects of tool geometry (nose radius and rake angle) and process parameters (speed, feed and depth of cut) on the performance characteristic surface roughness. Experiments were conducted using orthogonal array and central composite face centered (CCF) design. A mathematical prediction model for SR has been obtained in terms of factors mentioned. 3D response graphs were drawn to study the interaction effects of surface roughness. Prediction for orthogonal array is done through artificial neural network. Optimization of surface roughness was also carried out for CCF design using genetic algorithm. The efficacy of the CCF design over orthogonal array design is evaluated. Minimum surface roughness of 0.985 µm is obtained when A = 90m/min, B = 0.052mm/rev, C = 0.6mm, D = 4mm and E = − 12 O after application of genetic algorithm.