OPTIMIZATION OF SURFACE ROUGHNESS INTURNING USING STATISTICAL ANALYSIS AND MACHINE LEARNING

Surface roughness optimization plays a crucial role in ensuring the quality and cost-efficiency of machined products. This work uses machine learning and Response Surface Methodology (RSM) to predict and optimize surface roughness during mild steel CNC turning. Crucial process variables like cutting speed, feed rate, and depth of cut were investigated in order to develop accurate prediction models. Surface roughness (Ra, Rq, Rz) and cutting forces were recorded using a Mitutoyo SJ-210 and a Kistler 9257B dynamometer, respectively. Experiments were performed under varied conditions, and both statistical analysis and machine learning techniques were employed. Optimal machining parameters were identified as 300 m/min cutting speed, 0.15 mm/rev feed rate, and 0.5 mm depth of cut. The RSM and artificial neural network (ANN) models showed high prediction accuracy, with ANN achieving an R-value of 0.9943. The findings confirm the effectiveness of combining statistical and machine learning methods for surface quality improvement in CNC turning.