Akademik Veri Yönetim Sistemi
JOURNAL OF POLYTECHNIC-POLITEKNIK DERGISI, 2024 (ESCI)
CNC turning is commonly used to machine metal parts. The roughness values of the final surface are fundamental data for evaluating the output quality and determining the machining conditions. Investigating and predicting complex relationships between machining parameters and surface roughness is essential for improving the cutting process and output quality. Artificial neural network (ANN) models can be used to correlate cutting conditions with surface roughness (Ra) values due to their ability to learn and predict complex relationships. In this study, mathematical prediction models were developed using Multiple Linear Regression (MLR) and ANN algorithms to predict surface roughness values after turning. Turning experiments were conducted on DIN 1.2344 hot work tool steel material under dry -cutting conditions to generate training and test data sets for the algorithms. The control factors and levels were determined as cutting speed (300, 350, and 400 m/min), depth of cut (0.25, 0.5, and 0.75 mm), and feed rate (0.07, 0.1, and 0.13 mm/rev). Surface roughness values of the machined specimens were measured with a profilometer. The prediction models were validated with experimental measurements, and their performance was evaluated. It was calculated that ANN predictions had an accuracy of 87.6% compared to the actual values, while MLR predictions had an accuracy of 78.4%. The results showed that the ANN method performs more than the MLR method and can be used to predict surface roughness values.