Journal of Soft Computing and Artificial Intelligence

Yazarlar: Şehmus BADAY, Onur ERSÖZ

Konular:Bilgisayar Bilimleri, Bilgi Sistemleri

Anahtar Kelimeler:ANN,Cutting force,Cutting tool insert,Cryogenic heat treatment

Özet: Cutting force is one of most important criteria for evaluating machinability of workpieces. For this purpose, in present study, prediction of cutting forces obtained by turning AISI 1050 steel with cryo-treated and untreated CVD-coated cutting tool inserts with artificial neural networks (ANN) was investigated. Machining parameters such as feed rate, cutting speed and conditions of cutting tool insert were selected. These parameters were used for input parameters while cutting force was used for output parameter. The employed ANN structure was chosen according to network type, training function, adaption learning function and performance function as feed-forward back propagation, TRAINLM, LEARNGD and MSE, respectively. Thus, the estimation values of cutting forces attained from ANN model during training and experimental values coincide perfectly with the regression lines, which make the R2 = 0.99874 in training. For this reason, cutting force was explained by ANN with an acceptable accuracy in this study.

---