Prediksi kekasaran permukaan baja S45C terhadap parameter pemesinan dan getaran pada proses bubut menggunakan metode artificial neural network

Abstract

Based on machining characteristics, this study gives surface roughness modeling for machine parts. The artificial models used the Artificial Neural Network (ANN) modeling approach and multivariable regression analysis were used to create the prediction model. S45C steel was one of the materials utilized in this research. With a depth of cut 0.5 mm, the parameters are spindle (n) of 165, 330, 585, and 1170 rpm and feed (f) of 0.2 mm/rev. Utilizing TIBCO software, surface roughness values will be predicted. Equations derived from multivariable linear regression serve as the study's findings. At 1170 rpm spindle rotation and 0.5 mm of cut depth, the lowest surface roughness measurement of 1.114 (μm) was recorded. At spindle speed 585 and a cut depth of 2.0 mm, a roughness value of 2.999 (μm) was recorded as the maximum value. Roughness rises at spindle speeds between 585 and 900 rpm when cutting at shallower depths. The third modeling had the smallest error value, which was 11.21%, and surface roughness value using an artificial neural network with five simple multi-layer models.