Estimation of Surface Roughness Value Using Back Propagation Neural Network on Green Machining


M. Yanis†*, Arie Y.B, Aneka F, Nova Y


  Department of Mechanical Engineering-Faculty of Engineering-Sriwijaya University, 60662, Indonesia

Department of Chemistry-Faculty of Mathematics and Natural Science-Sriwijaya University, 60662, Indonesia

Corresponding Author Email: [email protected]

This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

This study discusses the prediction of surface roughness during the side milling of AISI 1045 material. Surface roughness is an indicator of surface quality which is one of the most frequently specified customer requirements in the machining process. Experimental and optimization studies were developed using the Artificial Neural Networks (ANN) method. The Levenberg-Marquardt algorithm on back propagation is used for training and testing to predict the effect of input variables on surface roughness. The machining variable input consists of cutting speed, feeding rate, radial and axial depth of cut. The machining condition uses coconut oil as the cutting fluid with a Minimum Quantity Lubrication system. The network developed is a structure with one and two hidden layers. The number of neurons 1 to 20 is assigned to this hidden layer. The network structure that offers the surface roughness value closest to the experimental value was 4-10-1 with MSE training 0.00467 (R2 = 98.23%) and testing 0.0052 (R2 = 99.47%). This shows that a network structure with more hidden layers does not increase network functionality. The average error percentage between experimental results and ANN predictions is 1.43% (training) and 6.87% (testing). The effect of machining variables on surface roughness shows that the surface roughness value decreases (smooth) with increasing cutting speed. And increasing the feed rate and depth of cut will increase the surface roughness value.