Comparison between artificial neural network and multiple regression for the prediction of superficial roughness in dry turning

Main Article Content

Yoandrys Morales Tamayo http://orcid.org/0000-0001-7456-1490
Yusimit Karina Zamora Hernández http://orcid.org/0000-0002-0112-1061
Paco Jeovanni Vásquez Carrera http://orcid.org/0000-0003-4734-8584
Mario Paúl Porras Vásconez http://orcid.org/0000-0002-4119-4812
Joao Lázaro Bárzaga Quesada http://orcid.org/0000-0001-9792-7379
Ringo John López Bustamante http://orcid.org/0000-0002-6519-1587

Abstract

The simple regression and artificial neural network methods are techniques used in many industrial. This work developed two models in order to predict the surface roughness in dry turning of AISI 316L stainless steel. In its implementation they were considered various cutting parameters such as cutting speed, feed, and machining time. The models obtained by both methods were compared to develop a full factorial design to increase reliability of the recorded values of roughness. The analysis can be checked by the values of coefficients of determination that the proposed models are able to predict surface roughness. The obtained results show that the neural networks techniques is more accurate than the multiple regression techniques in this study.