Author(s):
1. Miloš Madić, Serbia
2. Miroslav Radovanović, Faculty of Mechanical Engineering in Niš, Serbia
3. Dušan Petković, Faculty of Mechanical Engineering in Niš, Serbia
4. Predrag Janković, Faculty of Mechanical Engineering in Niš, Serbia
5. Miloš Milošević, Faculty of Mechanical Engineering in Niš, Serbia

Abstract:
In this paper, linear and quadratic regression models and artificial neural network model were developed to predict surface roughness for different values of cutting speed, laser power and assist gas pressure in CO2 laser cutting of mild steel. For the purpose of laser cutting experimentation Taguchi’s L25 orthogonal array was used arranging three factors at five levels. Surface roughness predicted values by both models were compared with the experimental values. The artificial neural network model was found to be capable of better predictions.

Key words:
CO2 laser cutting,regression analysis,artificial neural networks,modeling

Thematic field:
Production and Computer-Aided Technologies

Date of abstract submission:
06.05.2015.

Conference:
12th International conference on accomplishments in Electrical and Mechanical Engineering and Information Technology (DEMI 2015)

Final paper