Naslov
Prediction of surface roughness in plasma jet cutting process using neural network model

Autori
Peko, Ivan ; Nedić, Bogdan ; Đorđević, Aleksandar ; Đurić, Stefan ; Džunić, Dragan ; Veža, Ivica ; Janković, Marko

Vrsta, podvrsta i kategorija rada
Radovi u zbornicima skupova, cjeloviti rad (in extenso), znanstveni

Izvornik
Proceedings of 15th International Conference on Tribology, SERBIATRIB ‘17 / Mitrović, Slobodan - Kragujevac : University of Kragujevac, Faculty of Engineering, 2017, 520-525

ISBN
9788663350410

Skup
15th International Conference on Tribology, SERBIATRIB '17

Mjesto i datum
Kragujevac, Srbija, 17.05.2017. - 19.05.2017

Vrsta sudjelovanja
Predavanje

Vrsta recenzije
Međunarodna recenzija

Ključne riječi
plasma jet cutting, experiments, surface roughness, neural networks, modeling

Sažetak
Plasma Jet Cutting process is one of the most important nonconventional machining methods that uses a thermal energy of highly ionized gas to cut specified material. Its optimal combination of cut quality, productivity and costs in cutting various types of materials increase its demand in the market and usually position it in front of the other cutting processes. This process is controlled by technological parameters recommendations given by the manufacturers of the cutting equipment. These parameters recommendations usually do not lead to optimal cut quality but refer to manufacturers’ business targets. Accordingly, various experimental researches have been done trying to describe process and define technological parameters levels that lead to optimal cut quality. In this paper, mathematical model for prediction of surface roughness in plasma jet cutting process was developed. Process parameters whose influence was analyzed are cutting speed and arc current. Experiments were conducted according to Taguchi L9 orthogonal array design on aluminium sheet thickness of 3 mm. Mathematical model was created using a single hidden layer artificial neural network trained with the Levenberg- Marquardt algorithm. Prediction accuracy of developed model was verified using statistical measures such as Mean Squared Error (MSE) and correlation coefficient (R) between experimental and predicted values. To analyze the main and interaction effects of the plasma cutting parameters on the surface roughness 2D and 3D diagrams were generated from which optimal cutting regions were identified.

Izvorni jezik
Engleski

Znanstvena područja
Strojarstvo

POVEZANOST RADA