Proceedings on Engineering Sciences (Jun 2019)

PREDICTING CUTTING PARAMETERS BY APPLYING DEVELOPED NEURAL NETWORK AND LINEAR REGRESSION MODELS

  • Aleksandar ĐORĐEVIĆ,
  • Milan ERIĆ,
  • Miladin STEFANOVIĆ,
  • Slobodan MITROVIĆ,
  • Marko PANTIĆ,
  • Aleksandra KOKIĆ ARSIĆ,
  • Dragan DŽUNIĆ

DOI
https://doi.org/10.24874/PES01.01.064
Journal volume & issue
Vol. 1, no. 1
pp. 482 – 489

Abstract

Read online

The paper presents the methods for prediction of the cutting parameters. In order to test the workability of the material by process of scraping, from the aspect of the cutting temperature, an natural thermopar was placed just below the cutting edge of the plate. In this way, a simple, reliable, accurate and economical method for determining the workability of material by cutting is obtained. The feasibility study of several semi-finished products by applying the realized experiments was carried out. Different materials in processing with cutting discs with different coatings give different results, which are used to form neural network and linear regression models.

Keywords