COMPARISON OF CLASSICAL AND MODERN CALCULATION OF MACHINED SURFACE QUALITY

Ovim završnim radom obuhvaćeni su osnovni parametri površinske hrapavosti, parametri obrade tokarenjem te njihova povezanost s površinskom hrapavosti strojnih dijelova. Opisana je klasična metoda izračuna hrapavosti, te je dan prikaz rezultata jednog takvog izračuna. Zatim su navedene neke od suvremenih metoda izračuna hrapavosti, te je u nastavku objašnjen način rada neuronskih mreža. Prikazan je model kreirane neuronske mreže za izračun hrapavosti, zajedno s računalnim kodom potrebnim za kreiranje iste, te su na kraju dane usporedba klasičnog izračuna sa suvremenom metodom procjene umjetnim neuronskim mrežama i zaključak o provedenim analizama.This undergraduate thesis includes description of basic parameters of surface roughness, turning cutting parameters and their connection to the value of surface roughness of machine parts. It includes a description of classical methods for calculating surface rougness and the calculation results of an example for said method are displayed. Then some of the modern calculation methods are listed and operation principles for neural networks are explained. A model of created artificial neural network for surface roughness prediction is shown, along with the needed computer code and in the end the comparison of classical and modern calculation method are given, as well as a conclusion about conducted analyses