<p>Proces glodanja tvrdih (kaljenih) čelika, vretenastim loptastim glodalima,<br />najčešće se primenjuje u operacijama završne obrade kompleksnih<br />površina. Modelovanje ovog procesa i optimizacija njegovih parametara su<br />veoma važni, kao pomoć za razumevanje samog procesa tako i za rešavanje<br />praktičnih problema. Za modelovanje izlaznih performansi procesa obrade<br />i nalaženje optimalnih vrednosti ulaznih parametara, korišteno je<br />nekoliko klasičnih i prirodom-inspirisanih metoda. Od klasičnih metoda<br />modelovanja i optimizacije, u radu su primenjene: metodologija odzivne<br />površine-RSM, Taguči metoda i Taguči metoda sa sivom relacionom<br />analizom. Korištene prirodom-inspirisane metode modelovanja i<br />optimizacije su: genetski algoritam&ndash;GA, sivi vuk optimizer&ndash;GWO i<br />nedominantno sortirajući genetski algoritam II&ndash;NSGA II. Dobijeni pouzdani<br />matematički modeli izlaznih performansi procesa obrade i optimalni<br />ulazni parametri obrade, potvrđuju opravdanost primene svih navedenih<br />metoda u procesu glodanja vretenastim loptastim glodalima tvrdih<br />(kaljenih) čelika. Posebno treba izdvojiti rezultate dobijene pomoću metode<br />sivi vuk optimizer&ndash;GWO. Ova prirodom-inspirisana metoda je potpuno nova<br />metoda i do sada nije bilo literaturnih informacija o mogućnostima njene<br />primene u procesima obrade rezanjem.</p> / <p>The ball end milling process of hard (hardened) steel, is usually applied in the<br />finishing operations of machining comlex surfaces. Modeling of this process and<br />optimization of its parameters are very important as an aid to understanding the<br />processes and to solve practical problems. Several classic and nature-inspired<br />methods were used for modeling of the output process performances and finding<br />the optimal values of input parameters. From traditional methods of modeling and<br />optimization Response Surface Methodology-RSM, Taguchi method and Taguchi<br />method with Gray Relational Analysis, and nature-inspired methods of modeling<br />and optimization Genetic Algorithm-GA, Gray Wolf Optimizer-GWO and Nondominant<br />Sorting Genetic Algorithm II- NSGA-II, were applied in the paper. Reliable<br />mathematical models of the output process performances and the optimal input<br />parameters, confirm the validity of the application of these methods in the process<br />of ball end milling hard (hardened) steel. The results obtained by the method of the<br />Gray Wolf Optimizer-GWO should be particulary noted. This nature-inspired<br />method is quite a new method, and so far there was no literature information on<br />the possibilities of its application in the cutting processes.</p>

L’objectif de ces travaux de thèse est l’optimisation par une approche expérimentale d’un modèle de micro-fraise hémisphérique en carbure de tungstène revêtu, de diamètre 0,5 mm, destiné à l’usinage des aciers durs. Les données expérimentales obtenues résultent donc de l’usinage d’un acier 40NiCrMo16 à l’état trempé (54 HRC). Les résultats permettent de mettre en évidence certains phénomènes de coupe, d’usure et de comportement dynamique de l’outil liés au procédé, et leur évolution en fonction du type de fraise considéré et des conditions de coupe choisies. La géométrie de l’outil et son comportement dynamique sont ici principalement commentés. Dans un premier temps, l’étude de différentes nuances de carbure, préparation de surface ainsi que l’optimisation des géométries globale et locale des micro-fraises a permis de proposer un modèle optimisé pour notre partenaire outilleur Magafor. Dans un second temps, une approche numérique du micro-fraisage a été utilisée. Un premier modèle de calcul analytique des efforts de coupe a été testé et il a été mis en évidence les limites d’identification des coefficients spécifiques de coupe dans notre cas. Ensuite, une modélisation numérique par éléments finis du micro-fraisage a été réalisée afin d’appréhender l’étude du comportement dynamique des micro-fraises en fonction de la géométrie globale de l’outil développé. / The aim of this PhD work is optimized by experimental approach with 0.5 mm diameter micro-ball-end mills made from micro-grain tungsten carbide and PVD coated for hardened tool steels machining. The experimental data are obtained on machining of hardened steel (54/55HRC), typically used for the production of plastic injection molds. Results permit to highlight some cutting phenomena of wear and dynamic behavior of the process related tool, and changes depending on the type of milling considered and selected cutting conditions. The geometry of the tool and dynamic behavior are primarily discussed here. As a first step, the study of different carbide grades, surface preparation and optimization of global and local geometries of micro-cutters helped provide an efficient model for our partner Magafor toolmaker. In a second step, the modelling of micro-milling is discussed and an analytical model for cutting forces calculation is introduced. It was demonstrated the identifying limits of the specific cutting coefficients in our case. Then, a test of finite element modelling of micro-milling is made in order to estimate the potential of such a method for the study of the dynamic behaviour of micro-mills.

The diploma thesis is focused on machining using the ball-end shank mill. Content of the preliminary part of the work is a shank mill characteristic and a consequent part shows a splitting of ball-end milling cutters, its application, the cutting tool geometry and a characteristic signs of machining. The cutting force model of the ball-end mill is presented as well. A part of proposed model is the conversion of the resultant force too. Practical part is aimed at cutting force analysis of the ball-end mill and the main purpose of this part is a quantification of the cutting force for different work piece tilt angles while milling is performed.