OPPORTUNITIES AND REQUIREMENTS FOR THE NU MERICAL-SIMULATION OF COLD FORMING OPERATIONS - SELECTION OF THE MATERIAL AND THE FABRICATION CONDITIONS LEADING TO THE DESIRED LEVEL OF PROPERTIES

One of the main challenges facing the forgemaster is to predict the be haviour of steel during its deformation process, and after that, under its service conditions. Indeed, the design of optimized forging sched ules, by means of the classical trial + error process, is becoming inc reasingly hard to bear in a competitive environment, thus requiring ne w predictive methods such as numerical simulations. During the last fe w years, Finite Element software packages have made so much progress t hat, at least for two dimensional problems, they are no longer the lim iting factor to reliable computer simulations. They have pointed out t he limitations of numerous physical models, therefore requiring new ex perimental developments. At Irsid, we have identified three physical p henomena as critical for a reliable computation. First of all, the con stitutive law of steel under cold forging extreme conditions (0< epsil on< 6, 0< ($) over bar epsilon< 1000 s(-1), 20< T< 500 degrees C) is r equired to get realistic loads or local stresses. Then, friction condi tions have to be specified, in order to obtain a good assesment of the material flow between the dies. Finally, it is very important to unde rstand and describe properly the evolution of ductile damage leading t o fracture.In the determination of constitutive laws under cold formin g conditions, in view of numerical simulations, two problems have to b e solved. First of all, it is required to select the most general anal ytical form, able to describe the behaviour of all the considered stee l grades with a finite number of parameters. For this purpose, we sele cted a form proposed by F.J. Zerilli and R. W. Armstrong in 1987 and b ased on dislocations theory. Then, it is necessary to choose a suitabl e test, keeping in mind economical considerations. For this purpose, w e used the standard upset tests along with an original methodology to derive the strain hardening curve from the experimental recordings. Wi th this kind of experimental setting, it is not possible to study the effect of very high strain rates; therefore, we chose to use dynamic t orsion tests of tubes, realized with Hopkinson bars, in order to reach strain rates up to 1700 s(-1). The main difficulty arising with these experiments was the exploitation of crude data because of the strain localization phenomenon. To overcome these limitations, we used an inv erse method based on successive converging finite elements computation s. The friction problem in treated using two very different methodolog ies based upon the forming process considered. For forging operations, we are adapting a << cone upsetting >> test, initially proposed by Ko pp for hot forging applications. This new test will allow us a more co mprehensive treatment of the friction phenomenon, compared to the clas sical ring test, thanks to the possibility of varying the cone angle a nd therefore the normal pressure. For the wire drawing process, a mode l based on drawing load measurements is developed. This model relies o n an analytical analysis of such operations, using an improved slices method. In order to predict ductile fracture, we first implemented the most classical criteria into a finite element package, to assess thei r precision over a wide range of loading conditions. The main advantag e of this method lies in the possibility of overcoming the very restri cting assumptions required by analytical integrations. Our test showed that none of the selected formulations was general enough to solve al one cold forming problems. We are then using the most recent approache s based on porous plasticity theory. Special experimental procedures a re developed to vary the main factors governing damage evolution witho ut changing othe