The wear of hot working tools. Application to forging and rolling of steel E.Felder
In hot working, two main types of tool damage occur: thermal fatigue and abrasion induced by the sliding velocity of the metal on the tool. First, by a mechanical and thermal analysis, we quantify the influence of friction, geometry of the operation and presence of scale on the workpiece, on these two types of damage. We then build a model of abrasion, which is then applied to rolling and forging; we obtain a good agreement with experiments performed on flat forging dies. Finally, in the case of forging, we model the evolution of die surface
hardness, and give some experimental results about the influence of a nitriding treatment.
The design of hot working processes requires models providing an estimation of tool life, which depends on various types of damage; the main ones (1) are abrasion, mechanical and thermal fatigue, and plastic deformation. This paper is devoted to the description of a model of abrasion and a discussion of some factors governing the thermal fatigue in the hot working of steels. The prevailing damage type depends directly on the thermomechanical conditions at the tool-metal interface, mainly on the sliding velocity Au of the metal on the tool. Two cases m a y arise:
- Au=O In this case, only contact thermal fatigue occurs: due to the periodic contact between the tool (temperature 40°C-200OC) and the steel workpiece (temperature llOO°C1200°C), the tool surface undergoes cyclic heating and cooling; the thermal stresses thus generated produce a network of cracks, becoming deeper and deeper (2). - Au#O Thermal fatigue is still present, but moreover, oxyde particles carried into the contact by the workpiece induce abrasion and scratch the tool surface in direct proportion to Au and the contact pressure p.
50% ( R = 2 50 H,= 20 mm)
Fig 1 :damage m a p of flat rolling cylinders(4)
2 MECHANICAL ANALYSIS OF WORKING PROCESSES AND DIE DAMAGE M A P S The mechanical analysis of the metal flow by the theory of plasticity (3) provides the values of Au and p along the interface, which mainly depend on the friction stress T, the geometry of the process and the tensile yield stress of the metal 00. Let us consider two simple, but instructive examples which can be analysed by elementary, classical methods. 2.1 Flat rolling Figure 1 represents a wear map for flat rolling. The process is characterized by two parameters: the "TRESCA" friction coefficient m= 743 100 < 1 (3) and the aspect ratio of the roll bite A. If A is below a critical value, which increases with m, friction prevents any sliding and wear; above the critical value, sliding, hence abrasion, occurs. For given values of the roll radius R and the initial strip thickness Hi, the aspect ratio A increases with the reduction r=l-H2/H1 and we can defiie in the map of figure 1 five zones: - too low or too high reduction for industrial application - roll slippage, where friction is too low to carry the strip into the roll bite - at low reduction, the field of thermal fatigue - at high reduction, the zone of abrasion, the intensity of which depends on friction.
The interface conditions during hot forging depend on lubrication and temperature, and a dissymmetry between upper and lower die frequently occurs in industrial workshops. The sliding velocity on one of the dies is influenced by frictional conditions on both dies: for a given aspect ratio of the cylinder, figure 2 shows that the sliding speed on the upper die increases when friction ms on the upper die decreases, or when the friction on the lower die mi increases. Note that under very dissymetricalconditions (tTss<
can be greater than that obtained under symmetrical frictionless conditions (&=&4); this is confirmed by observations on plasticine workpieces. 3 THERh4OMECHANICAL BEHAVIOUR OF SCALE The damage type and friction are mainly influenced by the effective hardness of the workpiece scale which depends on its temperature and its chemical composition. This is a very complex matter, since the high temperature scale on steel is composed of three iron oxydes FeO, Fe304 and Fe2O3. However wustite FeO is the main component; as its hardness is lower than that of steel above 90O0C,scale can have a lubricating action if it cools down only slowly during each contact time At, which can vary between 5 ms (forging on drop hammer or final finishing rolling passes) and a few seconds (forging on hydraulic press or first roughing rolling passes). Figure 3, deduced from ring compression tests performed under dry conditions, demonstrates that the behaviour of the scale can be characterized by the ratio S of the scale thickness hc and the depth of heat diffusion into scale (3): S=hc ( 6 ~ A t -Oe5 ) (1) where is the thermal diffusivity of scale (=0.4mm2/s>.
Fig. 2 : influence of the lower and upper die friction coefficient on the sliding velocity on the upper die for upsetting of cylinders ( 5 ) (diameter to height ratio: A=2R/h=0.75)
Fig. 3 : influence of the contact pressure p and thermal conditions on dry friction and behaviour of scale (6)
255 Three zones can be distinguished on figure 3: - lubricating zone: S>2 the scale temperature remains high enough to induce a low TRESCA type friction, which is not very sensitive to p and At. As the real contact area
17; 0.95 E,= 0.35 fnm
approaches the apparent one, the scale can stick to the tools, but induces very low abrasion. - wear zone: S<0.05 the scale is strongly cooled by the contact with the tool, and induces a higher, COULOMBtype friction: the friction stress is proportional to p and not very sensitive to At; the scale is harder than the workpiece, brittle, not very adherent, but induces high wear. - intermediate zone: O.O5
HAMMER m ;=0.8, m,= 0.1
A / \
4 MECHANICAL MODEL OF ABRASIVE WEAR
Tests on mbometers most of the time follow ARCHARD'S law: wear volume SV is proportional to the normal load P and the sliding distance 61 (during time St): 6V= k P 61 (2) Let us consider an apparent contact area A. Noting that contact pressure p=PIA, wear depth Sh=6VlA, and sliding length Sl=Au 6t, we can easily deduce from eq. (2) an estimate of the wear depth at a point M on the tool surface: 6h = k j p Au dt (3) We state precisely the value of wear coefficient k in paragraph 5 below.
4.1 Application to the metal working processeS The mechanical analysis of the process provides Au and p (cf. paragraph 2). We first consider the rolling of a strip of length L in the high wear region (fik0.66; A>3, cf. figure 1). For this stationary process, the rolls undergo uniform wear and integration of eq.(3) along the arc of contact provides the radial loss (4):
The forging process is not stationary, and the analysis is more complex. As p and Au vary with time and along the interface, 6h depends on the number of forgings and varies along the interface. Figure 4 gives values of sliding length S1= Au dt ,the contact pressure p, die surface temperam 0s and wear depth profile 6h for the upsetting performed under dissymetrical conditions on a drop hammer and a mechanical press.
Fig. 4 : results of thermomechanical analysis of the upsetting of cylinders under dissymemcal conditions, and wear depth profile (7) steel cylinder 2% = 30 mm ;ho = 40 mm ;height reduction 75%; N=1000 forgings; Flat dies Z38CDV5 ;HV=4GPa upper die (lubricated) lower die (dry) 4.2 The wear rate coefficient k
In order to describe the results obtained in forging (7), we assume that the wear rate coefficient k has the form: k = KF K w HV -2.1 (5) First k decreases very markedly (exponent -2.1) as the Vickers hardness HV of the tool material increases, this is probably in close relation to the relatively small difference in hardness between the tool and the workpiece scale under the contact conditions (cf paragraph 3). Moreover, k
256 depends on the microstructural features of the tool material (amount of carbides for instance); this effect is described by the factor KW which can be written:
workpiece radius of the
ICw = 1 + 5.1 exp(-O.O85W) W = (%W) + 2(%Mo) + 4(%V)+ 0.5(%Cr)
_ _ - I-
-W is called the tungsten equivalent. BOTTOM DIE
Finally, the wear rate depends on the superficial films (lubricant film, transfer film formed by scale). The experimental values of the corresponding factor K are given in table 5.
s=0.015 to 0.05
lower die: dry;workpiece
Fig. 6 : theoretical and experimental wear depth profiles in upsetting (conditions: same as figure 4). Z38CDV5 flat dies ( H V 4 GPa) (7) a: mechanical press N = 1 0 forgings b: drop hammer; bottom die
Table 5 : influence of contact time on KF and friction
4.3 Exuerimental test of the model In figure 6 we compare the theoretical and experimental wear depths profiles under various forging conditions. Under drop hammer forging conditions, we assume that the workpieces are randomly shifted horizontally by +3mm
(the positioning system is not perfect).We can note some interestingfeatures from table 5 and figure 6: - the lubricant diminishes drastically the wear rate at all contact times. Therefore, although lubrication increases the sliding distance (cf figure 4), lubricated wear is very low under our experimental conditions (figure 6); in industrial workshops where contact conditions are more severe, lubrication most of the time increases wear. - for very small contact times such as with drop hammer (S=0.2), the scale of the lateral surface of the workpiece, which comes progressively in contact with the dies at the periphery, is not very abrasive (it remains hot). So at the periphery, under dry conditions, the wear depth is smaller on drop hammer than on press. - on the contrary, in the central part of the dry lower die, the scale of the plane surface of the workpiece is cooled during the time lag before forging, and the wear rate is very high.
b) Hamm o r
The values of the wear rate coefficient deduced from a few measurements performed in hot rolling with formula (4) are in good agreement with the values observed in hot forging.
5 EFFECTIVE HARDNESS OF FORGING DIES The thermal solicitations of the forging dies are much more severe than those of the rolls of rolling mills, and can produce marked metallurgical, hence hardness changes in the superficial layers.Hence, the damage rate may greatly vary with the number of forgings performed. Depending on the maximal values of die surface temperature OM (Fig. 7), superficial softening or hardening may occur.
em t I
Fig. 7 : thermal cycle of the die surface
257 5.1 Surface softening Let us fiist assume that the maximal surface temperature exceeds the tempering temperature of die material, but remains smaller than the temperature for start of austenitization AC 1 . The additional tempering can be calculated from the master tempering curve of the tool material (10,ll). As the hardness decreases, wear rate can increase drastically according to formula (5). moreover, in the central part of the dies, where the sliding distance and wear remain small, thermal fatigue can increase drastically as demonstrated in figure 8. Due to the progressive reduction of the die flow stress, the plastic extension occuring after each forging increases steadily.
Fig. 8 : evolution of the superficial thermal and stress cycles with the number of forgings performed on Z38CDV5 dies (9)
5.2 Surface hardening
At the periphery of the drop hammer die, the sliding velocity can reach very high values (a few m/s), and induce very high die surface heating. If the maximal surface temperature OM exceeds AC1, the hardening occurs by redissolution of the carbides in the matrix (all the more complete as OM is higher) during heating followed by martensitic transformation (all the more complete as the bulk temperature Om is lower) during cooling. We have built a model of this metallurgical change; the hardening occurs after a few forgings, increases with an increase in @Mand a decrease in Om (figure 9).
Fig. 10 : influence of nitriding treatment in hot forging with a mechanical press (conditions: same as fig. 6 ) a: transfer film on dry die N=1000 forgings b: maximal depth of cracks vs distance from centre
The high hardness of the "white layers" (6,ll) so produced avoids abrasion, but they are brittle and most of the time appear cracked. It is interesting to notice that hardening of die surface by nimding treatment has similar effects (figure 10): in press forging, the surface hardness of nimded dies remain high enough (HV>8GPa) to prevent any wear, but promotes thermal fatigue (figure lob). Under dry conditions, nitriding promotes formation of highly adherent transfer films which reduce the cracking rate in comparison with the lubricated die: lubricant hinders the adhesion of the scale and the thermal cycles are more severe.
6 CONCLUSION The prediction of tool life in industrial practice clearly reveals that hot working is not easy: changes of some process parameters usually can modify the tool damage pattern. We have seen for instance that high tool surface hardness induced by thermal cycles or surface treatment hinders absasive wear but promotes transfer films under dry conditions and thermal fatigue. Despite their simplicity, the models presented in this paper have been successfully applied to simple geometries. Work is presently in progress to introduce these models into Finite Element models of metalworking processes (12) to account for wear in the design of more complex operations. REEERENCES ( 1 ) THOMAS, A. : "wear of drop forging dies" in TRIBOLOGY IN IRON AND STEEL WORKS. Publ. by THE IRON AND STEEL INST. (LONDON) (1970), 135141
(2) STEVENS,P.G., IVENS,K.P. and HARPER,P. : 'Increasing work roll life by improved roll cooling practice. J. Iron Steel Inst. (Jan. 1971) 1-11 (3) BAQUE, P., FELDER, E., HYAFIL, J. and D'ESCATHA, Y.: MISE EN FORME DES METAUX. CALCULS EN PLASTICITE. Publ. DUNOD (PARIS) (1974) (4) FELDER, E. : "Interactions mCtal-cylindre en laminage" CESSID-IRSID lectures (MAIZIERES-LESMETZ) (1985), 85-171 (5) THORE,Y. and FELDER, E. :'htheoretical and experimental study of the interface conditions during the hot forging of steel using a dissymetrical flow model."J. Mech. Working Tech. 13 (1986) 51-64 ,I
(6) FXLDER, E. and BAUDUIN, P. :"usure et frottement ,I en forgeage B chaud. Mtcanique, Matkriaux, Electricitt (Jan. 1981) 4-15 (7) THORE,Y. ?Etude thtorique et exPCrimentale du frottement et de l'usure par abrasion des rnatric.7; de forge B chaud des aciers. Influence d'une nitruration. Thbse Dr Ing. (Ecole des Mines de Paris, CEMEF, 1984) (8) FELDER, E., RENAUDIN, J.F. and THORE,Y. : "The tribology of a simple hot forging process: influence of lubrication and nimding treatment.'boc. Conf. Advanced Techn. Plasticity (1984) Vol 1,231-234 (9) BAUDUIN, P. , FELDER, E. , BATIT, G. and MONTAGUT, J.L. :"Study of the thermomechanical It conditions and the wear of hot forging dies. Proc. 10th Int. Conf. Stamping (LONDON, JUNE 1980) (10) PAYSON, P. : THE METALLURGY OF TOOL STEELS. Publ. J. WILEY and Sons (LONDON) (1962) (11) FELDER, E. and COUTU, L. : "Aspects thermodynamiques de l'usure des mamces & forge B chaud de l'acier.'Bull. Cercle Etude des MCtaux 4, 14 (Dec. 1978) 153-176 (12) CHENOT, J.L. and ONATE, E. : MODELLING OF METAL FORMING PROCESSES. Publ. KLUWER (DORDRECHT) 1988