|
|
|
|
Chapter 4 DIRECT AND BACK EFFECTS 4.4 Effect of interaction conditions.
4.3 Mechano-rolling fatigue Another experiment was designed from the standpoint of tribo-fatigue. Its purpose was to investigate the direct and back effects in the metal-to-metal active system during mechano-rolling fatigue, this time friction was created once again in the zone of tension of a bending test specimen (cf. Fig. 3.2, à). Fig. 4.6 illustrates the results of tests of the carbon steel 45 (the specimen) / alloyed steel 25ÕÃÒ (the roller). The diagram ÀÂÑD is plotted in the following coordinates: pressure ð0 in the center of the contact site (the abscissa axis) – the amplitude sà of cyclic stresses in bending (the ordinate axis).
Fig.
4.6 – Diagram of limiting states of active
system in mechano-rolling fatigue The point À is the fatigue limit s–1 of steel 45 specimens determined by common tests for mechanical fatigue following the scheme in Fig. 3.2, c. The limiting state criterion is when the specimen breaks into two pieces because of the main fatigue crack in its vulnerable cross section. Hence, this point implies the mechanics of fatigue fracture. In general, the ordinate axis sà is the strength scale: this scale should accommodate results of fatigue tests of any components of structures made from any materials. The point D is critical pressure pf in rolling friction without slip, it was determined by common tests for friction (following the scheme in Fig. 3, b). The limiting state criterion is the appearance of pits of spalling of critical density along the rolling path. Hence, this point implies tribology. In general, the abscissa axis ð0 is the tribological scale: this scale should accommodate test results of any friction pairs the components of which are made from any materials. Curves ÀÂÑD are a diagram of limiting states of the active system during mechano-rolling fatigue, it as plotted from the results of wear-fatigue tests (following the scheme in Fig. 3, à). Hence, it implies tribo-fatigue. The limiting state along the portion ÀÂ is predominantly due to the development of the main fatigue crack, meanwhile the processes of appearance of pits of spalling are attendant. Therefore, the direct effect occurs satisfactorily described by the equation
where mð
– the parameter of rolling
hardening; it is mð =
0.92 in the conditions of the experiment. On the opposite, the limiting state along the portion ÑD is governed by the critical concentration of pits of spalling, meanwhile the development of mechanical fatigue microcracks is an attendant damage. It is the back effect satisfactorily described by the equation
where ms
– the parameter of cyclic hardening; it is ms
= 0.65 in the conditions of the experiment. The portion ÂÑ is transient; the kinetic processes of interactions between the phenomena of friction (with wear) and mechanical fatigue evolve at larger parameters of loading sà and ð0 close (or equal) to critical (s–1, pf). The limiting state under these conditions of tests can be reached concurrently based on two criteria. Examination of the ÀÂÑD diagram leads to the following basic conclusions. (1) The fatigue limit of the specimen increases 1.5...1.6 times if the process of rolling friction occurs concurrently (the direct effect — the portion ÀÂ). The direct effect factor advanced in tribo-fatigue (3.2)
is, in fact, a strength characteristic; its maximum value in the test conditions is KD max = 268/165 = 1.62. Factor (4.11) is incorporated, naturally, into equation (4.9). (2) The critical (ultimate) pressure in rolling friction increases 1.2...1.25 times if cyclic stresses are concurrently excited in the specimen (the back effect – the portion ÂÑ). The back effect factor advanced in tribo-fatigue (3.3)
is, in fact, a tribological characteristic too; its maximum value in the test conditions is KB max = 2200/1760 = 1.25. Factor (4.12) is incorporated, naturally, into equation (4.10). (3) The process of wear in rolling within the optimum range of contact pressures (ð0 » 400...1300 MPa) significantly increases the reliability of the system based on the criterion of fatigue resistance so that a tendency to wearless friction is unjustifiable in this case. (4) Tensile stresses during cyclic loading in the optimum conditions (sà » 50...100 MPa) are favorable because they lead to a significant rise of the reliability of the system based on the criterion of resistance to rolling friction. Improvement of the limiting state characteristics s–1ð and ðfs in the process of wear-fatigue tests versus the characteristics during rolling friction (ðf) and mechanical fatigue (s–1) can be explained from the viewpoint of mechanics by the following reasons: · addition of stresses with opposite signs (contact and bending) causing the shift of the mean stress cycle towards negative values and thus to the reduction of the maximum stress cycle; · hardening of the working portion of the specimen by surface plastic deformation; · appearance of favorable residual compressive stresses; · healing of the primary fatigue cracks by elastoplastic deformation in the process of rolling friction. The controlling parameter of wear-fatigue damage (cf. Fig. 4.6)
has the critical value
This critical value separates the regions of direct and back effects on the diagram of limiting states of the active system. If ysp < y–1f, we obtain the curve ÑD. If ysp > y–1f, we obtain the curve ÀÂ. The value ysp = ¥ (pure mechanical fatigue) corresponds to the point À, the value ysp = 0 (pure rolling friction) corresponds to the point D. Application
of fine experimental methods of research enables to study and get insight into
the specifics of complex wear-fatigue damage. Fig. 4.7 exemplifies the results
of studies (with the method of atom force microscopy) of the processes of
cracking on steel 45 specimens during rolling friction and wear-fatigue tests as
a function of the level of contact pressure ð0
and the value of the amplitude of cyclic stresses sà. Figure
(their dimension is ~35´35
mm2)
shows the morphology of cracks typical for the corresponding conditions of tests.
The histogram shows the relation between the critical depth h
of the damaged layer and the level of cyclic stresses (at unchanged contact
pressure ð0 = 2130 MPa).
These experimental data lead to the following conclusions.
Fig.
4.7 – Microtopography of surface damage during
rolling friction Any rise of contact pressure during pure rolling friction intensifies plastic deformation, hence it leads to deformation fragmentation of grains, initially to the appearance of discrete pores and cracks which later form chains. The system of the deformed grains, chains, pores and cracks is unidirectional and it is oriented along the rolling direction. This process leads to the formation of relatively large discrete pits of spalling. Delamination and spalling are two main types of wear. The critical damage depth of the layer is estimated at ~0.4...0.5 mm. During wear-fatigue tests similarly deformation fragmentation of grains, appearance of pores and cracks are observed. Yet the pattern of damage changes significantly. As the amplitude of cyclic stresses grows, the process of appearance of the second system of cracks accelerates, now they are transverse in respect of the rolling direction. That is why damage scatters and an almost regular net of intersecting pores and cracks appears, that fringes with finely dispersed particles (fragments of grains) of the material. The higher the cyclic stresses the denser is the net of pores and cracks and the finer and thinner are the separating particles. The critical depth of the damaged layer grows smaller to 0.05 mm. It prevents the appearance of larger and deeper pits of spalling, and they are not observed under these conditions. Surface crushing is the dominating wear process in this case. It is characterized by separation of finely dispersed particles from the working surface that result from multiple microshearing over intersecting planes and generation of a huge number of scattered microscopic pores and cracks and fine crushing of grains. This mechanism of complex surface damage is called the scattered effect of multiple microshearing (SEMMS). The above results enable to identify additional causes why wear-fatigue damage in certain conditions is less menacing than the damage in friction (at a similar contact pressure). 1. Superposition of the fields of contact and bending stresses leads to dissipation of more applied energy in a finer surface layer of the material and localization of the processes of cracking and wear in the layer. Deformation energy is expended faster for finer crushing of grain fragments and their multiple separation than for penetration of damage into the depth of the material. 2 Wear of the surface layer damaged by a net of pores and cracks exposes a new relatively sound surface highly resistant to fracture. The appearance of relatively larger (in response to the loading conditions) pits of spalling is thus delayed in time or even prevented entirely at the bottom of which dangerous micrconcentration of stresses and a dangerous main crack develop. 3 Approximately tenfold rejuvenation of the working surface is required by fragmentation, crushing and separation of metal particles during wear-fatigue tests for the damage to reach the same depth like in rolling friction, providing the contact pressure is similar in both cases. In this way, it has been established experimentally that wear-fatigue damage is a specific and peculiar type of surface damage of the main component of the active system. Its specific feature in these conditions is the surface crushing because of SEMMS over the intersecting planes of sliding. Its peculiarity is that the process does cause damage, but it is useful for it boosts significantly the reliability and durability of the active system. It is evident that in case of an optimal combination of the loading parameters sà and ð0 the active system reaches the state when its bearing capacity is maintained spontaneously (or controlled automatically) for a long time by the wear and removal of a fine damage surface layer in the friction zone. Summarizing it should be mentioned that the active system is a peculiar dynamic system, its behavior can and should be controlled, for example, by non-traditional method of the wear intensity control. It should
be remarked that the diagram of limiting states of the active system (cf. Fig.
4.6) differs cardinally from ultimate double-parametric diagrams known in
mechanics (for example, sa – sm,
cf. Fig. 1.17). As a rule, the diagrams of the limiting states of components of
structures and friction pairs are plotted using a single criterion of damage (fracture),
for example, the appearance of main crack of a definite length (for a structural
component) or the appearance of the critical concentration of pits of spalling (for
a friction pair). Meanwhile, the diagram of the limiting states of the active
system shown in Fig. 4.6, is based on three criteria: fatigue fracture over the
portion ÀÂ (direct effect), ultimate wear over the portion CD (back
effect) and the critical state based on both criteria concurrently over the
portion ÂÑ. It means that a single equation cannot describe
analytically the full diagram of limiting states of the active system; there
should be separate equations for the portion ÀÂ and CD. Of
course, these equations may be similar (cf., for example, (4.9) and (4.10)), but
their parameters should be specific (like they are in equations (4.9) and
(4.10)). Another experiment served the purpose of studying the manifestations of the back effect when contact pressure increases in multiple steps within a broad range of variations (Fig. 4.8, steps I, II, ..., XII). In the process of tests the convergence dñ of axes of the pair components of the system comprising the specimen from soft steel / roller from high-strength steel was measured during rolling friction (cf. Fig. 3.12, b) (when sà = 0) and during mechano-rolling fatigue (at sà = 0.8s–1 and sà = 1.0s–1). It is visible that (cf. Fig. 4.8) the process of accumulation of wear-fatigue damage decelerates substantially compared with the process of damage during rolling friction, the range of normal friction based on the contact pressure expands by approximately 14%. We will explain the difference between the process of addition and interaction between damages using these experimental data (see also p. 2.5).
Fig.
4.8 – Variations of dñ
during step-by-step contact pressure rise Assume that during the time t1 damages due to contact (wp) and off-contact (ws) loads accumulate as Fig. 4.9, à shows it: none of these criteria leads to the limiting state (wp << 1.0; ws << 1.0). If damages are added up (wp + ws = Sw), then in case of wear-fatigue tests the limiting state (Sw = 1.0) is reached during the time t2 < t1. Yet, evidently this prediction turns out to be untrue for the experimental data shown in Fig. 4.8. If it is considered that damages due to contact and off-contact loads interact
then the scheme adequately reflecting the experimental data in Fig. 4.8 looks like Fig. 4.9, b shows it. The limiting state during rolling friction is reached within the time t2, while during mechanical fatigue it does not occur even at t1 >> t2. During wear-fatigue tests the durability (t1) turns out larger than during rolling friction (t2). Whence a general conclusion follows: during wear-fatigue damage the deformation energy due to contact (Up) and off-contact (Us) loads do not add, they interact dialectically:
The result of such interactions is determined both by the loading conditions and the direction of the processes of hardening-softening (see p. 2.5). It follows from (4.15) that a particular case of interactions between damages occurs, or their addition, at l (s « p) = 1 (the sign of equality is assumed in condition (4.15)).
Fig. 4.9 – Diagrams explaining summation of (à) and interactions between (b) damages An unexpected
phenomenon was discovered during tests (cf. Fig. 4.8): residual
undulatory damages or immovable irregular plasticity waves along the rolling
path on the soft steel specimen (see the photo in the upper right-hand
corner of Fig. 4.6). Meanwhile the shape of the high-strength steel roller
remains unchanged in the contact zone, i.e. geometrically undistorted. Fig. 4.6
shows that in case of regular loading the ultimate pressure pfs
did not exceed
Fig. 4.10, à shows the scan of the rolling path with several irregular (congealed) waves of plastic surface deformation that appeared under these conditions of tests. Each wave is a combination of two peculiar semipunctiform craters and a lintel with the tip resembling a wavy ridge. Fig. 4.10, à shows the typical dimensions of the craters and lintels that lead to the following conclusions. a)
b)
Fig. 4.10 – Specific type of limiting state: surface undulatory damages (pits of spalling are shaded) (à) and distribution of microhardness along length L of rolling path (b) (L A Sosnovskiy, S A Tyurin, V A Yakovlev) None of the congealed deformation waves repeats: each crater and lintel has its own dimensions different from others. The step between craters is also variable. The relative plastic deformation in the radial direction is 4...8%, while it reaches 50...70% in the axial direction. Hence, the appearance of residual surface undulatory damage is due to the non-stationary process of elastoplastic deformation. The anisotropy of the mechanophysical properties of the material in local zones of the rolling path can be assumed to be responsible for the deformation anisotropy in these zones leading to the formation of discrete pits of spalling as sources of the nonsteady state. The stronger the deformation anisotropy and the larger the pits of spalling, the stronger is the dynamic force excited during local collisions of the roller with the specimen. Thus, the form of the translimiting state described for this case is due to the nonsteady impact fatigue processes. The method of microhardness was applied
to corroborate the conclusion about the anisotropic properties of the friction
surface. Fig. 4.10, b shows the distribution of the microhardness of the
material over the circumference of the specimen passing through the centers of
the craters (see line L in Fig. 4.10, a). It is apparent, that, on
the one hand, microhardness changes periodically according to the step of the
craters. On the other hand, the pattern of distribution of the microhardness is
substantially irregular reflecting the random nature of the anisotropy of
properties of local zones of the material along the path of rolling. Hardness is,
as a rule, much lower over the lintels than in the bottoms of craters. So, a significant deformation anisotropy
of the properties of the material in local zones of the path of rolling appears
and develops in the process of wear-fatigue damage. It becomes manifest in three
typical directions: the circumference, depth (the radius of the specimen) and in
the axial direction. It seems that it dictates introduction
of special characteristics of the local wear-fatigue damage process: the coefficient
of asymmetry
where rmin
and rmax – the minimum and maximum radii of one
diameter of the specimen, and the coefficient of irregularity
where rsma
and rlar – the smallest and largest radii of the
specimen during one revolution. Fig. 4.11 (cf. also Fig. 3.13) shows the
conventional designations of the radii of the specimen, also the relations
between the coefficients Ra and ha
and the level of cyclic stresses during the tests
of the active system steel 45/steel 25XTT for mechano-rolling fatigue by
changing the bending
loads in steps under the contact pressure p0 = 0.7pf
= const. It is apparent that the degree of irregularity (or anisotropy) of local
wear-fatigue damage grows according to the augmentation of cyclic stresses. Note
that the smaller are the values Ra and ha
the larger is the anisotropy of wear-fatigue damage.
Fig. 4.11 – Dependence of asymmetry coefficients and irregularity of wear-fatigue damage during tests for mechano-rolling fatigue of steel 45 / steel 25ÕÃÒ active system (L A Sosnovskiy, S A Tyurin) The procedure of different determination of the asymmetry and irregularity coefficients can be used, viz. they can be recorded in the order of magnitudes dñ:
It is quite apparent that the coefficients determined from formulas (4.18) are unequal to the corresponding coefficients determined from formulas (4.16) and (4.17). Selection of the type of presentation of the coefficients R and h is dictated by the purpose of a specific analysis. |
