Bayer R.G. Mechanical Wear Fundamentals and Testing, Revised and Expanded

Подождите немного. Документ загружается.

great deal of ﬂexibility in the load, sphere radius, speed, and materials used, as well as

with the conditions surrounding the test (e.g., lubrication, temperature, and humidity).

To a large degree, these parameters can be adjusted to simulate magnetic recording

applications. However, the basic geometry of the co ntact situation and stress system

are signiﬁcantly different than in typical applications (e.g., those shown in Fig. 9.72).

The method used to rank magnetic tapes in terms of their abrasivity is similar to that

used for printer ribbons, namely, to determine a wear coefﬁci ent. This wear coefﬁcient

is the volume of wear divided by the product of the normal load, speed, and time of the

test. The higher the value of this coefﬁcient, the more abrasive the tape is.

The controllable parameters and the use of a wear coefﬁcient in this test provide a

more general description of the wear situation and makes the results more suitable to

general models and theories for wear. In the SCAT test, many of these parameters

are indirectly controlled and not speciﬁcally identiﬁed. This illustrates some of the differ-

ences typically associated with the two categories of tests. In general, the phenomeno lo-

gical test tends to have a potential for greater applicability or generalization than the

operational test.

With appropriate values of parameters, the spherical test can rank tapes of signiﬁ-

cantly different abrasiveness that is consistent with ﬁeld data. However, the wear rates

in the test are typically several orders of magnitude higher than observed in practice

(e.g., 10

times higher). While this is the case, the wear rates from the SCAT test

are typical of those observed in practice. This difference between the two tests is the

result of poor simulation of the contact situation in the spherical test and good simula-

tion in the SCAT test. As this implies, the spherical test provides a high degree of

acceleration and allows tests times to be reduced from hours, required with the SCAT

test, to minutes. At the same time such a difference introduces more concern regarding

correlation with the ﬁeld and the ability of the test to provide accurate rankings. This

illustrates another typical difference between phenomenological and operational tests.

With an operational test, there is usually less concern with correlation, since the degree

of simulation is high, and a minimal amount of effort is needed to establish correlation.

With a phenomenological test, there is generally more concern in this area and more

effort is required to establish correlation. In this particular case, there is a higher degree

of conﬁdence in the rankings provided by the SCAT test than those provided by the

spherical test, because of the higher degree of simulation in the SCAT test.

Figure 9.74 The conﬁguration of a phenomenological test used for magnetic head applications.

9.3.3. Coin Wear Test

One type of test that has been used to simulate the wear of coins is a tumbling test. Test

coins are placed inside a suitably designed drum and tumbled (80,81). The basic concept

of the test is to simulate the rubbing that coins experience against one another in pockets,

change drawers, etc. This is achieved by placing coins is a plastic drum, which is lined with

a rubber-backed cloth. The interior of the drum has an axial ridge, which provides agita-

tiontothecoinsasthedrumisrotated(Fig.9.75

). To simulate possible chemical effects

associated with hand ling of coins, the co ins can be coated or dipped with artiﬁcial per-

spiration. Basically the test method consists of tumbling the coins for a period of time

and determining mass loss. This is done at intervals so that a wear curve can be developed,

which is used to establish an average wear rate (i.e., wear per unit time). In addition

to directly using the mass loss as a measure of wear, thickness reduction is also used

particularly when comparisons of different materials are involved. This is done by divi-

ding the mass loss by the product of density and the nominal surface area of the coin

(including the side area).

This test was investigated and used to evaluate the wear rates of different coin mate-

rials (80). As part of the investigation of this test and its correlation to the ﬁeld, the inﬂu-

ence of several of the test parameters were investigated to select optimum and desirable

values for these parameters. The goal was to select values which result in producing similar

wear characteristics to that found in the ﬁeld, to minimize test time, and to reduce scatter

in the data. Included in these were the effects of the number of coins in the drum, cylinder

size, drum rotational speed, and the inﬂuence of chemical agents. It was concluded that the

number of coins in the drum, the size of the drum, and total linear distance of rotation

inﬂuence the wear. Total linear distance is the product of the inner circumference, revolu-

tions per unit time, and time. It was found that different sizes of coins would wear at dif-

ferent rates, depending on the number of coins tumbled, up to a total of 12 coins after

which the same wear rate was obtained. Thus, 12 coins are used in the standard test.

For a given drum size it was found that the wear curve was linear and that the amount

of wear was simply dependent on total linear distance. However it was found that the wear

rate was higher with smaller diameter drums; hence a smaller drum and higher rpm was

Figure 9.75 ‘‘A’’ shows the overall apparatus used to study coin wear. ‘‘B’’ shows an individual

drum that is used to tumble the coins. (From Ref. 126, reprinted with permission from ASME.)

chosen as the standard test. It was also found that the chemical agents and the manner of

their application inﬂuence wear behavior, which lead to the well-controlled use of artiﬁcial

perspiration as a standard part of the test procedure.

In addition to these sensitivity evaluations during the initial portion of the develop-

ment of the test, the morphology of tested coins was also compared to that of ﬁeld-worn

coinstoinsurethatitwasrepresentative.AsillustratedinFig.9.76

, the appearance of

tested coins was similar to used coins. In addition since ﬁeld-worn coins tended to show

increases in hardness with time in use, tested coins were also examined for an increase

in hardness. A similar behavior was found with tested coins, supporting the case for simu-

lation. Furthermore, this increase in hardness suggests that a major factor in the wear is

coin=coin interactions. This is consistent with the results of the sensitivity studies. The

increased wear rate observed with larger number of coins and small drum size indicate

that wear by the drum liner is minor. The standard test conditions selected tend to

enhance this type of interaction.

This is an accelerated test in terms of time or use. Comparison of test wear rates

and ﬁeld wear rates indicated acceleration factors in the range of 100–1000 times. The

actual value is dependent on the conditions of use (e.g., degree of coin-usage in the

society and the general environmental co nditions associated with that society). For

example, comparison with Canadian coins indicates a value of 600, while a similar com-

parison with El Salvadorian coins indicates a value of 300. A partial reason for this

difference is probably associ ated with the difference in coin-usage in the Canadian

and El Salvadorian societies (i.e., more coin-usage is considered likely in El Salvador).

Also, the environmental conditions of the two countries are signiﬁcantly different.

Figure 9.76 Examples of coins worn in the tumbling apparatus. (From Ref. 126, reprinted with

permission from ASME.)

El Salvador has a hotter, more humid, marine-type atmosphere than does Canada.

The laboratory test s indicate that such environmental conditions will tend to increase

wear rate.

While this test does not provide a characterization of wear behavior in terms of more

fundamental parameters, such as load, speed, or sliding distance, it not only provides a

ranking of materials but it also provides a quantitative assessment of ﬁeld performance

through an acceleration factor. These are frequently features of operational tests. With

these attributes, such tests are very useful for speciﬁc applications. On the other hand,

it is difﬁcult to apply the results of such tests to other applications or to general wear

behavior.

9.3.4. Test for Rolling with Misalignment

This test was developed to address problems associated with a linear stepper motor used

inarobot(Fig.9.77

). In this motor, the stator acts as a rail upon which the armature

moves back and forth. Conventional ball bearings are used as wheels for the armature,

with the outer race being the wheel surface, which engage the stator. In this type of

actuator, the wheels serve to support the weight of the armature and any other weight that

is being trans ported and to provide guidance with a minimum of friction . The magnetic

coupling between the stator provides accelerations and deceler ations parallel to the sur-

face of the rail and the armature. Some degree of misalignment between the ball bear-

ings and the surfaces of the stator was likely to be present in most assemblies and

considered to be a signiﬁcant factor in wear life. To obtain adequate life, there was a

need to select and optimize several design parameters in terms of their inﬂuence on wear.

This included the selection of material s and the evaluation of lubricants. In addition, it

was also necessary to select bearing size and the contour of their outer race, evaluate the

effect of load, and determine tolerances needed to control alignment. All of these can

affect wear and are interrelated tribosystem parameters.

The wear situation may be described as rolling wear with some slip. Based on the

general trend, it is expected that the wear rate will increase with the amount of slip

Figure 9.77 Diagram of the rolling contact situation occurring in a linear stepper motor

application.

occurring in the contact and that this will be a major factor in overall wear behavior.

There are two potential sources of slip: one is misalignment, with the amount of slip

increasing with the degree of misalignment present and the other is slip during starting

and stopping of the motor. Since the rollers do not provide traction for starting or stop-

ping, the latter source was considered unlikely and actual motor tests with different

ratios of start=stops to total distance traveled conﬁrmed this. Consequently, it was

concluded that it was not necessary to simulate start=stop behavior in a wear test used

to evaluate materials and design parameters.

A test apparatus and technique was developed to investigate the inﬂuence of the var-

iousdesignparametersinthiswearsituationandtoselectmaterials(82)(Fig.9.78

). It con-

sisted of a driven central cylindrical wear specimen, which rotates about its axis. This

cylinder simulates the rail or stator. Pressed against the wear specimen are three ball bear-

ings located 120



apart at three different locations along the axis of the wear specimen.

These can be varied and simulate the rollers in the application. The apparatus was de-

signed so that the normal load, rotational speed, and roller alignment could be varied as

well. Hence, the contact situation in this app aratus was very similar to that in the appli-

cation. There were some differences, which were not considered signiﬁcant. One apparent

difference is that in the test apparatus the contact is between two cylinders, while in the

application it is between a cylinder and a ﬂat. This difference is minor since the general

nature, i.e., a line contact, and stress dist ributions of the contacts are the same according

to Hertz contact theory. Another difference was in terms of the relative wearing action

that the two members experience. The bearing surfaces in the tester experience about

2.5 times more wear acti on than the wear specimen. In the application, the bearings

Figure 9.78 Conﬁguration of the tester used to investigate the wear between the roller and the sta-

tor in a linear stepper motor application. (From Ref. 127, reprinted with permission from ASME.)

experienced a varying amount, depending on the movement associated with the robotic

action and the ratio can vary from something less than 1 to more than 10, with higher

values being more typical. This was not considered to be signiﬁcant since the ratio in

the test was in the range of that for the application. The use of this type of wear apparatus

was considered to be more advantageous than motor tests, because it provided adequate

simulation with a greater degree of control of key design parameters of load, alignment,

speed, ball bearing geometry, materials, and lubrication.

The rollers were positioned in the test apparatus such that they did not interfere with

one another. Their wear tracks are separate and, since 120



separate them, wear debris

from one track would not contaminate another. As a result, it is possible to simultaneously

conduct tests at three different alignments at the same time. This has an advantage in that

it facilitates the assessment of misalignment effects as a function of the other design para-

meters. For example, the developers of the test utilized three standard angles for much of

their evaluations, namely 0



, 0.117



and 0.235



of axial misalignment. Examples of some

ofthedifferencesseeninthismannerareshowninFig.9.79

The test method consisted of characterizing and measuring the amount of wear as a

function of number of revolutions. This was done for different combinations of design

parameters. For the wear specimen, the maximum depth of the scar was used as a measure

of wear. This was determined by means of proﬁlometer trace s through the wear scar,

which gives a natural reference surface for this type of measurement. Since the outer races

of the ball bearings are usually signiﬁcantly harder than the surfaces of the rail materials,

little wear is produced on the roller surfaces. However, surface modiﬁcations did occur.

These were characterized by optical and SEM micrographs, EDX, and roughness

measurements. Features such as the occurrence of transfer, oxidation, polishing, and

smoothing were noted and used in the overall wear assessment of the system. These tech-

niques were used to examine the wear specimen as well, and similar notations were made.

Figure 9.79 The effect of misalignment on the wear of the center specimen in the rolling test.

(From Ref. 127.)

In addition to the measurements and characterization of the wear scars, wear deb ris

was collected and characterized in terms of size, morphology, and composition. These

observations were then integrated into an overall wear assessment of the system.

The maximum depth was used since the scar proﬁle is typically tapered because of

the misalignment. As is pointed out in Ref. 82, misalignment produces a moment at the

interface, which results in nonsymm etric loading of the interface. As a result, a nonuni-

form wear scar tends to occur. A typical proﬁle for a misaligned contact situation, along

withaloadingdiagram,isshowninFig.9.80

A wide range of wear behavior was observed in this test. Depending on the amount

of misalignment, loads, and materials involved mild or severe wear resulted. Wear scars,

which have a fretting corrosion morphology, were obtained, as were ones with a morpho-

logysimilartogrosssliding.Figure9.81

shows some of the conditions observed in the test.

Figure 9.80 Illustration of the effects of misalignment on the loading between the roller bearing

and the center specimen and the proﬁle of the wear scars on the center specimen. (From Ref. 127.)

Figure 9.81 Examples of the wear scars occurring on plated center specimen in the rolling test with

misalignment. The amount of misalignment in the test increases from left to right, 0



, 0.117



, and

0.235



. Flaking of the plating is evident in the center scar of ‘‘B’’. (From Ref. 127, reprinted with

permission from ASME.)

Similar wear behavior, including the occurrence of tapered wear scars, was found in motor

tests, providing veriﬁcation of the simulative aspects of the test. Of the many parameters

investigated with this test, alignment was the overriding factor. Misalignment was found

to affect both the magnitude and nature of the wear. The magnitude of wear increased

with increasing amounts of misalignment. Transitions not only in the type of wear but

from mild to severe wear were also found to be a function of misalignment. The experi-

mental data indicated a critical angle for misalignment, above which a rapid increa se

inwearratewasobserved(Fig.9.82

In addition to the experimental aspects of the study, a theoretical model for such

behavior was developed and used to explain the general trends observed (82). In this

model, a critical angle is identiﬁed. Misalignment above this value results in slip over

the entire contact region. Below this value, the compliance of the surfaces is sufﬁcient

tolimitthesliptoregionswithinthecontactzone.Thefollowingequationforthiscritical

angle identiﬁes the parameters involved:

3:625mP

wbE

ðvÞþc

ðvÞ

 log

hi

ð9:23Þ

where m is the coefﬁcient of friction; P is the normal load; w is the length of the contact; b

is the width of the contact; E is the reduced modulus; n is Poisson’s ratio. Values for c

andc

aregiveninTable9.5.ThisrelationshipisillustratedgraphicallyinFig.9.83.

This test was used to investigate a number of design parameters and aspects as well

as the evaluation of materials and lubricants. The test was used to determine the allowed

rangeofmisalignmentforthedesign,thesizeoftherollers,whetherornotcrownedrollers

shouldbeused,andtoevaluatetheeffectofedgeconditions.Sincethetestingwasdone

with speeds and loads appropriate to the application, selections made on the basis of this

test were directly applicable to the application and consistent with its load capacity and life

requirements. The high degree of simulation allowed the estimation of wear life by the

extrapolation of the data.

Figure 9.82 The general relationship between wear depth and misalignment observed in the rolling

tests. (From Ref. 127, reprinted with permission from ASME.)

A phenomenological rolling wear test, which could be conducted with slip (Sec.

3.2.10), could pro bably have been used to rank materials in terms of their wear resistance

under combined conditions of rolling and sliding. However, that test canno t provide all

the information that the operational test did. The various effects of misalignment that

were found to occur in the application and signiﬁcant in terms of life cannot be studi ed

in the phenomenological test (e.g., critical angle and loading alterations), since these

depended on the unique conditions of this application. The slip in the rolling wear test

is the result of different rotational speeds for the rollers, not from misalignmen t. In the

test performed the effect on wear results from the combined inﬂuence of misalignment,

loading, and material properties. As a consequence, the phenomenological test only pro-

vides rankings and cannot be used directly in addressing load capacity and life require-

ments. This illustrates a typical difference between these two types of tests.

Table 9.5 Values for C

and C

of Eq. (9.23) for

the Critical Slip Angle

Poisson’s ratio

0 0.1177 0.2022

0.1 0.1198 0.2224

0.2 0.1202 0.2426

0.3 0.1188 0.2628

0.4 0.1156 0.2831

0.5 0.1107 0.3033

Source: Ref. 126.

Figure 9.83 The relationship between critical angle of misalignment, friction, modulus, and load in

rolling. (From Ref. 127.)

9.3.5. Bearing Tests

The wear life of a bearing can be inﬂuenced by a variety of factors, which are difﬁcult to

address in phenomenological tests. Also, there can be several wear points and aspects of

the wear that potentially determine life and that are inﬂuenced by different factors and

interactions. For example, in the case of roller or ba ll bearings the mixture of rolling

and sliding that takes place at various locations in the bearing is a determining factor

in the life. In these types of bearings, there are two general wear situations which poten-

tially determine life. One is the contact between the cage and the balls or rollers; another is

the contact between the balls or rollers and the races. Which one determines life and the

life itself is signiﬁcantly affected by the general loading conditions (e.g., mixture of axial

and radial loads, nature of time varying loads), type of motion (e.g., oscillatory or rota-

tional), and preload. Other factors are also involve such as geometrical tolerances, lubrica-

tion, and thermal expan sion. Frequently, these elements interact in a complex way to

determine the nature and the location of the contacts among the various elements of

the bearing (83–86). With journal bearings, the clearance between shaft and bearing

can be a factor in wear behavior, as well as alignment, type of loading, and motion (e.g.,

frequent or rare stop=starts). In journal bearings, these elements can directly inﬂuence the

wear in terms of contact pressure or location of the contact and indirectly inﬂuence wear

by their inﬂuence on the type of lubrication which occurs (e.g., boundary or elastohydro-

dynamic(EHD)).

Again, the effects of these parameters can be co nvoluted and interactive.

Consider the situation with respect to clearance; temperature effects clearance and

clearance inﬂuences heat dissipation and therefore tempe rature. Clearance, per se, also

can effect load and pressure distributions and the formation of triboﬁlms. These factors

in turn can inﬂuence friction, which inﬂuences temperature, etc. (87–89 ). These complex

interactions and their inﬂuence on the nature and location of the wear are difﬁcul t to

simulate in phenomenological tests. As a result, bearing wear tests are used to investigate

these complex aspects, determine design information, and verify performance.

A wide variety of bearing testers, differing in complexity, instrumentation, and size

havebeenusedforthispurpose.SeveraloftheseareillustratedinFigs.9.84

and9.85. The

basic element of these testers is a bearing conﬁguration representative of an application or

typeofapplication.Figure9.86

shows this generic representation. Many of the bearing

tests contain instrumentation to measure both friction and temperature. A variety of wear

measures are used in these tests, depending on a large degree on the nature of the bearing

and the failure criteria. For journal bearings common ones are various forms of measuring

increases in clearance between shaft and bearing (e.g., such as end-play, elongation of

hole, etc.), dimensional and roughness changes on shaft and bearings, and volumetric

and mass changes; the last of these provides a more fundamental characterization. Some-

times increases in friction level and operating temperatures have been used as failure cri-

teria in these types of evaluations. With ball and roller bearings, the measures tend to be

indirect. Frequently, vibration characteristics are used to monitor bearing performance,

since vibrations tend to increase as wear takes place. Also, play or slop in the bearing

can be used as a measure. More basic wear measurements on the various components

of the bearing can be used as well (e.g., depth of wear scar on race, mass loss of cage).

However, in many cases, the useful life of the bearing is associated with very smal l geome-

trical changes on these components. For example, when increased vibration level s asso-

ciated with the end of life are detected, the wear on the rollers and races may only

manifestthemselvesasincreasesinroughness,suchasshowninFig.9.87

. Increases in

temperature and friction in these bearings are also used as measures of life in some cases.