Astm stp 1105 1991

STP 1105

Tribological Modeling for

Mechanical Designers

Kenneth C Ludema and Raymond G. Bayer, editors

As M

1916 Race Street

Philadelphia, PA 19103

Downloaded/printed by

University of Washington (University of Washington) pursuant to License Agreement. No further reproductions authorized.

Library of Congress Cataloging-In-Publication Data

Tribo]ogica] modeling for mechanical designers / Kenneth C Ludema

Raymond G. Bayer, editors.

(STP ; 1105)

Papers from a symposium held in San Francisco on May 23, 1990,

sponsored by ASTM Committee G-2 on Near and Erosion.

Inc]udes bibllograohica] references and indexes.

ISBN 0-8031-1412-5

1. Tribo]ogy--Rathematical models--Congresses.

II. Bayer, R. G. (Raymond George), 1935- III.

G-2 on Erosion and Wear. IV. Series: ASTH special

publication ; 1105.

TJ1075.A2T725 1991

821.8'9'015118--dc20

and

I. Ludema, K. C

ASTM Committee

technical

91-8238

CIP

No part of this publication may be reproduced, stored in a retrieval system, or transmitted,

in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise,

without the prior written permission of the publisher.

NOTE

The Society is not responsible, as a body,

for the statements and opinions

advanced in this publication.

Peer Review Policy

Each paper published in this volume was evaluated by three peer reviewers. The authors

addressed all of the reviewers' comments to the satisfaction of both the technical editor(s)

and the ASTM Committee on Publications.

The quality of the papers in this publication reflects not only the obvious efforts of the

authors and the technical editor(s), but also the work of these peer reviewers. The ASTM

Committee on Publications acknowledges with appreciation their dedication and contribution of time and effort on behalf of ASTM.

Printed in Baltimore

October 1991

Downloaded/printed by

University of Washington (University of Washington) pursuant to License Agreement. No further reproductions authorized.

Foreword

This publication, Tribological Modeling for Mechanical Designers, contains papers presented at the symposium of the same name held in San Francisco, CA on 23 May 1990. The

symposium was sponsored by ASTM Committee G-2 on Wear and Erosion. Professor Kenneth C Ludema of the University of Michigan in Ann Arbor, MI and Raymond G. Bayer of

IBM in Endicott, NY presided as symposium chairmen and are editors of the resulting

publication.

Downloaded/printed by

University of Washington (University of Washington) pursuant to License Agreement. No further reproductions authorized.

Contents

Overview--K. c LUDEMA AND R. G. BAYER

WHAT MECHANICAL DESIGNERS NEED IN TRIBOLOGICAL MODELING

Comments on Engineering Needs and Wear Models--R. G. BAYER

Design of Plain Bearings for Heavy Machinery--w. A. GLAESER

WHAT IS AVAILABLE IN TRIBOLOGICAL MODELS

(MOSTLY FOR WEAR)

The Structure of Erosive Wear Models--s. BAHADUR

Success and Failure of Simple Models for Abrasive Wear--J. LARSEN-BASSE

Wear by Chemical Reactions in Friction Contacts--J. L. LAUER

Tribological Models for Solid/Solid Contact: Missing Links--s. L. RICE AND

F. A. MOSLEHY

DATA BASE AND SIMULATION ISSUES FOR TRIBOLOGICAL MODELING

Friction in Machine Design--K. G. BUDINSKI

Considerations on Data Requirements for Tribological Modeling--A. w. RUFF

Classification of Metallic Materials from a Viewpoint of Their Antiwear

Behavior--T. SASADA

Wear Transition Surfaces for Long-Term Wear Effects--c. s. YUST

PRINCIPLES OF MODEL MAKING AND USE

IS Modeling in Tribology a Useful Activity?--J. R. BARBER

Wear Modeling: How Far Can We Get With Principles?--M. GODET,

Y. BERTHIER, J. LANCASTER, AND L. VINCENT

Cultural Impediments for Practical Modeling of Wear Rates--K. C LUDEMA

vii

127

143

153

165

173

180

Downloaded/printed by

University of Washington (University of Washington) pursuant to License Agreement. No further reproductions authorized.

Overview

Purpose

The symposium for which the following papers were written was

organized out of the recognition that those tens of thousands of mechanical

designers who design consumer products need far better information than

they now have when they design mechanical components for wear life.

They have equations (tables, graphs, guidelines, etc.) for the analysis of

stresses, for vibration modes and natural frequencies, for rates of heating

and cooling, and for most other phenomena - but very little for the wear life

of products.

The needs of designers may best be seen in the dichotomy between the

mechanical sophistication of machines and devices, and the fact that these

devices are most often discarded because of mechanical wear. Tribological

adequacy seems to be one of the last considerations in the design process

if it is explored at all, and probably for good reason - it is very complicated.

Tribological design requires knowledge of materials (including lubricants),

surface making processes, running-in procedures and assembly

procedures. The designer is handicapped because neither friction nor

wear are intrinsic properties of material in any form, but rather are highly

dependent on the mechanical system and how it is run. Most designers

have been caught in attempting to upgrade products only to find that the

new product fails too often. Some then attempt a test program, only to find

that there is no correlation between test results and the functioning of

production items.

Proaress in wear modeling

The impetus for developing useful information on the wear properties of

material comes mostly from those in research, referred to here as research

tribologists. Their first priority Is to maintain research activities and write

scholarly papers. By the nature of their work tribologists select relatively

impractical materials and experimental parameters, and interact mostly with

others who do the same. However, some tribological concepts have

diffused into general design practice. The most common are equations for

designing fluid film bearings. Further mature concepts have made their

way into the design of rolling element bearings, belts, gears, pumps, etc.

such that predictions can easily be made of functional product life.

Whereas many mechanical devises can be built up with such components,

many consumer products can not, because they must sell at the lowest cost.

The majority of designers are connected with consumer products.

This is the third symposium on modeling for wear resistance, each with

different sponsors. The first was held at Columbia University in New York

City, December 17-19, 1986 (1) and the second was held at Argonne

National Laboratories (2). These were attended mostly by researchers, and

by invitation. These were serious efforts and much good information was

exchanged. It may be seen from the proceedings of these symposia that

each of the specialties in tribology communicates in very different and

esoteric language, compared with the needs of designers.

vii

Downloaded/printed by

University of Washington (University of Washington) pursuant to License Agreement. No further reproductions authorized.

The third symposium, the ASTM Symposium of May 23,1990 reported

here, sought to meet the needs of designers. Authors were invited to show

how the great chasm between research language and designers needs

could be bridged. Perhaps the extent of the chasm may be seen in that

only two authors from industry submitted papers. The great majority of the

papers were written by research tribologists. The latter were written from

the perspective of a physicist, a chemist, several in materials engineering,

on specialist in solid mechanics and five mechanical engineers. The latter

are "near" the design process, but do not often design consumer products.

Overview of the papers of the symposium

To a great extent the authors report that they have a long way to reach in

order to reach designers. Designers have an equally long reach, but they

have no better idea than do research tribologists which direction to reach.

Our authors made a valiant effort to propel us toward sensible wear models.

Most authors agreed on the nature of the problem and some offered

specific improvements in the understanding of wear. In particular, some of

the points made were the following, with editorial comment:

1. Research papers in tribology contain information that is rarely

applicable to practical problems. The reasons may include:

a. Terminology is a major point of confusion in the field. This is

probably a consequence of the presence of several very different

academic disciplines in the field.

b. Research papers focus on very few of the operating variables and

phenomena in real machines that control wear. Research papers

range from the "near applied" to the fundamental, the latter often

from the point of view of the atomic and molecular structure

through the sliding interphase region.

c. Attempts to harmonize the methods in the various specialized

areas in tribology are largely philosophical and not well directed.

d. Research results as presented seem to imply that the dual

phenomena of friction and wear are uncoupled from each other

and from considerations of the mechanical properties of the

machinery holding the sliding pairs.

e. Research results rarely provide information on the changes that

occur at interfaces (debris formation and migration, eg.) over

time.

2. The greatest advances in tribology have been made in capital products

and machinery that are expected to last for a long time. The design of

consumer products involves minimum cost for material and

manufacture, which involves variables (surface roughness, materials

variation, etc.) that have been inadequately studied.

3. Several wear models do exist but these are extremely limited in scope

and applicability. Unfortunately, the limits of applicability of these

models are rarely published. In fact, the literature would suggest, by

virtue of the lack of comment, that the available models are universal

in application. This is particularly misleading in designing weal" tests

when only those variables that appear in simple models are thought to

be the controlling variables in all sliding systems.

viii

Downloaded/printed by

University of Washington (University of Washington) pursuant to License Agreement. No further reproductions authorized.

4. Designing for wear resistance outside of the scope of the current limited

models should be done primarily by empirical methods in the next

decade. This is so because it is not reasonable to expect the many

relevant and disparate variables in wear to be rationalized in the next

decade, whether in the form of broadly applicable equations, models,

algorithms or handbook entries. The same applies to wear tests as

well as to the selection of materials.

The empirical method includes:

a. Gathering data from practical sliding elements over a reasonable

range of controllable variables. The entire system, including the

machinery surrounding the sliding surfaces, must be thoroughly

characterized.

b. A data base of research results, for equally well characterized

laboratory systems should be (and is being) gathered

c. Bench wear tests should be done but only after the results of the

bench test are known to correlate very well with the results from

the practical system being simulated.

d. Special attention should be paid to wear debris and other residue

- the manner in which sliding systems retain or flush out debris,

which will depend on, among other things, specimen shape,

vibration characteristics and duty cycles. Efforts should be made

to trace the chemical and mechanical "pathways" by which the

debris and residue was formed, transformed or ejected.

A significant fraction the efforts of research tribologists should be devoted to

such empirical work.

Overall, tribology is seen to be a very broad and complicated topic.

There is a major problem in communication across the field which should

be addressed in the next decade. Research tribologists should devote

some of their efforts to making their results useful, but designers should

indicate what they need from research tribologists. Many more symposia

on wear modeling must be held.

1. Approaches to Modeling of Friction and Wear, Proc. Workshop on the

Use of Surface Deformation Models to Predict Tribology Behavior ,

Eds. FF. Ling and C.H.T. Pan, Springer-Verlag

2. Proc. of the International Workshop on Wear Modeling, June 16 and 17,

1988, Eds, F.A. Nichols, A.I. Michaels and L. Northcutt,

DOE-Conf.-8806370, June 19,1989).

Ken Ludema

University of Michigan,

G.G. Brown Building

Ann Arbor, MI 48109-2125

Downloaded/printed by

University of Washington (University of Washington) pursuant to License Agreement. No further reproductions authorized.

What Mechanical Designers Need in

Tribological Modeling

Downloaded/printed by

University of Washington (University of Washington) pursuant to License Agreement. No further reproductions authorized.

Raymond G. Bayer

COMMENTS ON ENGINEERING NEEDS AND WEAR MODELS

REFERENCE: Bayer, R. G., "Comments on Engineering Needs

and Wear Models, = Tribolo~ical Modelin~ for Mechanical

Desis ASTM STP 1105, K. C. Ludema and R. G. Bayer,

Eds., American Society for Testin B and Materials,

Philadelphia, 1991.

ABSTRACT: Engineering applications can involve complex

tribological considerations other than the ranking and

selection of materials. These can include such aspects as

assessment of the, significance of a variety of design

parameters, determination of tolerances for these parameters,

comparisons of designs, and life projectlons. Several of

these are illustrated in terms of actual situations

encountered by the author, along with the methods used to

address these aspects in those cases. The typical needs of

such applications of tribology are identified, and current

approaches in wear modeling and wear testing are compared to

these needs. Inadequacies are identified and suggestions as

to how to address these are made. In addition, the

differences between a material engineering and a more general

design approach to engineering problems are identified, and

the significance of these in terms of modeling ~nd testing

are discussed.

KEYWORDS: wear models, wear tests, wear design, wear

prediction, wear transitions

INTRODUCTION

The subject of wear models is one of current and perhaps perennial

interest. One can go back fifty or so years and find models for wear.

More recently there have been informal and formal meetings sponsored by

various technical and governmental agencies of wear models (1,2,3).

This volume on the proceedings of ASTM's Symposium on Tribological

Modeling for Mechanical Designers reflects the current interest in this

area. In these meetings, and in discussions with tribologists and

engineers regarding wear models, a variety of meanings for the term

"wear model" can be found along with a variety of expectations for

Raymond G. Bayer is a senior engineer at International Business

Machines Corporation, Systems Technology Division, Endicott, NY. He is

co-chairman for this year's Symposium on Tribological Modeling for

Mechanical Designers.

Downloaded/printed by

University of Washington (University of Washington) pursuant to License Agreement. No further reproductions authorized.

4 TRIBOLOGICAL MODELING FOR MECHANICAL DESIGNERS

models. In some cases a phenomenological description of the wear

process is considered a model. In others it is a quantitative

relationship for a physical mechanism. Still in others it is

specifically a quantitative relationship between design parameters and

wear. The term wear model has also been applied to a qualitative

description of wear behavior and a resulting set of design guidelines.

In some cases it has even been applied to a wear tester or robot. It

is interesting to note that these concepts of what constitutes a wear

model tend to vary with the type of involvement with tribology. For

example, an empirical physical scientist would tend to the first

interpretation. A theoretician would tend to think in terms of the

second, while with a cut-and-dry engineer the last interpretation might

be found.

The uses of or expectations from a model also vary with the type of

involvement with tribology. With the fundamentalist the understanding

that is associated with a model is the likely goal of the model. For

the mechanical or design engineer the model should provide guidance in

the selection and development of a design. Materials engineers and

scientists, on the other hand, would expect a model to guide them

either in the selection or the development of materials by providing an

understanding of the material parameters which affect wear behavior.

These observations imply that in any discussion of wear modeling it

is appropriate to identify from whose perspective the subject is being

approached and what are the intended uses of the models. In this paper

the perspective will be from a mechanical design engineer who wants a

model to relate wear performance in his application to design

parameters. While in this broad category there can be significant

variations in the uses of models or the type of model required, there

are some general features that can be identified. One is that the

general situation involves more than material optimization. Another is

that a predictive element be provided by the model. In this paper a

general aspect of this type of use of models will be illustrated in

terms of a specific example. From this some general attributes of

models for these types of applications will be identified. The current

state of tribology and tribological modeling will then be reviewed with

respect to these needs and attributes. Following this some con~nents on

the relationship of wear testing to these needs and modeling also will

be made. Then, in conclusion~some suggestions for future work and

modeling development will be made.

ENGINEERING APPLICATION

A case study involving a printed circuit edge connector system can

be used to illustrate the type of concerns with wear that can arise in

a design engineering environment. In Figure 1 the card edge connector

system is illustrated. The metallurgy of both the spring and the tab

are composed of two or more electroplated layers, the outermost layer

being Au or some other noble metal to resist corrosion and provide

stable contact resistance. A thin film of lubricant is also applied to

the tab surface. The failure criteria for this connector is the

exposure of the base metal, which would allow corrosion to take place.

Downloaded/printed by

University of Washington (University of Washington) pursuant to License Agreement. No further reproductions authorized.

BAYER ON ENGINEERING NEEDS AND WEAR MODELS 5

METALIZEDI LAYER

INSERTION/EXTRACTION

CONTACT SPRINGS

FILLED RESIN /

Figure 1 -- View of Card Edge Connector System.

Motions Due to Vibration Can Occur in Any Direction.

Metalized layer consisted of a surface layer of Au or Pd-Ni

above an intermediate layer of Ni. Substrate was Cu.

In the original application of this connector system, the only source

of relative motion between the tab and the spring was during insertion

and extraction of the card. The wiping action of the spring against

the tab during the engagement of the contact is a design feature

intended to remove contaminants from the active contact surfaces. The

range of the load between spring and tab used in the design is also

governed by criteria to insure good contact resistance behavior. Since

the contact system was expected to experience only several dozen card

insertions and actuations over life, wear evaluations were handled

empirically. The design was successfully released and used.

As frequently happens in an industrial environment, an additional

application was considered. In this case the mass of the cards to be

used increased by approximately seven times from those considered in

the original application. As a result the possibility of having

relative motion between the tab and the spring during shipping and

machine operation had to be considered. In fact, some vibration

testing with actual hardware showed very quickly that not only could

such motions occur but that they could also produce a significant

amount of wear and failure. This situation generated a series of

questions, such as the following:

How much motion can the present design tolerate?

How much improvement in the wear resistance is needed?

What improvement will different metallurgies provide?

How does insertion wear interact with vibration wear?

Is lubrication adequate for the vibration situation?

Is there a better lubricant?

What are the effects of design tolerances on this wear?

Do the vibration tests used in company qualifications provide

adequate simulation of this type of wear?

Downloaded/printed by

University of Washington (University of Washington) pursuant to License Agreement. No further reproductions authorized.

6 TRIBOLOGICAL MODELING FOR MECHANICAL DESIGNERS

Do those tests need to be modified?

Is there a frequency effect on the wear?

If the amplitude is reduced by a factor of .5 or .i, how much

improvement in wear will occur?

What are the effects of hardness variations in the materials

on the wear?

While a similar set of questions can be associated with the

initial application, there is one significant difference. In the

initial situation the wear condition allowed the questions to be easily

addressed empirically. With this new situation, the complexities and

duration of vibration testing, as well as variables and unknowns

associated with a vibration environment, made it desirable to develop a

model relating key parameters. With a model, testing can be reduced

and the effects of various design changes can be assessed analytically.

The development of the model started with a few assumptions. One

was that the wear was the result of sliding. The second was that it

was the total amount of sliding experienced that controlled the amount

of wear. In effect this means that the rubbing from insertion, shock

and vibration could be added. The thirdwas that the wear could be

described by an equation of the following form,

h = K x pn x S n (i)

where h is the depth of wear; P, the load; S, the amount of sliding; K,

a factor dependent on lubrication and materials. Potentially the

exponents m and n can depend on lubricants and materials as well.

Empirically this equation was verified, as well as the other two

assumptions. The following equations, describing the wear, were

determined:

h K x pO.5 S0.2 = x , thick Ni under-plate (2)

p4.5 $0.2,

h = K' x x thin Ni under-plate (3)

In addition to these relationships, a simple wear test, utilizing

a reciprocating ball-plane apparatus (4), was developed to determine

values of K and K' for several materials systems and lubricants.

Figure 2 illustrates the wear test, the method of analysis, and the

resulting data. In these tests, loads ranged from 125 to 250 gm;

cycles, 1 to i0~; stroke length, .35 to 6 ~m; repetition rate,

i0 cycles/minute; environment was room ambient.

With this information available, various design options were

considered. For example cost versus performance trade-offs were

examined. The reliability for various field scenarios could be

assessed and various strategies developed to address them. In

addition, the model provided an analytical way of determining the need

for design modification to reduce vibrations and an analytical means of

assessing the impact of design modification on the wear. Finally, with

this model, appropriate qualification tests for the product were

identified.

Downloaded/printed by

University of Washington (University of Washington) pursuant to License Agreement. No further reproductions authorized.

BAYER ON ENGINEERING NEEDS AND WEAR MODELS 7

CARD

SECTION r J

LOAD, P I/ s[ ,NO O D

~ TO BEAM

OSCI LLATING PLATEN

AMPLITUDE, D

TEST CONFIGURATION

LOAD

10/lm ~P1

I I /

1 N i 10 4 LOG N

CYCLES WEAR CURVE PRODUCED

Ki =

(DNi)2 pn

ALOG h

n = ,~LOG N

~Ki

i=1

= WHERE

Figure 2 -- Method for Determining Wear Coefficient

In the design situation, the general advantage of this type of

model and associated data is that development time is reduced, hardware

cost for tests with prototypes or in simulations is reduced, various

design options and needs can be addressed analytically, and wear

considerations can be factored into the design process at the

beginning, not after hardware is built and tested.

Downloaded/printed by

University of Washington (University of Washington) pursuant to License Agreement. No further reproductions authorized.

8 TRIBOLOGICAL MODELING FOR MECHANICAL DESIGNERS

This specific example illustrates some of the general features

that are desired for an engineering model. These may be sun~arized as:

1 - a completely defined analytical expression relating wear to design

parameters and application factors; 2 - a data base for material wear

coefficients involved in that relationship. A frequently desired

attribute for the analytical expression is that it be simple. However,

with the advances in computational and evaluation capability, e.g., FEM

and CAD techniques, this is probably not a necessary requirement.

WEAR THEORY AND MODELS

To the designer, the ideal situation would be to be able to go

directly to some reference source, e.g. data base, book, or expert

system, and find the engineering model that is applicable to the

specific situation, along with relevant data for different materials.

The next best situation is that the designer can contact tribologists,

who in turn can identify the needed model, or perhaps reduce a more

general model to this specific case, and identify the associated data

base. Ideally this type of support by the tribologist is

provided analytically with, perhaps, some short, simple wear tests to

determine wear coefficients for specific materials. Consider what

either the designer or the tribologist finds when this is attempted.

In the wear literature and information, some engineering types of

relationships can be found. For the more general of these there is

usually no consensus on their validity or range of applicability. For

the more specific models the range of applicability appears limited and

usually does not match the situation of interest. In addition the data

bases associated with these models are frequently limited. Beyond the

specific information related to engineering models, there is

considerable information available regarding wear phenomena and wear

mechanisms. Much of that information is related to situations which

are not directly relatable to an application for a variety of reasons.

For example, the conditions of the wear test associated with the

information could be significantly different than the condition in the

application. There might be insufficient details in the paper to make

use of the data. The information might be qualitative or

phenomenologically oriented, or the quantitative data taken may be

insufficient for extrapolation. Further, the range of applicability of

the information is frequently not addressed nor is the relevance of

that information with respect to other phenomena. Physical models for

various mechanisms or modes of wear can also be found. Some are

descriptive and some are characterized by analytical relationships.

However, their relationships to design parameters are usually not

developed. Also, many of the more recent articles specifically point

out that wear is a system property; that there are many mechanisms by

which materials wear; that there are transitions in wear behavior; that

the individual wear mechanisms can exist in parallel and interact in a

sequential, as well as in a parallel, manner.

This complex, confusing (to the non-tribologist), and often

incomplete condition of wear knowledge typically results in the

tribologists performing a study, which involves testing, to develop or

Downloaded/printed by

University of Washington (University of Washington) pursuant to License Agreement. No further reproductions authorized.

Thư viện tri thức trực tuyến

Nội dung xem thử

Mô tả chi tiết

Tài liệu tương tự (6)

Astm stp 837 1984

Astm stp 15c 1962

Astm stp 531 1973

Astm stp 642 1978

Astm stp 849 1984

Astm stp 1601 2017