ORNL-TM-2727.txt

 

 

          
 

 
  

LIBRARY
~ENTRAL RESEARCH |
~ " DOCUMENT COLLECTION

S 76
OAK RIDGE NATIONAL LABORATORY

operated by

UNION CARBIDE CORPORATION
for the

U.5. ATOMIC ENERGY COMMISSION

ERGY RESEARCH \LIBRARIES

(VEABA

3 445k 0513

"fa

ORNL- TM- 2727

THE MECHANICAL BEHAVIOR OF ARTIFICIAL GRAPHITES

- £3

AS PORTRAYED BY UNIAXIAL TESTS

W. L. Greenstreet
J. E. Smith
G. T. Yahr
R. S. Valachovic

“

 

NOTICE This document contains information of a preliminary nature
and was prepared primarily for internal use at the Ook Ridge National
Laboratory. It is subject to revision or correction and therefore does
not represent a final report.

 
 

 

 

 

 

LEGAL NOTICE

This report was prepared as an account of Government spfinsored work. Neither the United States,

nor the Commission, nor any person octing on behalf of the Commission:’

A, Maokes any warranty or representotion, expressed or implied, with respect to the accuracy,
completeness, or usefulness of the informotion contained in this report, or that the use of
any information, up.poruru's, method, or process disclosed in this report mey not infringe
privately owned rights; or . .

B. Assumes any liabilities with respect to the use of, or for damoges resulting from the use of
ony information, apparatus, method, or process disclosed in this report.

As used in the above, "‘person acting on behalf of the Commission'’ includes any employee or

contractor of the Commission, or employee of such contractor, to the extent that such employee

or contractor of the Commission, or employee of such contractor prepares, disseminotes, or
provides access to, any information pursuont to his employment or contract with the Commission,

or his employment with such contractor,

 

 

 

i
ORNL-TM-2727

Contract No. W-T4O5-eng-26

Reactor Division

THE MECHANICAL BEHAVIOR OF ARTIFICIAL GRAPHITES
AS PORTRAYED BY UNIAXTAL TESTS

. Greenstreet
. Smith

. Yahr

. Valachovic

Ty =
e

DECEMBER 1969

OAK RIDGE NATIONAL LABORATORY
Ozk Ridge, Tennessee
operated by
- UNION CARBIDE CORPORATION
for the
U.S. ATOMIC ENERGY COMMISSION

      
 

  

AIES
LOCKHEED MARTIN ENERGY AESEARCH LIBRA

(R

2 b
\ 3 yusk 0533802 &

T
 
i1l

}Contents
Page
N o T = Y o 1
TNETOAUCELON et s e oo vn st oe s ne e snneeneneseneensnnseneennsnnns 1
Mechanical Behavior ...veiiisiineiirersesscnnersesnsennans [ 2
Synopsis Of Reported RESULES «evrrerrennereerennneeenns 2
Additional Investigations ............. ettt 14
Conclusions .....ev.... e ettt et 41
ACKNOWI1EAEEMENTS « vt et ennerrsenneesosecnnassssennesesennnanas Lo

RETFETEIICESE « v v et e aen s st aneansensesassnssenesnnssnsenennns L3
THE MECHANICAL BEHAVIOR OF ARTIFICTAL GRAPHITES
AS PORTRAYED BY UNIAXTAL TESTS

W. L. Greenstreet
J. B. Smith
G. T. Yahr

R

. S. Valachovic

Abstract

Stress-strain behaviors which are representative of nuclear-
grade, or equivalent, graphites are described and discussed.
oince c¢cyclic loading test results provide important insight into
the characteristics of graphite behavior, emphasis is placed on
results of this type. Monotonic loading curves are also consid-
.ered, and both stress versus longitudinal strain and stress
versus lateral strain curves are given for monotonic and cyclic
loading. The new results given in this report are combined with
results published in the literature to provide a unified descrip-
“tion of the observed behavior.

Keywords: graphite, materlals testlng, mechanical prop-
erties, compression, tensile properties, stresses, deformation,
heat treatment. -

Introduction

' Many of the mechanical‘behavior charaoteristics of graphite have been
studied. This is espeeielly true of'nuclear—grade, or equiyalent,_grsph—
ites es a class of materials. These graphites are generally made from
petroleum-coke and coal-tar pitch, but the same overali characteristics
are exhibited by graphites made from other binder and filler_combinations |
as Well It is the purpose of this report to bring out some of the salient
_features of observed mechanlcal behav1or in & unified fashlon

Important behavioral characteristics are descrlbed in the literature.
These have been 1dent1f1ed from monotonic and cyclic loadlng tests, but
there are certain aspects of cycllc loadlng behav1or that elther have not
been studied or have not been studied in detail. Therefore, addlt;onal
investigations were made to overoome this deficieney. Three types of cyclic
loading were considered, including (1) cycling of specimens, which were
preloaded in compression, between zero and a constant stress level less

than the preload stress value, (2) cycling in compression wherein the
specimens were only partially unloaded, and (3) cycling alternately in
campression and tension. Studies of the first two types have not been
reported previously, and studies of the third were used to experimentally
examine the validity of an often stated premise regarding reloading be-
havior. The results from these investigations when combined with avail-
able data provide important background information for constitutive equa-

tion development.

Mechanical Behavior

 

Despite material variations from grade to grade, from block to block
of a given grade, and from piece to piece within a given block, there are
definite characteristics of mechanical behavior associated with artificial
graphite as a class of materiéls. Most graphites are anisotropic, and
stress-strain data generally show that there is rotational symmetry of the
anisotropy in an element. That is, the material may be classified as
transversely isofropic. |

The anisotropy is a result of the preferred orientation of the coke.
particles used in the manufacture and the orientations of the‘cryétallites
within the particles. Although the forming method, that is, molding or
extruding, governs the preferred orientations of the particles, this
method influences the character of the anisotropy only in terms of actual
measures Of deformation resistance. Therefore, the characteristics to be

discussed are independent of forming method.

Synopsis of Reported Results*

 

It is common knowledge that, when a graphite specimen is loaded in
simple tension or compression, nonlinear stress-strain behavior is exhib-
ited from essentially zero stress to the failure stress. What is the
characteristic of this nonlinearity? By loading to a stress level less
than that at failure and unloading, one obtains a loading curve, OA, and

an unloading curve, AB, as shown schematically in Fig. 1. Since the

 

*Information summarized here as well as summaries of other informa-
. . . . - 1 ‘
tion on graphite are given in a survey report.
ORNL IWNG. 69-L105

STRESS

 

 

c B STRAIN

Fig. 1. Schematic Drawing of a Stress-Strain Diagram for Graphite.
unloading curve does not retrace any part of the initial loading curvé,
the behavior cannot be classified as nonlinear elastic. The nonlinearity
must be described in some other way.

It may be seen that the behavior exhibited is reminiscent of time-
independent elastic-plastic behavior ascribed to metals. When the speci-
men is fully unloaded, there is a residual strain, OB, as for materials
that undergo elastic and essentially time-independent plastic deformations.
The unloading curve, AB, is also nonlinear. On loading the specimen a
second time, the resultifig-curve is again nonlinear and a hystéresis loop
is formed between the unloading and reloading cufves.' Loading beyond the
stress corresponding to that at the unloading point, A, gives a curve
which becomes asymptotic to the cuffie that would have existed. had unload-
ing never occurred.¥ There are no abrupt changeé in the slopes of any of
the segments of the entire curve.

The tensile strengths and the initial slopes of the stress-strain
curves (elastic moduli) are generally greater in the with-grain direction
than in the across-grain direction. There are pronounced differences be-
tween both stress and strain at fracture in tension and in compression:
Complete stress-strain diagrams for simple tension and compression are
plotted together in Fig. 2 (Ref. 2) to.illustrate these differences in the
tensile and compressive behavior. The material is EGCR-type AGOT graph-
ite,f and the data are for the with-grain (parallel tb the extrusion axis)
direction. Typically, fracture strainé on the order of 0.1 to 0.2% and
1.0 to 2.0% are found in tension and compfession, respectively, for nuc-
lear-grade, or equivalent, graphites.

Arfagon and Berthier® performed compression test étudies on 216 spec-
imens made from extruded, petroleum-coke, industrial graphite. Three types
of tests were used: (1) simple compression; (2) cyclic tests (loading-
unloading-reloading) in which the cycles were made from increasing Stfess

levels spaced at equal intervals; and (5) cyclic tests befween zero stress

 

*This is an often stated premise which is based on many observations.
A direct experimental examination of this premise will be discussed later
in this report.

TEGCR—type AGOT graphite is a coarse-grained, nucleaf-grade, extruded
graphite which was made by Carbon Products Division of Union Carbide Corp.

 
ORNL DWG. 64-11422

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2000
STRAIN (%) ‘
-1.4 -1.2 -1.0 -0.8 -0.6 -0.4 -0.2 TENSl}E
0.2
COMPRESSIVE
ra -2000 =
2
/ o}
w
. w
/ @
ot -4000 »
— -
-6000
Fig. 2. Complete Uniaxial Stress-Strain Curves Parallel to Extrusion

Direction.
and a fixed maximum. The slopes of the stress-strain curves at the origin
for continued loading (type 1) were the same order of magnitude as given
by sonic measurements. Seldin® also found good comparisons between sonic
moduli and the slopes at zero stress as measured from tensile and compres-
sive stress-strain curves for several grades of molded graphite.

Arragon and Berthier found that cyclic compressive tests of type 2
cause the apparent density to increase. The hysteresis loops, which are
formed by loading, unloading, and reloading, increase in size with in-
creased maximum stress, and the slopes of straight lines connecting the
unloading and reloading points decrease with increased stress. (The slope
of a line connecting the two points of one cycle is termed the "paraelas-
tic modulus.") | |

A schematic diagram, which depicts the essential features of the be-
havior observed from a test of fype 2, is shown in Fig. 3. The envelope
curve corresponds to that for simple compression. Arragon and Berthier3
also discovered that straight lines drawn through the unlcading fioint and
the point of zero stress for each ¢ycle converge at a single point as
shown in the figure. The coordinates of this point are both negative
(taking compression as pog}tive), and it was reasoned that the existence
of this point is a manifestation of the history of the virgin specimen.

In each case, the‘reloading curves asymptotically approach the enve-
lope curve after each cycle of loading, unloading, and reloading. This
and the other details of the behavior, as described above, are_genefally
typical of the graphites being considered. Corroborations can be found by
studying the results for extruded graphite reported by Losty and Orchard®
and for molded graphites by Seldin.®

In the case of test type 3, which was used in the study by Arragon

3 each specimen was subjected to 12.cycles. During the first

and Berthier,
cycles, the total deformations at the unlocading and reloading points in-
creased with increased cycle number, but, after the sixth cycle, these
deformations were essentially constant, afid the hysteresis loop was re-
traced on each subsequent cycle. The form of the hysteresis lcop remained
constant throughout the cyclic loading as did the paraelastic modulus.

There was a slight increase in apparent density during cycling.
ORNL DWG. 67-2981

STRESS

 

 

 

STRAIN

Fig. 3. Schematic of Compressive Stress-Strain Diagram.
8

A schematic diagram of the type behavior observed during the first
few cycles is shown in Fig. 4. Additional illustrations of this behavior,
as obtained by the present study, are given in Figs. 5 and 6. The latter
figures show results obtained from EGCR-type AGOT and from RVD¥ graphite,
respectively.

The initial slopes of the stress-strain curves, or Young's moduli,
for nuclear-grade, or equivalent, graphites are about the same in tension
and in compression. This is generally true for both the with-grain direc-
tion and the across-grain direction. However, exceptions may be fofindf%y
comparing across-grain data for some graphites.  Close comparisons of
tensile and.compressive curves for a given direction also reveal that
there is a tendency toward gfeater deformation resistance in bompreséion
than in tension. This especially is true for the across-grain direction.

The preceding discussion was limited to stress versus longitudinal
strain behavior. ©Stress versus lateral strain curves for these graphites
have different curvatures in tension and compression; the diagrams for
tension are concave,toward the stress axis, while those for compression

are convex. These observations were first reported by Seldin.*

Schematic
diagrams of the tensile and compressive behaviors are shown in Fig. 7.

In uniaxial compression tests, the transverse—tOalongitudihal strain
ratios are essentially independent of stress, yielding almost conétant
values. However, the ratios in tension are functions of stress, decreasing
as the stress is increased. Figure 8, which shows the strain ratios for
EGCR-type AGOT graphite as functions of longitudinal strain,2 provides a
clear illustration. In this figure, the subscript 3 refers to the paral-
lel, or with-grain direction, while the subscripts 1 and 2 refer to two
orthogonal (across-grain) directions in the plane of isotropy. (Trans-
verse isotropy is assumed.) The first of the double subscripts indicates
the difection of applied stress and the second indicates the direction of

induced strain. The vertical bars in the figure represent standard devia-

tions.

 

¥RVD is an extruded graphite manufactured by Carbon Products Division
of Union Carbide Corp.
CRNL IWG. 67-2980

STRESS

 

 

STRAIN

Fig. 4. Schematic Diagram Showing Cyclic Behavior.
10

QRRL IMG. 69-L106

 

4000

3500 //////

I

 

 

3000

 

N
wn
O
o

 

 

 

 

STRESS ( pst)
N
o
O
O

1 500

 

 

1000

 

 

 

500

 

 

 

 

 

 

 

 

0 o1 02 03 04 0.5 0.6
LONGITUDINAL STRAiN,EA(%)

Fig. 5. Cyclic Stress-Strain Curves for a With-Grain EGCR-Type AGOT
Specimen (12 Cycles).
11

ORNL TMC. 69-k104

e

8000 7

 

9000

 

 

7000

 

6000

 

 

5000

STRESS (psi)

 

 

 

4000

 

 

3000

2000 //
1 000

0 0.2 0.4 0.6 0.8 1.0 12
LONGITUDINAL STRAIN.EA(%)

 

 

 

 

 

 

 

 

 

 

 

Fig. 6. Cyclic Stress-Strain Curves for a With-Grain RVD Specimen

(S-Cycles).
12

- ORNL DWG. 67-3LT7

 

 

 

 

 

 

 

 

 

 

 

TENSION
4

4 3+
n
e

- a2
w
, &
%

— l —

02 0 0 o A 2 3 4
LATERAL € LONGITUDINAL €
STRAIN (PER CENT)

COMPRESSION
‘ E‘ —"-"

e
7
0
w
x
-
"

| Lt L)l

oz . ol 4 8 12 16 20
LATERAL € LONGITUDINAL €

STRAIN (PER CENT)

Fig. 7. ©Schematic Drawings of Longitudinal and Lateral Stress-Strain
Curves.
STRAIN RATIO

0.20

0.15

0.10

0.05

0.15

0.10

0.05

0.15

0.0

0.05 ]

13

ORNL-DWG 64-10983

 

 

 

T

]

 

 

 

 

 

 

 

 

ITT

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

L

o T -
J l ==
|
K ]y. + il
{ 1 ‘
2 ‘| 2 Hi2 v F21
|
H13 . 123
- ' s 1
[ K43, 123

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

LONGITUDINAL STRAIN IN COMPRESSION (%)

Fig. 8.

Il1 7 = _ I
R HH
| | HHHH
| _—
1 | 1 ’ B30, H32 K31, 132
0.5 1.0 1.5 20 0 0.05 0.10 0.4%

LONGITUDINAL STRAIN IN TENSION (%)

Strain-Ratio Curves from EGCR—Type AGOT Specimens.
1L

When a specimen is loaded and released, the transverse residual
strain is positive regardless of whether the load is a tensile or a com-
pressive one. Thus, the volume of a specimen pulled in tension and re-

leased is increased since all linear dimensions are increased.

Additional Investigations

 

The investigations reported in this section were made using speci-
mens with the design shown in Fig. 9. This particular design was selected
so that the specimens could be tested in either tension or compression.
The 0.100-in.-diam longitudinal hole was bored in each specimen to allow
for making permeability measurements, and a smooth machine finish was
used, without grinding or polishing.

Each specimen was marked for gage location using a HB grade drafting
pencil to avoid scratching the surface and to provide indications thatr
would not be removed by heat treating the specimen af 3000°C (to be ais-
cussed later). The latter aspect is important because the specimens were
heat treated and reinstrumented‘fwice in ‘some cases, and it was necessary
to mount the gages in the same location each time. They were instrumented
on opposite sides at the midpoint of the gagé length with strain gages
oriented in the axial and circumferential directions. Budd Metalfilm,
type C6-121-A, strain gages with a 0.125-in. gage length were used.

| Tensile tests were pérformed using steel clam-shell,fixturés attached
to the ends and 0.20-in.-diam, stainless steel, stranded-wire cables for
transmitting the. load to the specimen. A specimen, ready for test, is
shown in Fig. 10. The campressive tests were carried out using a subpress
with a miniature load celi placed in the subpress below the specimen. By
placing the load cell at this location, inaccuracies in readings due to
plunger friction were eliminated. The test setup for compression testing
is shown in Fig. 11.

For those cases in which the specimens were heat treated, a 100 kw,
3000 Hertz induction furnace was used for this purpose. The inside dimen-
sions were a diameter of 9 in. and a length of 14 in., and the specimens
were heated in Argon at atmospheric pfessure. The time to reach the 3000°C

temperature level was 45 minutes. The furnace was cooled from 3000 to
15

ORNL-DWG 69-5079

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

3.50 in.
0.50 .
) e 2.500 i
<:::1{32—1'n. R. (both ends)
j
Lo.10
L Dia.
0.625-in. Dia. ’
2-in. R. j;;”' R. L—1.125-1n. Dia.
1.250 in. -
_flZB.szs in.

 

Fig. 9. Drawing of Test Specimen.
 

 

16

PHOTO TE4ET

 

Fig. 10. Specimen Prepared for Tensile Testing.
 

bd
=
o
3
0
L
B
@
=
o
;n
]
Q
i
=
B
o
o
&
O
B
i)
v
B
o
oy
)]
~
i
=
u
B
-
o
)]
2y
o

1.

 
18

2500°C in approximately 15 minutes, from 2500 to 2000°C in about 1 hour,
and from 2000°C to room temperature in 12 to 14 hours.

In the first series of tests, specimens from EGCR-type AGOT graph-
ite were subjected to the following program of loading. Each specimen
was loaded to a given stress level in compression and unloaded. It was
then cycled ten times between zero and a stress, 00,'which was in the

range given by
1
5 0 < o, < o,

where O is the maximum stress level, that is, the stress corresponding.
to the first unloading point. In each case, the specimen was subsequently
loaded to failure. Two with-grain and two across-grain specimens were
tested, and both lateral, or transverse, and longitudinal, or axial,
strain versus stress diagrams were obtained. Compressive behavior was
chosen because relatively large stresses and strains can be induced with-
out failure.

Figure 12 shows the results for the with-grain direction. Here again
it may be seen that, when the specimens were loaded to failure after
being cycled, the stress-strain curves asymptotically approached exten-
sions to the initial loading curves that would have been obtained by con-
tinued loading beyond the unloading stress, o - The hysteresis loops
exhibited by the stress-strain curves do not change in size and coincide
rather than change position along the strain axis during the first few
cycles, as was shown in Fig. 4.

The cfirves for the across-grain specimens are shown in Fig. 13.

- Except for very slight deviations in the initial portions of the first
few reloading curves; the hysteresis loops are again retraced on each
cycle. Asymptotic approach to a monotonic loading curve upon loading to
failure may also be seen. |

Partiasl unloading characteristics were examined using a second set
of four with-grain and four across-grain EGCR-type AGOT graphite speci-
mens. A specimen was loaded in compression to a given stress level, O
and unloaded. It was again loaded to a higher compressive stress level,

O, unloaded, reloaded to o, & second time, and unloaded. A companion
19

ORNL DWG. 69-2977

 

 

3500 '
/
// _—

| /fl/ 4
A/ //
ol S /]
/. 1Y
ol [y

 

 

 

n
14d
Q
o

 

 

n
o
o
o

 

STRESS (pm)

 

 

 

 

1000

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0 02 04 06 08 (O 12 0 -002 -004 -006 -008 -0.0
LONGITUDINAL STRAIN, €, (%) , LATERAL STRAIN,E (%)
4000 —

 

 

 

 

| i
Wi

/| /,

/// -/

 

 

N
o
O
o

 

 

STRESS (p=m1)
3

 

 

 

 

 

 

oF e

 

 

 

 

 

 

 

 

 

 

 

 

 

1500

1000

. 500
0

0

 

 

Q4 06 0.8 1.0 1.2 0O -002 -004 -006 -008 -0.0
LONGITUDINAL STRAIN, €, (%) LATERAL STRAIN,CI_(%)

Fig. 12. Cyclic Curves for Two Preloaded With-Grain EGCR-Type AGOT
Specimens. 4
20

ORNI, DWG. 69-8209

 

 

sor_ > -
N 711 L7
o Y/ 7

00 /,//- - B // //
- [/ ' [ p
Y LY
f 4

 

 

 

 

 

 

 

 

STRESS (psi1)

 

 

 

 

500

 

 

 

 

 

 

 

TT TT

 

 

 

 

 

 

 

 

 

 

I.2. 1.4 0 -O.dZ -004 -Q06 -008 -010 -0I2

0 02 o LATERAL STRAIN, € (%)

06 08 1.0
LONGITUDINAL STRAIN.E, (%)

 

 

 

 

3500

 

 

g

 

 

2
Q

7
/I
7
1/ /1)
/

o 0.2 0.4 0.6 0.8 1.0 1.2 1.4 0 -002 -004 -006 -008 -010 -0J2
LONGITUDINAL STRESS €,(%) LATERAL STRAIN,€ [ (%)

 

 

8
o

STRESS {ps!)

 

 

1500

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

ar

 

 

‘ FPig. 13. Cyclic Curves for Two Preloaded Across-Grain EGCR-Type AGOT
Specimens. '
21

specimen was then loaded in compression to O partially unloaded, re-
loaded to the higher compressive stress level, O partially unloaded,
reloaded to on,-and unloaded. | ‘ |

The stress-strain diagrams for two companion with-grain specimens
are shown in Fig. 14. The diagrams for a second set of with-grain speci-
mens are shown in Fig. 15, and Figs. 16 and 17 show the results for sets
of across-grain specimens. |

The cyclic behav1ors of the fully unloaded speclmens follow the pat—
terns described earller An examination of the stress versus longltudlnal
strain curves‘revealS'several features. The hysteresis loops increase in
size and the paraelastlc moduli decrease w1th increase in unloadlng stress
level the total deformatlons increase upon loading a second tlme to the
higher compressive stress level, o n’ and the reloading curves became
asymptotic to extensions of the initial loading curves when the specimens
are loaded beyond O The stress versus lateral strain curves are very
similar in character to the stress versus longitudinal strain curves.

In the case of partial unloading, hysteresis loops are again formed,
and the same general features that were observed for full unloading may
be seen. The hysteresis loops are smaller, but thé segments of each curve
are entirely nonlinear. The asymptotic approach to the extension of the
initial loadlng curve in each case corresponds to that for full unloading.

Since the detalls of the asymptotic approach feature are important
to the development of mathematlcal analogs to the materlal behavior, the
qurves in Figs. 14 through 17 were carefully studied to provide guanti-
tative data. When a specimen is loaded to a stress level such as g m’
unloaded either fully or partlally, and reloaded, the relcading curve
undergoes an accelerated change in slope as the.stress approaches the
stress, © ; at the point of unloading. This accelerated change persists
until the reloadlng curve essentially begins to trace what would have been
the extension to the initial loading curve provided unloading had not
occurred. The quantitative data obtalned were the stress values at the
extréemities of the accelerated change reglon The lower llmlt is denoted
by o/ and the upper limit is referred to as o”.

The values for o/ and ¢’/ were obtained by plotting the ratio of

longitudinal strain to stress as a function of stress. The data on either

a
22

ORNL IWG. 69-2981R

 

 

3500

 

 

§

 

 

 

 

8
\%
=~

 

 

 

 

i /
E Ay L
71/ /

Y
L
)

o 01 02 ©03 04 03 06 - 0 -002 -004 -G06 -008 -0.0 -QI2
LONGITUDINAL STRAIN, €, (%) o LATERAL STRAIN, €_ (%}

e

 

 

 

10/

 

 

 

 

 

 

 

 

 

 

 

 

 

 

10/

 

 

(a) Full Unloading

 

 

 

 

 

 

 

 

2000

VI AT ] il
/ 1]

§
N

N\
BN

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

500 A :
o / / okt / ‘ 1010
o 0.1 02 03 ©04 05 08 . : 0 -002 -004 -006 -008 -QI0 =012

LONGITUDINAL STRAIN, €, (%) LATERAL STRAIN, € (%)

(b) Partial Unloading

Fig. 1. Cyclic Curves for Fully and Partially Unloaded With-Grain
EGCR-Type AGOT Specimens. '
23

ORNL DWG. 69-2982R

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

3500
3000
2500
é [
@ 2000 A
; / i
® 1500 /////// // 1
1000 //// e //fl '
500 [—
Nd/4 o Il ,,..
0.1 02 03 04 05 06 0 -002 -0D4 -006 -Q08 -0 ~-0.12 -0i4
LONGITUDINAL STRAIN, €, (%) . LATERAL STRAIN, €; (%)
(a) Full Unloading
1500
3000
_ 2500
: 7 /
w» 2000 /
@ I
: ) 1
@ 1500 / : ‘!
1000 // : // ;
500 /
0 1028 i 08¢
6 o1 02 03 04 085 06 0 -002 -004 -Q06 -008 -OI1 -042 -Ol4
Lor‘:GlTUDINAL STRAIN, €, (%) LATERAL STRAIN, € (%)
(v) Partial Unloading
Fig. 15. Cyclic Curves for Fully and Partially Unloaded With-Grain

EGCR-Type AGOT Specimens.
ORNL DWG. 69-2983R

 

 

3500

 

 

3000

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

i L7 /
L 7/
. s /1y
T T )
LovarruomaL Seam. €, () Larcan, sTRan, €, ()
(a) Full Unloading

 

 

 

 

 

 

 

 

 

 

f 2000 71
E 1500 /77 é/ /4 /
T Y
-~ Ay

<]

" 1 Y 1.

0 0.1 02 O3 04 05 06 07 08B 09 1.0 0 -002 -004. -006 -008 -0.10
LONGITUDINAL STRAIN, €, (%) LATERAL STRAIN, €, (%)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

(v) Partial Unloading

Fig. 16. Cyclic Curves for Fully and Partially Unloaded Across-Grain
EGCR-Type AGOT Specimens. ‘ -
S TRESS {(pn)

STRESS (ps:)

o5

 

 

 

 

 

 

 

 

3500
3000
2500 p?
2000 ////;f”,/’/y'
1500 1 / / /
vd CNANA
1000 v A /'/
| A
500 A /A///
N
104
oo 21 - o2 c3 pa ©05 06 07 08 09 10
LONGITUDINAL STRAIN, €, {%)
(a) Full Unloading
3500
3000
2500
]
2000 //
1500 1/ 4
AN
1000 L /'
300 '/ /// :
104e
oo ol 02 03 04 05 06 07 0.8 Q9 1.0
LONGITUDINAL STRAIN, €, (%)
(v) Partial Unloading
Fig. 17.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

EGCR-Type AGOT Specimens.

ORNL DWG. 69-2980R

 

 

 

 

 

el
1/
"/

.7

 

 

 

 

 

 

 

 

 

1//

0 -a0z2 -Q04

 

-006 -008 -0

LATERAL STRAIN, €. (%)

 

 

 

T

 

/

 

/

va

 

0 -Q02 -004 -008 -0.08 -O! -or
LATERAL STRAIN, €, (%)

/

 

 

 

 

 

 

 

 

1044

 

Cyclic Curves for Fully and Partially Unloaded Across-Grain
- 26

side of O are described by straight lines with greatly different slopes.
The transition region then ccrresponds to the region where accelerated
change in slope of the stress-strain diagram occurs, and the limits may
be determined. A typical example of the plots used for determining o’
and o// is shown in Fig. 18.

The two stress levels, denoted by o/ and o’/, and the value of o,
for each stress versus longitudinal strain curve are listed in Table 1.
Also included are the values for the ratios 0’/Gm and o’fl/om_.

The dots on the stress~strain diagrams of Figs. 14 through 17 mark
the limits of the transition regions. It may be seen from the table and
the figures that there is consistency between the values. The value of
0, is the same for all cases. In five of the eight cases, the 0’/0m ra-
tios are the same, and all ratios fall within the range from 0.93 to 0.97.
The upperlextremities of the segments for full unloading are at higher
stress levels than those for partial unloading, which shows that asymp-
totic approach is achieved sooner in the latter case. For full unloading,
the OJ//Gm . ratio ranged from 1.04 to 1.07, while this ratio ranged from
1.01 to 1.03 for partial unloading.

The final test series reported herein was used in an attempt to ex-

perimentally verify the asymptotic approach premise and to provide data

Table 1. Data Obtained From Curves for
/ Partial Unloading Studies

 

. /
Specimen Om o o/’

 

. ‘ / r7
Number Orientation (psi) (psi) (psi) ¢ /Gm o’ o,
101 WG 2000 1850 2130 0.9% 1.07
101a WG 2000 1950 2040 0.97 1.02
102 WG 2000 1900 2130 0.95 1.07
1023 WG 2000 1900 2050 0.95 1.0%
103 AG 2000 1850 2080 0.93 1.0k
103a AG 2000 1900 2020 0.95 ° 1.01
104 AG 2000 1900 2100 0.95 1.05

10La AG 2000 1900 2025 0.95 1.01

 
27

ORNL DMG. 69-10965

2.8 _ : - —

 

 

2.6

 

2.4 : : ‘

(10 &Y 1b)
\

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2.2 /
z ¢
< 9
o
'._.
w| v
|8 7
ZE20 .
S| o«
= o
~
T s —?f//

o/
c/
2
1.6
0 500 1000 1500 o' 0,0 2500 3000
STRESS (psi1)
Fig. 18. Plot for Determination of the Extremities of the Accelerated

Change in Slope Region for a Fully Unloaded Across-Grain EGCR-Type AGOT

Specimen.
28

on cyclic behavior under alternate compressive and tensile loadings.
Studies on the validity of the asymptotic approach premise were motivated
by Seldin's? demonstrations that test specimens can be made to reproduce
their original stress-strain responses in the longitudinal and lateral
directions by heat treating them at their graphitization temperatures.
Provided it is generally true, this is a revolutionary discovery in graph-
ite testing, especially in the potential it provides for removing uncer-
~tainties in data analyses and interpretations that arise because of the
almost inevitable variations in graphite properties.

Pursuant to making the final studies regarding the cyclic behavior
of graphite{ Seldin's method for removing deformation history was examined
using specimens made from ATJ, RVD, and EGCR-type AGOT graphites. The
maximum particle sizes are 0.006, 0.015, and 0.032 in., respectively.
The first two are molded graphites. Each specimen was subjected to the
following sequence of steps.

1. Heat treat at 3000°C for 60 minutes.

2. Load in tension to a predetermined stress level and unload.

3. Heat treat at 3000°C for 60 minutes.

L. Load in compression to a stress level greater than in Step 2 and
unload.

5. Heat treat at 3000°C for 60 minutes.

6. Load in tension to the stress level of Step 2 and unload.
The first heat treatment was used to establish the graphitization tempera-
ture in each case. |

Stress-strain curves for two with-grain specimens of ATJ graphite are
plotted in Fig. 19, where both the longitudinal and lateral strains are
shown. The curves for two across-grain ATJ specimens are plotted in Fig.
20. All stresses, tensile and compressive, and the corresponding longi-
tudinal strains are plotted as positive in all figures depicting results
from the heat treatment studies, and the transverse data are plotted in
an analogous manner. Note that the two strain scales for a given speci-
men are not the same; hence, the differences between lateral strain curves
are accentuated. Curve No. 1 is for the first loading in tension, No. 2
is for the compressive loading, and No. 3 is for the second loading in

tension.
29

ORNL DWG. 69-4103

3200
e : : ‘
2800 : // ’ : /
2400 ! ’ . . ’/,3 A /
L g L/

 

 

 

 

 

 

 

 

 

 

STRESS '(“pnl
g
'l
N
§
N
N

 

o
o)
o

 

 

AN /1/
A /711 | 1V

 

 

 

20!

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

400
201 /

°
° Qaos QI0 0I5 02 025 03 Q35 0.4 043 001 0 -00I -002 -Q03 -004 -005 -006 -Q07
LONGITUDINAL STRAIN, €A (%) LATERAL STRAIN, ET('/n)

- z /
T | IRy AV,
/// | IS
A/ M
s00 A/ | / /i/
470 /)

400

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

STRESS

 

~

 

 

1200

 

 

 

 

202

 

 

 

 

 

 

202

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

045 ‘ ool C -001 -+002 -003 -004 -005 -006 -QO7

o 005 0.1 005 02 025 030 035 040
LATERAL STRAIN, €; (%)

LONGITUDINAL STRAIN,E, (%)

Fig., 19. ©Stress-Strain Curves for Two With-Grain ATJ Specimens.
30

ORNL DWG. 69-2975

 

 

3200

 

 

 

 

2400
= -
. -7
" 7y /

 

 

 

 

STRESS {pa+ )
g
N

 

 

 

 

v

-

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

,/
«00 / — - //

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

oy
o
o Q0% at s 0.2 25 03 03% 04 0ad 003 002 00 0 -001 -002 -0035 -D0O4 -005 -006 -
LONGITUDINAL STRAIN , €, (%)} LATERAL STRAIN.€ (m} 004 -005 -006 -QO07
. T
3200 -
2800 /
2400 Ve ' .‘Z
- "2
- /)
22000 Y Pt
3 Y
a
W
EIGDO ’A /
1200 / 4 M/
800 ,/ /// /
a4 J
400 // / //
. ros // .
o 005 01 Q5 02 02% 03 Q3% 04 045 003 002 00 0 -001 -002 -003 -004 -GO5 -006 -007

LONGITUDINAL STRAIN, €, (W LATERAL STRAIN, €4 (%)

Fig. 20. ©Stress-Strain Curves for Two Across-Grain ATJ Specimens.
51

The success of the heat treatment method for removing deformation
history is determined by comparing curves 1 and 3 for each specimen.
Since the second with-grain specimen failed during the second tensile
loadlng, comparisons are more difficult to make for the curves shown in
the lower part of Fig. 19 Figures 19 and 20 show that the hlstory was
almost entirely removed by heat treatment. Although there is a small
decrease in deformatlon resistance between the first and second tens1le
loading, the agreement between the curves is very good. ans is true
for both the with-grain and across-grain directions.

Some of the.beheyiora;flaSPects which were discussed in tne previous
section may be seen from these curves. The'deformation resistances are
essentially the same in tension'and in compression, with a very slight
_tendency toward greater deformation res1stance in compression than in
tension for the across- grain direction. The durvatures of the stress
versus lateral strain curves in tension are different from those in com-
pression, but the initial slopes are nearly equal. Finally, positive
residual lateral strains are incurred as a result of loading in tension.

The stress-strain curves shown in Figs. 21 and 22 are for two with-
grain RVD specimens and one across-grain specimen. Curves 1 and 3 for
the with-grain specimen are in good agreement. Again, slightly less
deformatlon resistance is found for the second loading in tension.

Although the differences between curves 1 and 3 are greater for the
across-grain specimen than in the above cases, the‘agreenent is gocd.
Greater deformation resistance in compression than in tension may also
be seen in Fig. 22. | | " |

Results obtained from EGCR;type AGOT specimens are given in;Figs.
23 and 24. The curves for the with-grain ‘specimen again show that the
heat treatment was. réasonably effective. It was less:effective for the
across-grain specimens, however, as may be,seen by examining Fig. 24.
The second across- -grain specnnen was not loaded to the same stress level
in tension during the first and third loadings, but thls does not detract
significantly from the comparisons. Again, the across-grain specimens
exhibit greater deformation resistance in compression than in tension.

The above results for the three graphites show that, in general,

heat treating does almost entirely remove the prior deformation history
 

3200

2

 

2800

 

2400

 

§2000

 

& 1600

 

1200

 

800

/

 

400

 

/

 

 

 

 

 

 

 

 

 

o/

 

o

0

3200

2800

STRESS (pm)
nN n
g 8

3
©
o

g

800

400

o

003 0.

015 02 025 0.3
LONGITUDINAL STRAIN, €, (%)

033 04

0.43

 

7

 

7

=3

 

 

I

 

 

)
/|

 

 

 

 

/

=
NN

 

 

 

 

 

 

 

 

Jor

 

o

0.03

Fig. 21.

0.1

045 02 025 0.3
LONGITUDINAL STRAIN €, (%}

0.3 04

0.48

Stress-Strain Curves

ORNL IWG. 69-8210

 

 

AL
//,-, }
N

/)
1Ly
///

-0.03 -0.04 -00% -006

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

A

004 003 002 OO0l 0 -0.01 -2.&
LATERAL STRAIN, €, (W

 

 

 

 

 

 

frr
/8
1

/L)

A
0.04 003 002 00/ O -00 -002 -003 -0.04 -0.03 -0.6
LATERAL STRAIN, € (%)

 

 

 

302

 

 

 

 

 

 

 

 

 

 

 

 

for Two With-Grain RVD Specimens.

ct
2000

2000

o
'O
o

11200

STRESS (psi )

800

400

 

 

L

 

7

/

 

-

Z4

 

 

/

z

AN

 

 

 

 

 

 

 

 

N\

 

304

 

 

005

ol 0i5 0.2 025 03 035
LONGITUDINAL STRAIN, €A (%)

Q.49

045

ORNL, DWG. 69-2972

 

 

/

 

 

¢t

 

/|

 

 

 

 

 

1Y

 

304

 

 

 

003

002

oo . O

-QO0l

-002 -003

LATERAL STRAIN,€ (%)

Fig. 22. Stress-Strain Curves for an Across-Grain RVD Specimen.
3k

ORNI, DWG. 69-2973

 

 

1200

 

 

1000 [— .
" ‘3

 

 

 

 

AEEy/a N/
"’ / /4

 

 

 

 

 

 

 

 

 

w LS /I
/;/ W

 

 

 

 

 

 

 

 

 

200
P // P
o V4
0 002 004 006 008 010 0004 O -0004 -0.012
LATERAL STRAIN, EI (%)

LONGITUDINAL STRAIN,E, (%)

 

1200 /

| /
1000 ’ / -

S/ T Y/
A/ /)
7T

/]

0

o 0.02 004 0.06 008 010 0004 0 -0.004 -0.012
LATERAL STF\’AIN,ET (%)

LONGITUDINAL STRAIN, €, (%)

 

 

 

 

 

o
©
©

N
S
~.

 

 

 

srm-fs (pst)
N
SO\

 

H
Q
O

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Fig. 23. Stress-Strain Curves for Two With-Grain EGCR-Type AGOT

Specimens.
 

 

 

 

600
2

500 V4
= / 3
g. 400 " v
& /
& 300 A/ 7
[
o)

 

200 /

 

 

 

W/

 

 

T

 

 

 

0 0.02 004

 

 

 

 

 

 

 

 

 

 

 

 

 

006 008
LONGITUDINAL STRAIN, €, (%)
600 -
s00 ? // ;
3 200 A
8 | / 7
W 300| - /.// _
& / /
200 ,////
o A/
) / 2T
o .
0O 002 004 006 008 O

LONGITUDINAL STRAIN, €, (%)

O.1

35

ORNL DWG. 69-2971

 

 

 

 

 

 

 

 

 

 

 

T

 

 

000

]
.

-0004 -0.008 -00I2

LATERAL STRAIN, ET (%)

-00lI6

 

 

%—_
L™
s
.\

 

=
P —

 

 

 

 

 

 

 

 

/

2T

 

 

0.004

0 -0004 -0008 -00I12 -0.016

LATERAL STRAIN, € _(%)

Fig. 24, Stress-Strain Curves for Two Across-Grain EGCR-Type AGOT

Specimens.
36

so far as stress-strain response 1s concerned. The greater differences
between the first and second tensile loadings for across-grain RVD and
AGOT specimens may be associated with opening of microcracks too large to
be healed by heat treatment. This conjecture is the most plausible for

- the coarse-grained AGOT material.

In addition to the investigations just described, studies were made
to examine the influence of hold time at 3000°C. EGCR-type AGOT graphite
specimens were used, and the hold times were 20, 30, and 40 minutes (in
addition to the 60 minutes hold time used in cbtaining the results already
discussed). The results were inconclusive because no definite trend with
hold time was established. However, it was decided that the longer hold
time is preferred to assure greater possibility for reproducible results.

Returning to the dual purpose series for examining the asymptotic
approach feature and for obtaining data on cyclic behavior under alternate
compressive and tensile loadings, three (two with-grain and one across-
grain) EGCR-type AGOT specimens were tested. After an initial heat
treatmgnt at 3000°C for 60 minutes, the first loading for each specimen
was used to establish a one-cycle envelope curve. FEach was loaded in
compression near that required for failure, unloaded, loaded in tension
to a given stress level, again near that required for failure, and un-
loaded. During preliminary tests it was found that a specimen will fail
in tension when the tensile strain, as measured from the zero stress
point, corresponds to that at failure for a virgin specimen loaded in
tension only, regardless of the position of the zero stress point.

The envelope curves are shown as short-dashed lines in Figs. 25, 26,
and 27. The apparent discontinuities in the slopes of the curves between
the tensile and compressive portions were caused by changing fixtures to
apply loads of different signs.

Each specimen was heat treated a second time under the same conditions
as above and retested. It was loaded in compression to a given stress
level, unloaded, loaded in tension, unloaded, loaded in compression to a
higher stress level than the first, and unloaded. The corresponding
curves are shown by the long-dashed lines in the figures. Finally, the
heat treatment and loading sequence was repeated using lower compressive

. stress levels. The solid curves in the figures were obtained in this way.
37

ORNL DWG. 69-29T70

 

 

 

 

 

 

-3000
-2500 7 ¥
/:/ %
/’ ; ‘ ) \\\\
-2000 7 q
— // / ’I . \ \
: 1 W

 

 

1500 J | N '
/17 \

-1000 /}, /, ’/, !\‘\\\
// / Il | \\\ \
-500 //" II \\\\ \

 

STRESS (
N,
~

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1
0 ”4” \
/1) 7 ‘
/i \
[y 5P \ \
500/ —L \
o -0.1 -2 -03 -04 -05 -06 o0B 006 004 002 0
LONGITUDINAL STRAIN, €A {%) LATERAL STRAIN, €. (%)

Fig. 25. Multi-Cycle Curves for a Heat-Treated EGCR-Type AGOT Speci-
men (With-Grain).
38

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

-3000
- 2500 ——
T
- /
d’ ,
-2000 i ~
/// ,'
v / ’
= e J
%1500 4 ,
T ,1 ,I
3 /%// / A
x-1000 7r) L/ -
- / / ,l
m 717,
/// / /’
~-500 A 2
// / /,’
C/ /)
0 / /: ,/l
/7 £ s
// ,I,/
¥
500 £
6P
1000
o -01 =-02 -03 -04 -05

LONGITUDINAL STRAIN,GA (%)

-06

ORNL IWG. 69-29T4

 

 

 

 

 

 

 

 

 

 

 

 

 

006 004 002 0
LATERAL STRAIN, € (%)

Fig. 26. Multi-Cycle Curves for a Heat-Treated EGCR-Type AGOT Speci-

men (With-Grain).
39

 

 

 

 

STRESS (ps¢)

 

 

 

 

 

 

 

 

 

 

 

-3000
-2500 s
. ] ) ”’II‘
-2000 ”/ II’
th ’l
. /‘l ’
4 /
-1500 # -
. 1 . A
/f / /'I
fl/ / ’
-1000 f 2
/7/ /s
[/
- 500 [—HptF—rt
iAo
l/ / S
0 ll / /
Yoour
l . "I’
/i
6T
500 '
0 -02 -04 -06 -08 ~-1.0 -2
LONGITUDINAL STRAIN, €, (%)
Fig. 27.

Specimen (Across-Grain).

ORNL DWG. 69-2976

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Q06 004 Q02
LATERAL STRAIN, ET (%)

Multi-Cycle Curves for a Heat-Treated EGCR-Type AGOT

 
Lo

The maximum tensile stress level was essgsentially constant for all load-
ingé on a given specimen.

In all cases, the stress level for reloading approaches that at the
unloading point before a transition in slope occurs. The curvature of a
stress versus longitudinal strain curve generally changes rapidly in the
immediate vicinity of the unloading point, while the change for a stress
versus lateral strain curve occurs over a much wider stress interval.
Behavioral details for the stress versus longitudinal strain curves shown
as solid lines are - difficult to detect because of the small scale. In
fact, the initial loading and relocading curves were separated in the fig-
ures for clarity. The discontinuities in slope attributable to changing
of loading fixtures are apparent in almost all cases. The hysteresis
loops and other aspects of the behavior are in accord with the results
from other cyclic tests.

Consider the stress versus longitudinal strain curves. The‘asymp-
totic approach feature at the lower stress levels is shown, but there is
some deviation.as the unloading stress level is increased. However,
Figs. 25, 26, and 27 show that the upper portions of the dashed curves
lie above the envelope curve in two caseg and below this curve in one
case. On the whole, asymptotic approach to an initial loading curve is
demonstrated by these data. |

Additional support is given by the stress versus lateral strain
curves of Fig. 25, where the agreement with the envelope curve 1s excel-
lent. However, the stress versus lateral strain curves in Figs. 26 and
27 indicate that asymptotic approach is at most closely approximated.
Conversely, because the restoration of the original stress-strain response
by heat treating was shown to be less than perfect by the studies made to
examine this technique, fully satisfactory agreement would be difficult
to achieve.

As in the cases of the longitudinal strain curves for the two speci-
mens which show greater differences, the deviations from the envelope
curves for the lateral strains are in one direction. The two subsequent
loading curves in each case appear to define an envelope curve although

it is different from the one established during the first loading.
al

A better selection for studies of this kind would have been ATJ
graphite specimens since heat treatment was shown to bé more effective
for this than the other two materials examined. As demonstrated here,
discovery of the heat treatment method allows for detailed comparisons

where none could be made formerly.

Conelusions

This report describes the stress-strain responses of nuclear-grade,
or equivalent, graphites under monotonic and cyclic loading conditions.
Both stress versus longitudinal and stress versus . lateral strain data are
considered, and it .is shown that consistent patterns of behavior are
identifiable from the results. These include differences in deformation
resistance in tension and compression and differences in strain-ratio
data for the two types of loading.

When a specimen that has been preloaded in compression is cycled in
compression between zero and given maximum stress levels, three behav-
ioral patterns are observed. For the case in which the maximum stress
levels are continually increased and all are greater than the preload
value on each subsequent cycle, the hysteresis loops become larger and
the paraelastic moduli decrease with increasing maximum stress. Cycling
between zero and the preload stress level produces hysteresis loops that
do not change with cycle number and the paraelastic moduli remain the
same. However, the loops translate along the strain axis for.the first
few cycles, after which the loops coincide. Fioally, cycling between
zero and a fixed maximum value which is less than the preload stress pro-
duces hysteresis loops’that are essentially invariant both in size and
location with cycle number. In the case of a specimen which is loaded,
partially unloaded, and reloaded, hysteresis loops are formed by the
stress versus longitudinal strain and stress versus lateral strain curves.
Also, the reloading curves approach what would have been extensions to
the initial loading curves more rapidly than for the case‘of full unload-
ing. The approach to the extension of a continued loading curve on re-
loading is often assumed. Here, this postulate is shown to be supported

by data obtained from tests designed to examine its validity.
Lo

The use of heat treatment at the graphitization temperature essen-
tially removes the effects of prior deformation, so far as stress-strain
response is concerned. However, these heat treatments were more success-
ful in rembving the prior deformation histories of ATJ and RVD graphites
than of EGCR-type AGOT graphite. Careful use of this technique is shown “
to be of unique benefit for making comparisons of the stress-strain be-

haviors of a single graphite specimen under different loading histories.

Acknowledgements

 

The authors gratefully acknowledge the assistance of J. M. Napier of
the Oak Ridge Y-12 Plant in heat treating the specimens; J. M. Chapman
and J. P. Rudd for their help in instrumenting and testing the specimens;

and H. A. MacColl for preparation of the figures.
1.

2.

L3
References

W. L. Greenstreet, "Mechanical Properties of Artificial Graphites —
A Survey Report," USAEC Report ORNL- M527, Oak Ridge National Labo-
ratory, December 1968.

W. L. Greenstreet, J. E. Smith, and G. T. Yahr, "Mechanical Propertles

“of EGCR-Type AGOT Graphlte," Carbon, T7(1): 15-45 (1969).

P. P. Arragon and R. M. Berthier, "Caractérisation mécanique du
graphite artificiel,"” pp. 565-578 in Industrial Carbon and Graphite,
Society of Chemical Industry, London, 1953.

E. J. Seldin, "Stress-Strain Properties of Polycrystalline Graphites
in Tension and Compression at Room Temperature,” Carbon, W(2): 177-

191 (1966) .

H. H. W. Losty and J. S. Orchard, "The Strength of Graphite," pp.
519-532 in Proceedings of the Fifth Conference on Carbon, held at

-Pennsylvania State University, Vol. 1, MacMillan, New York, 1962.
45

ORNL-TM-2727

 

 

Internal Distribution
1. S. E. Beall ' 34, G. B. Marrow, Y-12
2. H. W. Behrman, RDT 35. J. G. Merkle
3, 5. E. Bolt | 36. A. J. Miller
4. J. W. Bryson 37. S. E. Moore
5. J. P. Callahan 38. J. M. Napier, Y-12
6. 8. J. Chang : 39. A. M. Perry
7. W. H. Cook o - k0. C. E. Pugh
8. J. M. Corum | | 41. J. N. Robinson
9, W. B. Cottrell 42, M. W. Rosenthal
10. F. L. Culler 43, A. W. Savolainen
11. W. G. Dodge 4h, M. J. Skinner
12. W. P. Eatherly Ls-s4, J. E. Smith
13. A. P. Fraas 55. I. Spiewak
14-23. B. L. Greenstreet 56. J. L. Spoormaker
24. R. C. Gwaltney 57. D. A. Sundberg
25. P. N. Haubenreich 58. D. B. Trauger
26. P. R. Kasten 59-63. R. S. Valachovic
27. C. R. Kennedy 64. M. S. Wechsler
28. K. C. Liu 65. G. D. Whitman
29, M. I. Lundin 66-75. G. T. Yahr
30. R. N. Lyon T6-77. Central Research lerary
31. H. G. MacPherson 78-79. Y¥-12 Document Reference Section
32, R. E. MacPherson 80-84. Laboratory Records Department
3%, H. C. McCurdy 85. Laboratory Records, ORNL R.C.
External Distribution
86. S. A. Bortz, IIT Research Institute, Chicago
87. H. L. Brammer, SNPO-C, NASA Lewis Research Center, Cleveland
88. L. C. Corrington, SNPO-C, NASA Lewis Research Center, Cleveland
89. J. F. Cully, SNPO-A, c/o USAEC, P. O. Box 5400, Albuquerque
90. R. J. Dietz, Los Alamos Scientific Laboratory
91. D. M. Forney, Air Force Materials Laboratory (MAC) Wright-
Patterson Air Force Base, Ohio
92. C. W. Funk, Aerojet-General Corp., Sacramento
95. J. dJ. Gangler, Materials Engineering Branch, RRM, NASA, Washington,D.C.
o4. Harold Hessing, SNPO-A, CMB Division, Los Alamos Scientific Lab.
95. A. N. Holden, Westinghouse Astronuclear Laboratory, Pittsburgh
96. Gary Kaveny, Aerojet-General Corp., Sacramento .
97. J. J. Lombardo, SNPO-C, NASA Lewis Research Center, Cleveland
98. L. L. Lyon, Los Alamos Scientific Laboratory
99. D. P. MacMillan, Los Alamos Scientific Laboratory
100. M. M. Manjoine, Westinghouse Astronuclear Laboratory, Pittsburgh
101. J. E. Morrissey, USAEC, Washington
162. R. E. Nightingale, Pac1f1c-Northwest Laboratory, Richland
105. W. G. Ramke, Air Force Materials Laboratory, Wright-Patterson
Ai

ir Force Base, Ohio
1042105,
106.
107.
108.
109.
110.
111.
112.
113.
11k,
115,

116-130.
131.

ZQOE';UM"&:!';UEQ

L
. Rowley, Los Alamos Scientific Laboratory

. Scheib, SNPO, USAEC, Washington

. Schroeder, SNPO-C, NASA Lewis Research Center, Cleveland

. Schwenk, SNPO, USAEC, Washington

. Seldin, Parma Research Center, Cleveland

. Singleton, Westinghouse Astronuclear Laboratory, Pittsburgh
. Smith, Los Alamos Scientific Laboratory

. Spence, Parma Research Center, Cleveland

. Swanson, Westinghouse Astronuclear Laboratory, Pittsburgh
Thielke, SNPO, NASA Iewis Research Center, Cleveland

Tu Lung Weng, Parma Research Center, Cleveland

Division of Technical Information Extension (DTIE)

HgHOWQOHGOR 0N

- Laboratory and University Division, ORO

Yy
'