Title no.
97-M35
Dynamic Properties of Mass Concrete Obtained from Dam
Cores
by David W. Harris, Caroline E. Mohorovic, and Timothy f? Dolen
This paper summarizes the results of a US. Rureau of Reclamation
research projext designed to provide a broad database of material
properties for mass concrete tested at strain rates that correspond to
seismic (dynamic) and static loadings. Laboratory tests were per-
formed on cores drilledfrom dams and tested at strain rates that sim-
ulated dynamx and static loading conditions. Ratios of dynamic and
static compressive strength, modulus of elasticity, compressive failure
strain, Poisson’s ratio, and splitting tensile strength are summarized.
The effects of core size and sample saturation are also considered.
Results of this study indicate that dynamic-static strength ratios
aregreater than onefor both the compressive and tensile strength tests.
The dynamic-static ratios of modulus of elasticity andfailure strain
for the compressive strength tests, however, are generally less than or
equal to one.
Keywords: axe tests; dynamic modulus; dynamic strength; earthquake
propertIes; mass concrete.
RESEARCH SIGNIFICANCE
This paper provides a large database of test results for mass
concrete represented by drilled core tested at both static and
dynamic (seismic) loading rates. Dynamic and static strength
and elastic property results from several large-scale test pro-
grams with a total of over 470 test specimens are summarized
and compared to uncover trends in the data.
INTRODUCTION
The dynamic properties of mass concrete are an important
consideration in the analysis and review of the safety of struc-
tures such as concrete dams. Concrete tests can be designed to
predict the behavior of a structure under various static and
dynamic loading conditions. In laboratory tests, different
dynamic conditions are modeled by varying the strain rate at
which the test is performed. Thus, the strain rate of the tests is
key to the interpretation of results.
The dynamic properties of concrete have been studied for a
wide range loading rates. Much of the work has been summa-
rized as a state-of-the-art paper by Bischoff and Perry (199 I).
Strain rates and their conceptual equivalent situation are tabu-
lated as follows (Bischoff and Perry 1991)
Strain rate,
Condition
trim/mm (in/in.) per s
-6
Creep 1o-x to 10
4
Static 1o-6 to 10
-9
Earthquake Lo-J to 10
Hard Impact 10” t o 10’
Blast 10% to lo8
The American Concrete Institute (ACI) defines mass con-
crete as a volume of concrete large enough that measures must
be taken to address the heat of hydration of the cement and the
volume change of the material. The U.S. Bureau of
290
Reclamation’s concrete dams are mass concrete structures that
typically have a nominal maximum size aggregate of about 6 in.
(150 mm). Reclamation is particularly concerned \vith the per-
formance of these structures when subjected to earthquake
loads. For approximately 1.5 years, its laboratory core test pro-
grams have typically included dynamic tests performed at
strain rates corresponding to seismic loads.
This paper summarizes the results of a Reclamation
research project designed to pro8de a broad database c!f the
behavior of mass concrete from existing dams under dynamic
loading conditions that simulate earthquake loadings.
Laboratory tests performed on cores at both tiaditional, static
loading conditions (strain rates of IO‘” t o 10~’ in./in.
[mm/mm] per s) and dynamic loading conditions (strain rates
of 10.~ in./in. [mm/mm] per s) are compared. Dynamic and
static measurements of compressive strength, modulus of elas-
ticity, compressive critical strain (strain at ultimate compres-
sive strength), Poisson’s ratio, and splitting tensile strength are
summarized. In some cases, the core size or moisture condition
among similar samples was varied to determine if either of
these parameters affect test results.
Data from previous and current test programs conducted at
the U.S. Bureau of Reclamation Materials Engineering and
Research Laboratory, Denver, Colo. are provided. Results from
Reclamation’s past test programs that include dynamic and
static compression, splitting tension data obtained in a similar
manner, or both, are summarized. An additional 103 cores from
two dams were tested under dynamic and static compression
and splitting tension loading conditions.
TEST PROGRAM
‘Sample description
The dams that provided test cores for this study are sum-
marized in Table 1. All but two are currently owned by
Reclamation. Pine Flat and Englebright Dams are currently
owned by the U.S. Army Corps of Engineers (COE), and
Folsom Dam, though now owned by Reclamation, was owned
by COE at the time of testing.
T h e 1998 current test data consist of data from N’arm
Springs Dam and Roosevelt Dam. Results were obtained from
inventoried core and were specifically tested for this study from
May to September 1998 at Reclamation’s Denver laboratory.
The remaining data, hereby referred to as the historical data,
were extracted from previous test programs that were con-
ducted at Reclamation’s Denver laboratory All tests included
in the historical data and tested at Reclamation were performed
according to laboratory standards similar to those used for the
current tests and with contemporary and similar testing appa-
ratus. This constitutes data obtained from cylindrical cores
ACI ‘TmPds Journal v 97. No. 5. May-June Yooo.
MS No. ~-08l received April e;, ,999. and rewewed under Institute pubhcatton
pohcies. Copyright 0 WOO, American Concrete Institute. All rights reserved. mclud-
mg the making of copies unless permIssion IS obtamed from the copyright proprietors.
Pertment dlscussion wll be publnhed m the March-April 9001 ACIMatprzaLi JoumJ
if recewed by December I, 4000
ACI Materials Journal/May-June 2000
 obtain a specimen with a length-diameter ratio of 4.0 whenev-
er possible.
The static compression and static splitting tension test
results for Folsom, Pine Flat, and Englebright Dams were sup
plied by the COE when the dynamic tests were performed at
and b y Reclamation, for the purpose of comparison with
Reclamation’s dynamic test results.

drilled from dams that were tested according to present-day
laboratory standards, and at the strain rates defined for seismic
(dynamic) and static compression and splitting tension tests.
These data were obtained from files of the U.S. Bureau of
Reclamation Laboratory and from papers published by Gaeto
(l&3+) and Peabody and Travers (1986 a,b).
The test data reflect the great variability among mass con-
crete mixture proportions. The structures were placed using
mass concrete construction techniques that include nominal
maximum aggregate sizes of 3 . 0 i n . ( 7 5 mm) or larger.
Concrete mixtures reflect the state-of-the-art concrete technol-
ogy at the time of construction, and were partly controlled by
local conditions. Aggregates were obtained locally, and reflect
the geological history of the area. Such characteristics make
mass concrete dependent on both the time period and location
at which the concrete was made.
Sample preparation
All specimens considered in this study, with the exception of
the static compression and static splitting tensile test results
for Folsom, Pine Flat, and Englebright Dams, were prepared
and tested at the Bureau of Reclamation’s Materials
Engineering and Research Laboratory, Denver, Colo. Cores
were typically obtained by Reclamation’s regional drill crews,
packed on site, and shipped to the Denver facility. Test speci-
mens were cut to length using a diamond-impregnated saw to
Historical data
After drilling, cores are sealed in plastic to best maintain
their in-place moisture content. At the Denver laboratory,
cores are maintained either wrapped in moist cloth and covered
with plastic, or in a constant climate-controlled humidity room
maintained at a relative humidity not less than 95%
(ASTM C 511) prior to testing. The drilling and testing pro-
grams are usually completed within a few months.
All core samples from the historical data had a diameter of
6 in., except the core from Monticello Dam, which had a diam-
eter of 10 in.
Current data
The test program for this study was designed not only to
provide the specified comparisons of dynamic and static
material properties, but also to study the effects of the surface
moisture condition of the core and the core size on these
results. The cores from the Roosevelt and Warm Springs Dams
used in this study had been stored and air-dried from 2 to 5
years.
To study the effects of surface moisture condition, the 1998
samples from Roosevelt and Warm Springs Dams were tested
at an air-dried and in a surface-saturated, surface-dry (saturat-
ed) moisture condition. Since the core had been exposed for
several years, air-dried samples were tested in the moisture
condition at which they were found. Saturated specimens were
tested after having been submerged in lime water for at least
4 0 h, as suggested in the ASTM Procedure C 43-94. T h e

Table l-Concrete dams used for core sources

Llonticello Dam, Calif. Thick arch 6 in. Vertical

I
\Varm Springs Dam, Oreg. Thin arch 1919 106 ft 6 in. Vertical
I I I
Vertical/
Hoover Dam, Kev./Ariz. Thick arch 1936 726 ft 9 in.
horizontal

Stewart Mtn. Dam,‘Ariz. Thin arch 1930 ( 207 ft - Vertical

I I 1 1 I

Folsom Dam’, Calif. Gravity 1956 340 ft Vertical

Roosevelt Dam, Ariz. 4 in. Vertical/

Mod. - buttress Mod. - 1995 280 ft
(modification) (mod.) horizontal

Englebright Dam+, Calif. Variable arch 1941 280 ft -

I

Pine Flat Dam+, Calif. Gravity 1954 440 ft - -

/ I I

‘M’arm Springs, Folsom. and Stewart Mountain Dams were not constructed by Reclamatmn
+Englebnght and Pme Flat Dams are owned by U. S. Army Corps of Engmeers.

ACI Materials Journal/May-June 2000 291

density of each sample was measured before and after satura-
tion to determine the absorption of the concretes.
To study the effects of core diameter size, both 6 and 12 in.
(152 and 305 mm) diameter cores from Roosevelt Dam were
tested. All Warm Springs Dam cores measured 6 in. (152 mm)
in diameter.
Sample populations
Several test populations were developed from the 103 test
specimens to isolate the moisture conditions and the core size
parameters. The 6 in. (152 mm) diameter Warm Springs core
was evenly divided to test at both the saturated (Group A) and
the air-dried (Group B) moisture state. Eight 12 in. (so5km)
diameter samples were available from the Roosevelt core. Eight
6 in. (152 mm) diameter saturated cores (Group l(a)) and eight
6 in. (152 mm) diameter air-dried cores (Group l(b)) were test-
ed and compared with the eight 12 in. (305 mm) diameter cores
that were also tested air-dried (Group 2). All 24 specimens
were extracted from a similar location in the dam. Each test
group included dynamic and static compression and dynamic
and static splitting tension tests.
To further isolate the effect of saturation on the remaining
6 in. (152 mm) diameter Roosevelt Dam core, the population
was evenly divided to perform additional dynamic and static
compression tests (Groups 3 and 4) and splitting tension tests
(Groups 5 and 6) at both the in-place, air-dried and saturated
conditions.
In summary, the 1998 test specimens were grouped as follows:
Tests performed
Static compressive strength was determined according to
the ASTM C 39, Standard Test Methodfor Compressive Strength
of Cylindrical Specimens. For the 6 in. (152 mm) diameter core,
the modulus of elasticity and Poisson’s ratio were determined
u s i n g 4 i n . ( 1 0 2 mm) epoxied electrical strain gages, two
secured laterally and two secured axially, as stated in ASTM C
469, Standard Test Methodfor Modulus of Elasticity and Poisson’s
Ratio in Compression. The 12 in. (305 mm) diameter core
required twelve 4 in. (102 mm) gages, bridged in four groups of
three gages according to procedures previously described.
Ultimate strain was measured from the axial gages, and strain
is reported in units of millionths. Static splitting tensile
strength was determined according to ASTM C 496, Splitting
Tensile Strength of Cylindrical Concrete Specimens.
Dynamic tests were performed according to the ASTM pro-
cedures for the static tests, with the exception of loading rate.
In the Denver laboratories, a uniaxial testing machine capable
292
of providing failures within a strain rate of lo-’ and a time
frame of 0.05 to O. 1 s was used. The equipment uses a hydraulic
ram mobilized by an oil pump.
For all tests, strain gages directly provided strain measure-
ments, while load was measured and converted to units of
stress.
The density of the 6 in. (152 mm) concrete samples was
determined from the specific gravity of the specimens. Specific
gravity was computed from the mass of the specimen in air,
divided by the specimen mass in air, minus the apparent speci-
men mass submerged in water. During submergence, samples
were suspended in water by the scale apparatus. Density was
then calculated by multiplying the specific gravity of the con-
crete by the density of water, which is dependent on the water
temperature.
The density for both the saturated and the air-dried 6 in.
(152 mm) samples were calculated in this manner. This test
method requires the sample to be submerged in water for only
a few seconds, and the air-dried samples were allowed to dry
for several days prior to testing. Therefore, the submergence
procedure of the density test shotld not have affected thf sur-
face condition of the samples.
The mass of the 10 in. (254 mm) core samples exceeded the
capacity of the laboratory equipment used to suspend the sam-
ples in water. Thus, density was calculated as ‘the mass of the
sample in air divided by the sample volume, which was deter-
mined from surface dimensions.
TEST DATA
Nature of test results
Table 2 to 7 and Fig. 1 to 5 summarize the average test
results of the historical and current data according to the
arrangement of the 16 test populations. The tables provide the
average and standard deviation of the data, as well as the ratio
found between the average dynamic and average static materi-
al properties. Averages are derived from the average results of
each population, and are provided to summarize trends of the
data. The number of tests performed for each test program
varies. The data from the Warm Springs and Roosevelt core
were divided into several test groups to isolate particular
parameters, and each group is reported separately in the aver-
age. Thus, the current data comprised of test specimens from
Warm Springs and Roosevelt Dams represent nine of the 16
test populations included in the average.
The averages should not be interpreted as indicative of mass
concrete in general. The averages should simply be considered
as a guide that describes the trend of the test populations pro-
vided in Table 2 to 7.
Ultimate compressive strength
Data for compression tests are summarized in Table ‘L and
Fig. 1. The average dynamic compressive strengths of the, 16
test populations are generally slightly higher than the average
static compressive strengths. The average dynamic-static com-
pressive strength ratio of these results is 1.07 with a coefficient
of variation (COV) of 20%. The ratio ranges from 0.73 to 1.45.
Due to the significant variation within the test population, the
dynamic to static compressive strength ratio for mass concrete
should be determined on a case-by-case basis.
A s i n d i c a t e d i n F i g . 1 , the dynamic-static compressive
strength ratios tend to decrease as the static compressive
strengths increase. The data are too dispersed, however, to
draw any statistical conclusions. No significant statistical cor-
relation was found among compressive strength and dynamic-
static compressive strength ratio.
Modulus of elasticity
For the 16 mass concrete test populations summarized in
ACI Materials Journal/May-June 2000
 50

.
. l
40 ~Ccompnrdvennneth~
vh2(cacdc9th0.98m - 1.28m)
. . . .
. l * .
. - 30
. l
.
.
5.

f 20
vh2(c0nde9ih0.67m-0.98 m)

10

Fig l--Average static compressive strength versus average dynamic

to static compressive strength ratio for 16 test populations in Table 2, Fig. +--Typical stress and strain curves for static and dynamic com-
L$tamic Properties Study, 1998. pressive strength test, Roosevelt Dam test cores, Dynamic Properties
Study, I 998.

10 7.0 30 40 50

*+mge staticuoduhuof- (ON)

Fig .+-Average static modulus of elasticity versus dynamic to static i 1 IS 2 2.5 3 1.5 4 4.5
modulus of elasticity ratiofor 15 testpopulations in Table S, Dynamic Avenge St& spliaillg Tauik stmgth @Pa)
Properties Study, 1998.
Fig. &-Average splitting tensile strength versus dynamic to static
splitting tensile strength ratio for 15 test populations in Table 6,
Dynamic Properties Study, 1998.

pressive strength test and a dynamic compressive strength test

0.0 0.2 0.4 06 0.8 1.0
Avenge Dynmidstaic Ratio - Modulus of EMicily
Fig. *Average dynamic to static properties ratios, compressive
strength versus modulus of elasticity, for 15 testpopulationsprovided
in Tables 2 and 8.
Table S, the dynamic moduli of elasticity tend to be slightly
lower than the static moduli of elasticity Although the average
dynamic-static compressive strength ratio is slightly greater
than 1, the average dynamic-static modulus of elasticity ratio is
0.89, with a COV of 17%.
As illustrated in Fig. 2, the ratio of dynamic modulus of
elasticity ranges from approximately 0.7 to 1.1 for all moduli
values. Results indicate that the modulus of elasticity is similar
at strain rates corresponding to static and seismic loading con-
ditions. Therefore, dynamic moduli did not tend to increase as
the dynamic strength increased relative to the static compres-
sion test. Figure s compares strength and modulus ratios. The
variation of the data in Fig. s do not indicate a clear correlation
among the strength and modulus ratios.
A typical example of stress and strain data for a static com-
are provided in Fig. 4. The test pair was extracted within one
concrete construction lift from Roosevelt Dam. The ultimate
dynamic compressive strength is slightly higher than the ulti-
mate static compressive strength.
The curves in Fig. 4 are essentially linear and similar in
slope from the origin to a stress, corresponding to approxi-
mately l/2 of the ultimate static compressive strength. After
1.2 1.4 1.6 this stress is reached, the curves diverge. The slope of the stat-
ic compressive strength curve becomes nonlinear, decreasing
asymptotically to the failure stress. Concrete typically yields
before it fails in this manner at static loading rates.
At dynamic loading rates, yielding is not observed in the
stress and strain data. The stress-strain response for the
dynamic test in Fig. 4 is predominately linear from initial load-
in#to failure of the specimen. This linear response is represen-
tative of the dynamic test results. Under compressive loads, the
yielding phenomena were typically eliminated from the test
data when the test strain rate was increased from the static load
rate to the dynamic load rate.
The modulus of elasticity is typically calculated as either a
secant or a chord modulus. A secant modulus is calculated from
the origin to a defined point on the curve, usually within SO to
60% of the sample’s ultimate strength. The chord modulus,
typically used in all Reclamation test programs and most
recent data, is measured according to ASTM C 469 between
the stress and strain pairs at 50 millionths strain and at 441% of
the ultimate compressive strength. Because the dynamic and
static stress-strain curves considered in this study were gener-
ally linear within the boundary conditions defined for these

ACI Materials Journal/May-June 2000 293

Table 2-Compressive strength test data ticity ratio of 2.15, which significantly deviated from all other
data. The value for static modulus of elasticity was provided by
Cmnpressive strength, Mpa
the COE in the mid-1980s. Details of the data were not avail-
stanc (St) Dynamic (Dyn) Average able to analyze the difference.
Average Standard Average Standard Dyn/St
w dewatvm w dewation rano
Critical strain
Critical strain data are provided in Table 4. For this study,
critical strain is defined as the strain measured at the ultimate
compressive strength of the sample. The average dynamic-stat-
ic critical strain ratio is slightly less than 1 for most test popu-
lations. The average ratio of dynamic-static critical strain of
eight ratios reported in Table 4 is 0.93, with a COV of 12%.
The average excludes the ratio of 1.58 from the Roosevelt
Group ~(a) test population, which deviates drastically from the
rest of the test population, and is considered an outlier.
The average ratio implies that concrete tends to be more
brittle under dynamic loading conditions. Critical strain is fur-
ther considered in the discussion of the effects of saturation.

Poisson’s ratio
Data for Poisson’s ratio are summarized in Table 5. The
average dynamic-static ratio of P&son’s ratio for the 1.5 popu-
lations summarized in Table 5 is 1.05 with a COV of 27%. The
1996 data for Warm Springs Dam, which provides a dynamic
Poisson’s ratio of 0.44, is excluded. The value is considered an
outlier because it significantly deviates from the rest of the test
population.
For most test populations, the average static Poisson’s ratio
*Number in parentheses indicates number of test specimens included m average.
and the average dynamic Poisson’s ratio are slightly higher
N/A = not available; number of data unknown or too small to prowde standard than 0.80. Increases or decreases of Poisson’s ratio do not cor-
dewation. relate with changes in average compressive strength. The aver-
age Poisson’s ratio ranges from 0.14 to 0.29 for the static com-
pression test population, and from 0.18 to 0.28 for the dynam-
Table 3-Modulus of elasticity test data ic compression test population.
1
Modulus of Eladtic~ty. GPa
I Splitting tensile strength
stanc (St) Dynamic (Dyn) Average Data for splitting tensile strength tests are shown in
Average Standard AWage Standard
deviation dewatmn
Dyn/St
r a t i o
Table 6. For 15 mass concrete test populations, the average
(W’ w
ratio of dynamic to static splitting tensile strength is I.++, with
a COV of 15%.
Figure 5 indicates that the dynamic-static splitting tensile
strength ratio tended to slightly decrease as the static com-
pressive strength increased. The dispersion of results at high-
er strengths in Fig. 5, however, makes it diicult to correlate a
relationship between dynamic-static splitting tensile strength
ratio and increasing splitting tensile strength. The variation in
data suggest that site-specific data should t+e used to evaluate
critical structures.
Ratios of splitting tensile strength to compressive strength
for both the static and dynamic test populations are provided in
Table 7. The average ratio of static splitting tensile strength
to static compressive strength is 0.10, with a standard devia-
tion of 0.03. The ratio of dynamic splitting tensile strength to
dynamic compressive strength ia 0.13, with a standard devia-
tion 0.04.

Roosevelt
Parameters studied
Group 3 Moisture condition ofcort-pairs of similar Warm Springs
Hoosevelt 1998 core populations and Roosevelt c&e populations (Groups
Group 4
l(a) and l(b); Groups s and % and Gkoups 6 and 6) that were
‘Number in parentheses indicates number of test specimens included in average. tested at an air-dried and a saturated cohdition are represented
N/A = not available; number of data unknown or too small to prowde standard
deviatmn. in summary Table 2 through 7.
In general, the T-day saturation of specimens tended to
decrease the static and dynamic compressive strengths, and
moduli, measurements of secant and chord moduli would be increase the static and dynamic splitting tensile strengths.
similar. Almost all dynamic-static critical &rain ratios were greater
Results for Folsom Dam were excluded from the moduli for the saturated cores than for& air-dried cores. Too few
study. The results provided a dynamic-static modulus of elas- data providing critical strain kt different surface moisture
,,

294 ACI MateWM ay-June 2000

Table 4-Compressive strength critical strain test Table 6-Splitting tensile strength test data
data

Test
Program

Roose\elt Group J 1360(H) 160 l550 (6) 500 i 0.99

.~~- -+p---mm~
Rwse\ elt Group 4 1150(R) Y10 NG N/A N/A

‘Sumber m parentheses mdicates number of test specimens included m average.

S’.i = not wallable, number of data unknown or too small to provide standard
de\ ,at,on.
SC = good test data not obtamed.

Table 5-Poisson’s ratio test data

‘Number m parentheses mdlcates number of test specimens Included in average.
N/A = not available; number of data unknown or too small to provide standard
Ponson’s Ratio
deviatmn.

Table 7-Ratio of splitting tensile and compres-

sive strength results
Static strengths. MPa Dynan,~ rtrcngths. .\IPa
and ratio and ratlo
Spllttlng Com- Sphttmg Cm-
tenSlIe PreSSWe Ratio tensile pressi~r IMl”
PK?JW (SST) (SC) (SST/SC) (DST) (DC) (DCWDC)
Deaduood s 14 se.1 “.OW $76 40.9 0 116

E’ephanr !2 L4 L7.h 0 155 5.45 1152 O.IS6

Butte
Monacello e.3 1 3O.R 0.07” %4X $3 (I 0.104

‘Sumber m parentheses mdlcates number of test specmxzns mcluded m average.

+Outber not mcluded in average of dynamicistatic ratms.
S A = \alue not awlable.

contents, however, are available to draw conclusions from these
results.
For these variations in material properties, the correspon-
ding effect of the submergence process on the density of the
samples was extremely small. The average density of all SO
specimens from Warm Springs Dam prior to soaking was
1 2 8 . 8 lb/ins (206’ k g / m ” ) . F o r t h e 1 6 saturated samples,
submergence did not change the density. The average density
of 68 samples from Roosevelt Dam was 147.0 lb./in.s (2354
kg/ms) and the density of the 26 saturated samples increases
about 1% after soaking.
Although the effects of submergence on core samples tend-
ed to vary among test populations, the saturation process does
seem to affect the elastic properties of the test specimens.

ACI Materials Journal/May-June 2000 295

These changes were observed even though increases in the
density due to soaking were extremely small. To most
accurately predict the material properties of a structure, mass
concrete samples should be maintained and tested as close as
possible to an in-place moisture state.
Core size-Core from Roosevelt Dam was tested and com-
pared for diameters of 6 in. (152 mm) (Group I(b)) and 12 in.
(SOS mm) (Group 2). The larger core generally indicated lower
strengths and strains for both static and dynamic compression
and splitting tensile loading, but higher moduli of elasticity.
ASTM C 42 specifies that core diameters must be at least twice
the nominal maximum aggregate size, and preferably, three
times that size. Because the maximum aggregate s&e at
Roosevelt Dam is approximately 4 in. ( 127 mm), the 12 in. (SO5
mm) diameter core test results are considered more represen-
tative.
Comparison of results to other published values
The results of this study are compared with data provided
by Bischoff and Perry (1991) and Raphael (1978 and 1984).
These works discuss several test programs that used cylindri-
cal concrete specimens tested at similar static and dynamic
(seismic) loading rates. The increases in compressive and ten-
sile strength that resulted from increased loading rate indicat-
ed in this paper are consistent with the other published find-
ings. Raphael (1984) reports an average dynamic-static split-
ting tensile strength ratio of 1.45, and a dynamic-static com-
pressive strength ratio of 1.3 1 for the same loading rate ranges.
Bischoff and Perry (1991) report that compressive strengths
typically increase from o to SO% at loading rates that simulate
seismic loads, when compared with static loads. These results
agree with the average dynamic-static ratios of 1.07 for the
compression tests, and 1.44 for the splitting tension tests in
this study.
Bischoff and Perry (1991) also report that increased com-
pressive loading rates tend to provide stress-strain curves that
remain linear up to higher compressive stress values, which
they attribute to a decrease in internal microcracking for a
given level of stress at higher strain rates. In this study, stress-
strain curves produced from dynamic compression tests were
also more linear up to the point of failure than those produced
from static compression tests.
The mass concrete data reported by Raphael (1978) indicate
that, in general, the modulus of elasticity tends to increase with
increasing loading rate. Bischoff and Perry (199 1) note that
there is some discrepancy in this conclusion when the tangent
method is used to calculate modulus of elasticity. Several
results reported in this work predict no change in modulus for
dynamic loading. The results of this study, which is based on
aged mass concrete cores at the dynamic loading rate specified
to model seismic loading conditions, generally resulted in
lower moduli values than the static tests.
CONCLUSIONS
1. Dynamic compression tests indicated higher values for
strength compared with the static compressive strength. For
the 16 test populations, the average dynamic-static compressive
strength ratio is 1.07, with a COV of 20%. This ratio does not
depend upon the magnitude of the dynamic and static com-
pressive strengths. The variability of the results indicate that
compressive strength measurements should be determined on
a case-by-case basis;
2. Dynamic splitting tension tests indicate a significant
296
increase in strength compared with static splitting tensile
strength values. The average dynamic-static splitting tensile
strength ratio for 1.5 test populations was 1.44, with a COV of
16%. This ratio does not depend upon the magnitude of the
dynamic and static compressive strengths. The variability of
the results indicate that splitting tensile strength measure-
ments should be determined on a case-by-case basis;
S. The dynamic compression tests indicate a decrease in
moduli of elasticity compared with the static moduli of elastic-
ity. The average ratio of dynamic to static modulus of elastici-
ty (using the ASTM standard for calculation) was 0.89, with a
COV of 17%. Consequently, the average dynamic moduli did
not tend to increase as the dynamic strengths increased;
4. The concrete tends to yield before failing under static, but
not dynamic, load rates. Dynamic tests provide stress-strain
curves that are generally linear in nature from the origin to
failure. The stress-strain curves of the static compression tend
to begin as linear, and then decrease in slope and flatten as they
approach failure;
5. For air-dried test specimens, critical strains are generally
lower for dynamic tests, indicating that the materials are more
brittle under dynamic loading co&Jitions; *
6 . Saturation tended to decrease the static and dynamic
compression strengths and increase the static and dynamic
splitting tensile strengths, and did not significantly change the
density of the sample; and
7. Larger-diameter core generally yielded lower strength
and strain values and higher moduli values.
CONVERSIONS
Calculations reported in this paper were originally calculated in
English units and converted to SI (metric) using the following:
1 in. = 45.4 mm
O.oo6894 p s i = 1 MPa
6.894 psi = 1 GPa
16.018 Ib/fts = kg/m3
ACKNOWLEDGMENTS
We would like to thank the U.S. Bureau of Reclamation Dam Safety
Office, Denver, Colo. for funding this study. Special thanks to David
Kellstadt for contributing his time and resources to complete this study
Thanks to all employees of the Materials Engineering and Research
Laboratory concrete group for contributions to and support of this project.
REFERENCES
Bischoff, I? H., and Perry, S. H., 1991, “Compressive Behavior of
Concrete at High Strain Rates,” Matmiczls andShuhms, V ~4, pp. 4~5-450.
Gaeto, T. A., 198% “Stewart Mountain Dam Concrete Analysi--l%%,“
REC-ERC-84-5, Division of Research and Laboratory Services,
Engineering and Research Center, U.S. Bureau of Reclamation, Denver.
Cola., pp. 1-4.
Peabody, M., and Travers, E, (undated but approximately 1986a).
“Testing of Englebright Dam Cores under Rapid Loading Conditions.”
Memo reporting test results, U.S. Bureau of Reclamation, Denver, Cola..
pp. l-9.
Peabody, M., and Travers, E, (undated but approximately 1986b).
“Testing of Cores from Folsom and Pine Flat Dams under Rapid Loading
Conditions,” Memo reporting test results, U.S. Bureau of Reclamation.
Denver, Colo., pp. l-16.
Raphael, . I . M., 1978, “The Nature of Mass Concrete in Dams,” Douglas
McHmy Internattonal Symposium on Ccnm-& and Gmcretc Shwhms, SP-55.
B. Bresler, ed., American Concrete Institute, Farmington Hills, Mich., pp.
ISS-159.
Raphael, J. M., 1984. “Tensile Strength of Concrete,” AC1 JOUR?Z.4L,.
Proceedings V. 8 I , No. 2, Mar.-Apr. pp. 158-66.
Annual Rook of ASTM Stan&r&, 1998, V. 4.04, Concrete and
Aggregates, American Society for Testing and Materials, West
Conshohocken, Pa., 1998.
ACI Materials Journal/May-June 2000