UDC 666.97.017: 620.115.

82

Determining concrete strength by using

small-diameter cores
J. H. Bungey MSC, DIC, CEng, MICE

U N I V E R S I T Y O F L I V E R P O O L : D E P A R T M E N T O F C I V I LE N G I N E E R I N G

SYNOPSIS a testing programme with small-diameter cores to

Estimates of concrete strengthare often made from investigate the influence of a number of common
compression tests on cores whichhave a diameter con- variables, and also to examine the results in relation
siderably less than the recommended 100 mm. The to recommended procedures for ‘large’ cores.
paper examinesthe resultsofa laboratory investigation
in which 44 m m diameter cores were cut and tested. Aims and scope of investigation
The influence of both specimenandaggregatesize Thegeneralproblems of coretesting are well
upon heightldiameter ratio and orientation effects is known, and the factorswhich influence the relation-
examined, andobservedrelationshipsbetweencore ship between the strength of a core and the corres-
strengthsandmeasuredcontrol cube strengths are ponding Actual Cube Strength are described fully in
compared withthose normally used for larger cores.It Part 5 of the Concrete Society Report(*),which also
is shown that core strengthis affected by both specimen summarizes past researchon thetopic, includingthat
size and aggregatesize, and it isproposed that conver- on ‘small’ cores.
sion to corresponding cube strengths should take this The principal factors which may cause differences
into account. The variability of results is also assessedin behaviour between ‘small’ and ‘large’ coresare as
in relation to the above factors, and an estimate made follows.
of the accuracyofpredicted actual cube strength thatis (1) Effects of size of specimen
likely to be achieved from testing cores of this size. Research by Ne~ille‘~’ and many other inves-
tigatorssuggeststhatmeasuredconcretestrength
Introduction generally increases as the size of the test specimen
The need forcompressive testing of cores to yield decreases, and that results tend to be more variable
an estimateof the strengthof suspect concreteis well with small specimens.The lattereffect has been con-
established, and recommendations for such tests are firmed for core specimensby Henzel and Freitag‘4).
contained in BS 1881: Part4:1970‘”. Also, the Con- (2) Effects of cutting
crete SocietyTechnical Report No. 11 of 1976(2) The ratioof cut surfacearea to volume increasesas
provides considerably more detailed evidence and core diameter decreases, hence the potential influ-
recommendations for both testing and interpretation ence of drilling damage upon measured strengthwill
of results. Both the above documents are based on be greater with ‘small’ cores.
cores of 150 or 100 diameter. However, it is fre- ( 3 ) Relationship between size of aggregate and
quently found to be totally impracticable to obtain diameter of core
cores of this diameter with the required minimum This will be more critical with ‘small’ cores, which
heighvdiameter of 1.0. This may be due either to may commonlyhave aggregate-sizelcore-diameter
limitations of member dimensions or to critical rein- ratios in excess of the suggested limitof 113, which is
forcement locations,and is especially relevant topre- recommended by theAmerican”), Germad6’ and
stressed concrete construction. Consequently, cores Australian(’) Standards. Where the aggregate parti-
of a considerably smaller diameter are often used, cles are large in relation to the size of the specimen,
despite very limited evidence of their reliability. It the effects of any aggregate loosened by cutting will
was, therefore, considered worth while to undertake be increased. Furthermore, the homogeneity of the
91
Downloaded by [ TUFTS UNIVERSITY] on [23/09/16]. Copyright © ICE Publishing, all rights reserved.
 Magazine of Concrete Research : Vol. 31, No. 107 :June 1979

material in the test specimenis effectively reduced in
comparison with alargerspecimen,and this may
influence the internal failure characteristics.
Since theyare interrelated,differentiation between
individualeffects will be difficult. For example,
increased strength attributed to small specimen size
may be offset by a reduction due to greater cutting
effects. Cutting damage may furthermore be influ-
enced by aggregatecharacteristics,andeach may
separately influence the mode of failure and the var-
iability of results.
Kesler(’) has shown thatconcretestrength is a
further factorwhich may influence the behaviour of a
core, and it is possible that this also may affect the
relative behaviour of ‘small’ and ‘large’ cores. The
complexity of these problems contributedTo the deci-
sion to confine the investigation to a single ‘small’
core size, and 44 mm diameter was chosen as being
typical of ‘small’ cores used in practice.
Advantage was taken of laboratory conditions to
standardize, as far as possible, many of the variables
which are known to affect the measured strength of a
core (i.e. compaction, reinforcement, drilling techni-
que, capping, moisture condition and testing techni-
que). Thetwo variableswhich willhave a majoreffect
upon the measured strength of a ‘large’ core are the
heighddiameter ratio(hld) and the orientationof the
direction of cutting and testing relative to that of
casting. Attention was, therefore, concentrated upon
examining the effectsof concrete strength and aggre-
gate size upon theserelationships andthe subsequent
correlations between core strength and cube strength
for the ‘small’ cores.
The differencesbetweenPotentialStrengthand
Actual Strength of concrete are well documented(*),
and have not been considered in this investigation,
since the relationship is not directly influenced by the
size of test specimen.
Details of test programme
A total of 23 mixes was used in the investigation,
covering a rangeof measured cube strengths between
10 and 82 N/mm*, and these are listed in Table 1.
Portland cementswere generally used, although
three mixes were of high alumina cement concrete,
and coarse aggregates were 10 mm or 20 mm max-
imum size irregular gravels. In all cases, 100 X 100 x
500 mm unreinforced prisms were cast and cured in
the laboratory, together with at least four 100 mm
control cubes. Nominal 44 mm diameter cores were
cut from the prisms by using a diamond-tipped core-
cutter, andwere trimmed and capped with a thickness
of up to2 mm of high alumina cement mortar togive
over-all heighvdiameter ratios between 1.0 and 2.0.

T A B L E 1: Summary of testspecimens.
Measured Maximum Total number
cube aggregate of cores
Mix strength size Type of h id
No. cement of cores Orientation*
(N/mm2) (mm) V
1 82 10 HA 1 2, 2.0 6
2 82 20 HA 1.2, 1.7, 2.0 6
3 75 20 HA 1.2, 1.7, 2.0 24
4 23 20 OP 1-2, 1.6,2.0 24
5 41 20 OP 1-2, 1-6, 2.0 24
6 75 20 OP 1.2, 1.6, 2.0 24
7 34 20 RHP 1.4, 1.8, 2 0 24
8 34 20 RHP 1.2, 2-0 8
9 21 20 RHP 1.2, 2.0 16
10 68 20 RHP 1.2, 2.0 16
11 33 10 RHP 1.5, 1.8, 2.0 20
12 33 10 RHP 1 2, 2.0 8
13 23 10 RHP 1 5 , 1.8, 2.0 24
14 23 10 RHP 1.2, 2.0 8
15 10 10 RHP 1-2, 1-8, 2.0 16
16 45 10 RHP 2.0 -
17 36 10 RHP 2.0 -
18 31 10 RHP 2.0 -
19 18 10 RHP 2.0 -
20 43 20 RHP 2.0 -
21 33 20 RHP 2.0 -
22 36 20 RHP 2.0 -
23 24 20 RHP 2.0 -
~

Totals 248

*See Figure 1

92
Downloaded by [ TUFTS UNIVERSITY] on [23/09/16]. Copyright © ICE Publishing, all rights reserved.
 Determining concrete strength by using small-diameter cores

vertically
drllled The cores weretested in compression immediately
after removal from water, with loading applied at a
directlon rate of 15 N/mmz min, as recommendedby BS 1881,
of casting
Part 4(l).
Those with an anticipated measured strengthof 40
N/mm2 or less were tested in a 6.6 tonne manually
operated Denisontesting machine,whilst a 100 tonne
capacity Avery machine was used for strongerspeci-
mens. The failureof each core was examined visually
to ensure that there was no significant cracking or
damage to the caps and that the failure mode was
horlzontally drilled symmetrical.Caseswherediagonal shearfailure
occurred were accepted for long cores, as recom-
mended by the Concrete Society Report(z), butwere
Figure I : Drilling orientation in relation to prism specimen. rejected if hld was less than 1.4.

Analysis and discussion of experimental

Although the majorityof cores were cut ain ‘vertical’ results
direction as defined in Figure 1, a number were cut
‘horizontally’ relative to the direction of casting. INFLUENCE OF HEIGHT/DIAMETER
Cores were not cutuntil long enough aftercasting R A T I O (hld)
to allow effective stabilization of concrete strength, For each mix, the average values of measured core
and thus minimize variations due to curing differ- strengths, with a specific orientation, were compared
encesbetween cores,
after
cutting,
and
cubes. for each differentvalue of hld and expressedin terms
Minimum agesof 14 days for rapid-hardening and28 of a core with hld = 2.0. The averages wereall based
days for ordinary Portland cements were considered on the resultsof at least fourindividual similar cores,
adequate for this purpose. although a number of prism specimens were often
A specially designed frame was used to hold the involved and thus the variability of results will be
prismsduringcutting, to limit relativemovement partly due to variations of concrete between speci-
between the specimen and rig, and cores werestored mens. The results of this analysis are illustrated in
under water for at least 48 h prior to compressive Figure 2, and show a very large scatter. Neverthe-
testing, to standardize moisture contents. Since the less an over-all least-squares regression line K = 0.54
cubes were in every case cured with the prisms and + 0.23 (hld) is shown. Although a distinction has
cores, the cube strengths obtained do not represent been made betweenthe two basic types of cement in
standard28 day strengthsbutare actual cube this Figure, separate analysis does not yield any sig-
strengths which relate as closely as possible to the nificant influence of this variable.Similarly, although
strength of the concrete in the cores at the time of not illustrated, the effect of drilling orientation was
testing. also found not to be significant in this respect.

IO

Y
m B S 1881
I
l-
W
z
W
a
=; 09
W
a
0
V
n
W
a K = 0.54 + 0.23(h/d)
3
m
5 0 8
I 0 HAC, 2 0 mm gravel
U
0
V HAC, 10 mm gravel
0
l- A Portlandcement. 20 mmgravel
4
IT
0 0 Portland cement, 10 mm gravel
l 1 I I
1.8 1 1.6 1.2 1.4 2
HElGHTiDlAMETER RATl0,hid

Figure 2: Influence of heightldiameter ratio.

Downloaded by [ TUFTS UNIVERSITY] on [23/09/16]. Copyright © ICE Publishing, all rights reserved. 93
 Magazine of Concrete Research : Vol. 31, No. 107 :June 1979

Maximum aggregate size is also indicated in Figure Whilst this can be due to acombination of factor^'^),
2, and separateanalyses give K = 0.57 + 0.215 (hld) since similar casting and testing procedures used were
+
for 20 mm, and K = 0.50 0.25 (hld) for 10 mm it is likely to reflect differing variability of the con-
results. The scatter must limit the reliability of these crete within individual batches. Thusthe ratio (coef-
expressions; however, the corresponding correction ficient of variation of cores)/(coefficient of variation
factors to a standard hld = 2.0 are summarized in of cubes) is more likely to represent differences in
Table 2 and compared with the recommendations of behaviour between the various typesof concrete due
references 1 and 2. Although the strength of 10 mm to core testing alone.
aggregate coresis apparently more influenced by var- In Table 4, average coefficients of variation are
iations of hld than is that of the 20 mm aggregate shown according to cement type, aggregate size and
specimens, it will be seen that the scatter on observed core orientation. Examinationof the corekube ratio
values is so great thatdiscrepancies in corrected core suggests that 20 mm aggregate cores exhibit a greater
strength dueto adopting an average correction factor variability due to cutting and testing than 10 mm
which is independent of aggregate size will be rela- aggregate cores.The scatter of results is such that this
tively unimportant. is notprovedconclusively:nevertheless,further
weight is given to this view by the appreciable differ-
V A R I A B I L I T Y O F TEST R E S U L T S ences in standard deviations of this ratio between 10
Coefficients of variation for each set of 'identical' mm and 20 mm aggregates.Orientationseemsto
vertically drilled cores of Portland cement concrete have littleinfluenceuponthevariability of core
are summarized in Table 3. It will be observed that results. However, it can be seen from Table 4 that
there is no significant change in variabilityof results cores from the high alumina cement mixes appear to
between the two extreme values of hld, irrespective bemore variable thanthosefrom the Portland
of aggregate size. Resultsfor 10 mm aggregate cores cement mixes.
show a slightly greater scatter than 20 mm results.
However, this may be misleading in view of the dif-
ferencesobtainedforcoefficient of variation of T A B L E 3: Effectof (hld) uponcoefficientof
measured cube strength between mixes (Table 4). variation of measured core strength (Portland
cement concrete).
Average coefficient of variation
T A B L E 2: (hid) correctionfactors. ('vertical' cores)

Aggregate

1
95 %
hld =
over-a" confidence
mean
10 mm gravel 1.2 11.4-1.8 12.0 limits
'Small' cores
20 mm gravel
20 mm gravel
(7 mixes)
1 1 7.8 18.1
7.3 I 7.5 2 4.7

'Large' cores
Concrete Society'2'
BS 1881, Part 4") I 0.80
0.92 I 1.0
1.0
10 mm gravel
(5 mixes)
1 I
I
7-3
I
7-1 19.81 8.8 -t 6.0

TAB LE 4: Summary of coefficients of variation of test results.

Average coefficient of variation
(%l
Type of cement Orientation Number
and of of Standard
size of aggregate core* mixes Corelcube deviation of
Cores Cubes
ratio corelcube
ratio

PORTLAND CEMENT
V 7 7.5 5.5 1.8 1.3
20 mm gravel
H 5 5.8 3.4 1.7 1.2
10 mm gravel V 5 8.8 7.0 1.3 0.6
H 6 9.5 6.4 1.3 0.5

HIGH ALUMINA CEMENT

20 and 10 mm gravel V and H 3 10.8 4.1 2.7 1.2

*See Figure 1.

94
Downloaded by [ TUFTS UNIVERSITY] on [23/09/16]. Copyright © ICE Publishing, all rights reserved.
 Determining concrete strength by using small-diameter cores

M E A S U R E DS T R E N G T H S strengths for Portland cement concrete are shown in

In practice, strength assessments are likely to be
based on at least four similar cores, andthe results of
this investigation have generally been assessed on this
basis.Nevertheless, to illustratethevariability of
individual results, Figure3 shows values obtainedfor
single ‘horizontal’ Portland cement cores with 10 mm
aggregate.
Further comparisons of measured core and cube
Figures 4 and 5 , for ‘horizontally’ drilled and ‘verti-
cally’ drilled cores respectively. In each case, core
strengths have been corrected to a standardhld = 2.0
by using the relationship K = 0.54+ 0.23 (hld)
establishedabove, and eachpointrepresentsthe
mean of four individual similar cores. A distinction
has been madebetween the two maximum aggregate
sizes, and least-squares straight-line relationships are

60
60 l
A 10 mm gravel - 0 20 mm gravel
k 50 2 50 -- A IO mm gravel /
E E
2 / 2
I / I
I I
t- 40 t- 40
c7 0
z z
W W
U U
t- cube = 1 14 x core t-
m (0

W 30
\\
W 30
U U
0 8A 0
V V
D n
p 20
A p 20
V V \ cube = 1.22x core
W
U
U
W
U
U
A??
0 0
V 10 0 10
/
/
/
J
0 10 20 30 40 50 60 0 10 20 30 40 50 60
M E A S U R E D CUBE STRENGTH-N/rnrnZ M E A S U R E D C U B E STRENGTH-N/mrn2

Figure 3: Relationship between corrected core strength (single Figure 4: Relationship between corrected core strength (mean of
cores, hld = 20) and measured cube strength (Portland cement 4, hid = 2.0) and measured cube strength (Portland cement
concrete) for ‘horizontal‘ cores. concrete) for ‘horizontal’ cores.

70
-
60
-- 0 2 0 mm gravel
A 10 mm gravel 80 0

E
Q
E
0

I
40 Q
r
t-
a
z
Kc 50
m
W . cube = 1.25 X core
U
g 30
n
W
c
U
W
U 20
U
0
V

10

0 10 20 30 40 50 60 70 80 90
MEASURED CUBE STRENGTH-NImm’

Figure 5: Relationship between corrected core strength (mean of 4, hld = 2.0) and
measured cube strength (Portland cement concrete) for ‘vertical’ cores.

95
Downloaded by [ TUFTS UNIVERSITY] on [23/09/16]. Copyright © ICE Publishing, all rights reserved.
 Magazine of Concrete Research : Vol. 31, No. 107 : June 1979

O' c- 1

60
- /l A Portlandcement (20mm aggregate)
0 highaluminacement i

E
E
2 50 .
I
I
l-
U
z
E 40 cube = 1.28 x core
F
m
W
U
95% limits for
30 . Portland cement cores
0
W
F
V
W
U 20
U
0
0

io

l I I l l I I I l
0 io 20 30 40 50 60 70 EO 90
MEASURED CUBE STRENGTH-N/mm2

Figure 6: Relationship between corrected core strength (mean of 12, hld = 2.0) and
measured cube strength for 'vertical' cores.

shown for each size. These cubelcore strength corre- that the core strengths are significantly lower than
lations are summarized in Table 5 , together with the would be expected forsimilar Portland cement cores.
corresponding 95% confidence limits on calculated
cubestrengths. To permitrealisticcomparison of General observations, and comparison of
results, mixes with a very high cube strength (<50 results with existing recommendations
N/mm2) have not been included in calculating these The resultsclearlydemonstrate that for 'small'
relationships. It was found, nevertheless, that for the cores (hld)effects are considerably greater than those
vertically drilled 20 mm aggregate cores alone, the suggested by BS 1881"). Comparison of the experi-
conversion factor is only marginally changed from mental results with the more recent Concrete Society
1.25 to 1.28 with 95% limits of 2 2 2 % when the recommendations(Table 2) suggests that 'small'
higher strengths are included. Figure 6 is based on cores may be marginally more sensitive tohld varia-
means of 12 'vertical' 20 mm aggregate cores, and tions than large cores, but there is no evidence that
shows thecorresponding 95% confidencelimits the over-all variability of core results is affected by
which are reduced to t 12% in that case. The results the hld ratio used.
for H A C mixes are also plottedin Figure 6 and show The averagecoefficient of variation for core results

T A B L E 5: Cubekorrected-core conversion factors (cube strength < 50 N/mm2).

~ ~ ~~~~

Maximum Experimental results Concrete Society

Orientation aggregate recommendations"'
size 10 mm 20 mm Combined (for large cores)
Conversion factor 1.25 1.05 1.15 1 1.15
'Vertical' 95% limits on
predicted cube
? 17% 223% ?23% t 6%
strength
(mean of 4 cores)

1 Conversion factor 1 1.14 1 1.22 1 1-17 1 1-25

'Horizontal' 95% limits on

predicted cube
?15% ? 17% ? 17% -.6%
strength
(mean of 4 cores)

96
Downloaded by [ TUFTS UNIVERSITY] on [23/09/16]. Copyright © ICE Publishing, all rights reserved.
 Determining concrete strength by using small-diameter cores

of approximately 8% compares well with that onactualcubestrength will bebetterthanthose

expected‘*’, although the range is large and no sig- indicated in Table 5. These are considerably worse
nificant difference can be detected between ‘horizon- than the values suggested for ‘large’ cores, being at
tal’ and ‘vertical’ cores. The indication that 20 mm least three times theConcreteSociety‘2’value of
aggregate coresmay be more variable than those with &(1 2 / d n )%, where n is the number of cores aver-
10 mm aggregate is likely to reflect weakening of the aged.Directapplication of theConcreteSociety
cut surface. Although damage may not be visible, any recommendations‘2’ to this set of results for ‘small’
pieces of aggregate which are cut, and thusonly par- cores would havepredictedactualcubestrengths
tially embedded in the matrix, may influence the fail- which were up to 30% different from the measured
ure of the core and it is to be expected that this will value. A similar analysis based on BS 1881, Part 4,
become more significantas the ratio of aggregate size yielded estimated cube strengthswhich are on aver-
to core diameter increases. HAC cores show a par- age 12% too high, and range between 50% above and
ticularly high variability, which may be dueto 20% below the measured value. Although only one
deterioration of the core matrix as a result of conver- typical ‘small’ core size has beenused here, theresults
sion(’O)being accelerated by drilling and subsequent demonstrate clearly thatexisting methods for obtain-
curing of the cores. This view is supported by the ing cube strengths from ‘large’ diameter cores must
lower than expected core strengths obtained (Figure be treated with great cautionwhen smaller diameters
6). Measurements of percentage conversion werenot are used.
made, but it is expected that this would have been Frominformation given by Warren(9) it can be
fairly low in the cube and prism specimens, and the shown that, on thebasis of the cube results,the con-
concrete may thus have been vulnerable to the heat crete used in this investigation is ‘good‘ or ‘average’ in
generated in the surface zone by the drilling. These terms of variability. Reduction of quality to
results,althoughfew in number,suggestcaution ‘mediocre’, asmay occur on site, couldpossibly widen
where HAC is involved, and it would seem that this the range of 95% confidence limits on actual cube
problem should be examined more fully. strengths predicted from means of four site cores by
The strengthof cores with 10 mm aggregate follows an estimated 2 5 % . This assumes no difference in
theanticipatedpattern,‘horizontal‘coresbeing cuttingproceduresbetweensiteandlaboratory,
approximately 10% weaker relative to cubes than which may be optimistic. In interpreting site results,
‘vertical’ cores. However, the
corrected
core other factors such as voids and reinforcement must
strengths are10% stronger than would be anticipated also be considered, and these may further influence
from theConcrete Societyrecommendationsfor strength predictions. Caution must thusbe exercised
‘large’cores(Table 5). Thisdiscrepancy will be when attempting to assess the accuracy of actual con-
further increased by a small amountdue to the differ- crete cube strength predictions from site-cut cores,
ences between h/d correction factors (Table 2), but and the range of confidence limits with cores of this
would seem tobe primarily associatedwith the reduc- size is unlikely to be smaller than suggested by this
tion in specimen size. laboratory investigation, even under good conditions.
The 20 mm aggregatecores are considerably The sensitivity of ‘small’ cores to the many factors
weaker relativeto cubes than10 mm aggregate cores, outlinedabovemustmeanthatthevariability of
which tends to confirm the anticipated influence of results is likely to beso large that theuse of such cores
the (aggregate size)/(core diameter) ratio discussed to assess the strengthof in situ concreteis of doubtful
above. Thisis further supported by the greatervaria- value in many practical situations.
bility of results, and is reflected in the wider band of
95% confidence limits on predicted cube strength.
The anticipated orientation effects are not apparent Conclusions
for 20 mm aggregateconcretes,butthestrengths For ‘small’ cores of 44 mm diameter, thefollowing
obtained from vertically drilled cores are lower and conclusions may be drawn.
show a greater scatter than would be expected. This ( 1 ) The effect of height/diameter ratio is considera-
may possibly be attributedto greater variationsof the bly greater than that indicated by BS 1881, Part 4‘”,
larger aggregate concrete near ‘top’ surfaces, even in and is closer to, but still marginally greater than that
laboratory specimens, and serves to emphasize the proposed by the Concrete Society Report‘*’ relating
need in practice to avoid taking cores from concrete to ‘large’ cores.
from the topof pours. In thelight of these resultsit is (2) The average coefficient of variation associated
suggested that, for ‘small’ cores of this diameter, 10 with a set of Portland cement concrete cores drilled
mm and 20 mm maximumaggregatesshould be and tested under laboratory conditions is of the order
treated separately when attempting to convert core of 8%, and is not significantly influenced by
results to cube strengths.If the experimental correla- height/diameter ratio.
tions for ‘vertical’ or ‘horizontal’ cores are used as (3) Where 10 mm maximum aggregates are used, the
appropriate, it is unlikely that 95 % confidence limits relative orientations of casting and testing have an

Downloaded by [ TUFTS UNIVERSITY] on [23/09/16]. Copyright © ICE Publishing, all rights reserved. 97
 Magazine of Concrete Research : Vol. 31, No. 107 :June 1979
effect upon measured strengthwhich is similar to that
anticipated for larger cores.
(4) The measured strength of cores with 10 mm max-
imum aggregate size is approximately 10% greater
relative to the actual cube strength than would be
expected for coresof 100 or 150 mm diameter. This
was not found to apply with 20 mm aggregates for
which ‘vertically’ drilled cores were approximately
10% weaker than anticipated.
(5) The strengths of cores from ‘uncoverted’ HAC
concrete may be lower in relation to cubes, and more
variable, than would be expected forsimilar Portland
cement cores.
(6) The sensitivity of ‘small’ cores tomany factorsof
mix properties and testing procedure leads to a high
variability of strength predictions, and methodsused
for estimating cube strengths from‘large’ cores can-
not be relied upon under these circumstances.
(7) Prediction of actual cubestrength should account
for both orientation of cores and aggregate size, in
accordance with Table 5. In this event, for a set of n
results the 95% confidence limits are unlikely to be
betterthan k ( 3 6 / d n ) % underlaboratory condi-
tions. The number of additional influences associated
with site conditions suggests that the use of ‘small’
cores toassess the equivalent cube strengthof in situ
concrete may be of doubtful value.
REFERENCES
1. BRITISH STANDARDS INSTITUTION. Methods of testing concrete
for strength. London. pp. 25. BS 1881:Part 4:1970.
98
Downloaded by [ TUFTS UNIVERSITY] on [23/09/16]. Copyright © ICE Publishing, all rights reserved.
2. CONCRETE SOCIEIY, THE. Concretecore testing for strength.
London, 1976. pp. 44. Technical Report No. 11. Publication
51.071.
3. NEVILLE, A. M . The influence of size of concrete test cubes
on mean strength and standard deviation. Magazine of Con-
crete Research. Vol. 8, No. 23. August 1956. pp. 101-110.
4. HENZEL, J. and FREITAG, W. Zur Ermittlung der Betondruck-
festigkeit im Bauwerk mit Hilfe von Bohrkemen kleineren
Durchmessers. (The determination of the compressive
strength of concrete in a structure with the aid of test cores
of small diameter.) Beton: Herstellung, Verwendung. Vol. 19,
No. 4. April 1969. pp. 151-155.
5. AMERICAN SOCIETY FOR TESTING AND MATERIALS. Standard
method of obtaining and testing drilled cores and sawed beams
of concrete. Philadelphia, 1968. Part 14. pp. 32-36. C42-68.
6. DEUTSCHEN NORMENAUSSCHUSS. Priifierfahren f i r Beton:
Bestimmung der Druckfestigkeit von Festbeton fertiger Bau-
werke und Bauglieder. (Testing methods for concrete: com-
pressive strength of concrete in structures and members.)
Berlin. p. 1. DIN 1048: Part 2: 1972.
7. STANDARDS ASSOCIATION OF AUSTRALIA. Method for securing
and testing cores from hardened concrete forcompressive
strength or indirect tensile strength. Sydney. pp. 12. AS 1012.
Part 14:1973.
8. KESLER, c. E. Effect of length to diameter ratio on compressive
strength - an ASTh4 co-operative investigation. Proceedings
of the American Society for Testing and materials. Vol. 59.
1959. pp. 12161229.
9. WARREN, P. A . Assessing the validity of the cube test result.
Egham, RMC Technical Services, 1976. pp. 14. RMC
Technical Report No. 78.
10. TEYCHENNE, D. c . Long-term research into the characteristics
of high alumina cement concrete. Magazine of Concrete
Research. Vol. 27, No. 91. June 1975. pp. 7%102.
Contributions discussing the abovepuper should be in the hands of
the Editor not later than 31 December 1979.