Shear transfer in

lightweight reinforced
concrete
Alan H. Mattock
Professor of Civil Engineering and
Head, Structures and Mechanics Division
University of Washington
Seattle, Washington

W. K. Li*
Structural Engineer
Surveyor, Nenniger and Chenevert, Inc.
Consulting Engineers
Montreal, Quebec

T. C. Wang*
Structural Engineer
Kramer, Chin and Mayo, Inc.
Consulting Engineers
Seattle, Washington

*Former Graduate Student, Department of Civil Engineering,

University of Washington, Seattle, Washington.

20
L ightweight concrete is being used
increasingly in precast concrete
construction. This has resulted in a
Synopsis
need for connection design data when A study is reported of the
lightweight concrete is used. Neither single direction shear trans-
the ACI Building Code, ACI 318-71,1 fer strength of lightweight ag-
nor the PCI Design Handbook 2 pro-
gregate concrete.
vides any guidance in this respect.
Push-off tests were carried
The shear-friction provisions con-
tained in Section 11.15 of ACI 318-71, out on specimens made from
which are used extensively in the con- sanded lightweight concrete,
nection design procedures developed in two types of all-lightweight
the PCI Design Handbook, have only concrete, and sand and grav-
been validated experimentally for the
el concrete.
case of shear transfer in normal weight
concrete.
Both initially uncracked
The study reported here was directed specimens and specimens
toward developing shear transfer design cracked in the shear plane
recommendations for use in the design before being subjected to
of connections in precast structures shear, were tested.
made of lightweight concrete. It was found that the shear
transfer strength of light-
Experimental Study weight concrete is less than
that of sand and gravel con-
The experimental program reported
here was designed to study the crete having the same com-
influence on single direction shear pressive strength.
transfer strength and behavior, of The shear-friction provisions
the type of aggregate used in making of Section 11.15 of ACI 318-
the concrete. The primary variable 71 may be used in the de-
in the tests was the type of aggre-
gate, four types being used: sign of connections in light-
1. Naturally occurring gravel and weight concrete providing
sand. the value of the coefficients
2. A predominantly coated rounded of friction fc, contained in
lightweight aggregate. Section 11.15.4, are multi-
3. A predominantly crushed angular plied by the following factors:
lightweight aggregate.
(a) For all-lightweight con-
4. A "sanded lightweight" aggregate,
in which most of the lightweight crete having a unit weight
fine particles were replaced with not less than 92 lb per
normal weight sand. cu ft, multiply µ by 0.75.
The test specimens were of the (b) For sanded lightweight
"push-off" type shown in Fig. 1, with a concrete having a unit
shear plane of 50 sq. in. area. When
weight not less than 105
loaded as indicated by the arrows, shear
without moment is produced in the lb per cu ft, multiply µ
shear plane. The reinforcement cross- by 0.85.
ing the shear plane was in the form of

PCI JOURNAL/January-February 1976 21

 •
Is closed
..ҟ •

IRIGIHi
• till hf A

-I.
D

N N N N
* 1

r r r r

L i Section A—A

!~- 12 ^I

Fig. 1. Push-off specimen.

welded closed stirrups. This was to en- Materials and fabrication

sure the effective anchorage of the rein-
forcement on both sides of the shear When planning the project, the advice
plane. The specimens were cast on their of the Expanded Shale, Clay and Slate
sides, so that at the time of casting the Institute was sought as to the types of
shear plane was horizontal. aggregate and concrete mix designs that
Ten series of push-off specimens should be used in order that the results
were tested, as indicated in Table 1. of the study would have the widest
The variables between test series were practical applicability.
the type of aggregate, the concrete Current production of rotary kiln ex-
strength, and whether or not a crack ex- panded lightweight aggregate is pre-
isted in the shear plane before the shear dominantly of the coated surface type
transfer test. The variable within each in the coarse fraction and a blend of
series was the reinforcement parameter, crushed and coated surface aggregate
p/2,. The number of stirrups provided in the fine fraction.
in the specimens within each series is For the sake of brevity, lightweight
set out in Table 2. aggregate having predominantly coated

22
 Table 1. Program of push-off tests.

Series Aggregate Type Design Initial

f at test Condition
c (psi)

A Coated lightweight aggregate 4000 Uncracked

and sand

B Coated lightweight aggregate 4000 Cracked

and sand

C Coated lightweight aggregate 2500 Cracked

and sand

0 Coated lightweight aggregate 6000 Cracked

and sand

j E Coated lightweight aggregate 4000 Untracked

only

F Coated lightweight aggregate 4000 Cracked

only

G Crushed lightweight aggregate 4000 Uncracked

only

H Crushed lightweight aggregate 4000 Cracked

only

M Natural gravel and sand 4000 Uncracked

N Natural gravel and sand 4000 Cracked

Table 2. Specimen numbering and reinforcement details.

Specimen No. X.0* X.l X.2 X.3 X.4 X.5 X.6

Number of
#3 Stirrups 0 1 2 3 4 5 6

Reinforcement
Area, sq in. 0 0.22 0.44 0.66 0.88 1.10 1.32

* This specimen appears only in Series A, E, G, and M.

"X" is the letter designating the particular series.

rounded aggregate particles, will in this lightweight aggregate concrete made of

report be referred to as a "coated ag- a "coated aggregate" in "sanded light-
gregate." Similarly, aggregate in which weight" mixes.
the various particle sizes are predom- The first four test series (Series A, B,
inantly obtained by crushing larger- C, and D), therefore involve this type
sized particles is referred to as of concrete. The second group of four
"crushed aggregate." Most structural test series (Series E, F, G, and H) in-
lightweight aggregate concretes use volved "all-lightweight" concrete, Se-
mainly sand for the fine fraction of the ries E and F using a "coated aggregate"
aggregate. Such concretes are referred and Series G and H using a "crushed
to as "sanded lightweight" concretes aggregate." The final two series (Series
and have a dry density of about 110 lb M and N) used a natural gravel con-
per cu ft. crete as a reference against which to
In view of the foregoing, it was de- compare the performance of the light-
cided that the results of the study weight concretes.
would have the widest applicability if The natural aggregate used was a
emphasis were placed on tests involving glacial outwash gravel obtained from a

PCI JOURNAL/January-February 1976 23

 Fig. 2. Sample of coated
lightweight aggregate.

Fig. 3. Sample of
crushed lightweight ag-
gregate.

Fig. 4. Sample of natural

sand and gravel.

local pit. The gradings of the coarse ture contents were not measured.
and fine fractions of each aggregate as The proportions of the lightweight
determined by sieve analysis, are set concrete mixes were chosen so that for
out in Table 3. The appearance of the a 3-in, slump the design strength would
aggregates is seen in Figs. 2, 3, and 4. be attained at 28 days, after 7 days
Trial batches were made of the light- moist curing and a further 21 days cur-
weight concretes, using proportions ing in air. The proportions of the sand
suggested by the producers of the ag- and gravel concrete were chosen so that
gregates. The mix proportions selected for a 3-in, slump the design strength
for the various lightweight aggregate would be attained at 4 days, after 2
concretes are given in Table 4. The mix days moist curing and 2 days curing in
proportions for the natural gravel con- air. Approximately 6 percent of en-
crete were arrived at through the test- trained air was included in all mixes
ing of trial batches and they are shown to conform to usual practice with light-
in Table 5. The weights shown in Ta- weight concrete. Complete details of
bles 4 and 5 are of aggregate as sup- the procedures followed are reported
plied, i.e., slightly damp. Exact mois- elsewhere.4

24
 Table 3. Grading of aggregates
(cumulative percent retained).

Aggregate Coated Crushed Natural

Lightweight Lightweight Sand and
Sieve Gravel
Size

Coarse Aggregate

3/4 in. 0 0 0

1/2 in. 0 19.4 53.8

3/8 in. 2.7 61.6 87.0

No. 4 92.5 96.4 99.9

Fine Aggregate

No. 4 1.8 0.4 1.8

No. 16 43.1 43.7 26.3

No. 50 81.9 74.2 76.2

No. 100 93.8 86.9 96.9

Table 4. Mix proportions for lightweight concretes

(lbs per cu yd)
Type I Lightweight Aggregate
Type of Concrete Portland Natural
Cement Coarse Fine Sand

Sanded Lightweight
coated aggregate)

Series A & B 500 740 227 1369

(f'C - 4000 psi)

Series C 400 744 220 1426

(f 2500 psi)

Series D 640 784 220 1091

(f = 6000 psi)

All-Lightweight

Series E & F 500 740 1049 --

(coated aggregate,
V = 4000 psi)

Series G & H 517 825 1284 --

(crushed aggregate,
4000 psi)

Note: In all mixes, water was provided to produce a 3-in. slump.

Table 5. Mix proportions for natural sand and

gravel concrete (lbs per cu yd)
Type III
Concrete Portland Gravel Sand
Cement

Series M & N 489 1798 1448

(fc = 4000 psi)

Note: - In all mixes, water was provided to produce a 3-in. slump.

PCI JOURNAL/January-February 1976ҟ 25

 Table 6. Data concerning test specimens.
Concrete
Specimen Stirrup Reinforcement f. (1) f ct (2) Dry Shear Stress
No. Yield Point Parameter Density (3)
(psi) (psi) (psi)
f (ksi) Pf (psi) (lb. /cu. ft) u

Series A

AD -- 0 4230 402 111 500

Al 47.7 210 3740 336 111 758
A2 53.6 472 4095 367 105 914
A3 53.2 702 3910 349 110 1020

A4 50.9 896 4100 352 108 1100

A5 50.9 1120 3960 351 108 119D
A6 51.8 1368 4250 397 110 1344

Series B

81 49.6 218 3740 336 111 450

B2 50.9 448 3360 318 107 652
83 4 50.9 672 3910 349 110 840
B4 i 49.1 864 4100 352 108 940
85 50.5 1111 3960 351 108 1000
B6 51.8 1368 4250 397 110 1154

Series C

Cl 49.6 218 2330 254 102 364

C2 53.6 472 2330 254 102 514
C3 50.9 672 2000 232 103 526
C4 52.3 921 2050 235 105 560
C5 53.6 1179 2330 269 106 640
C6 49.6 1309 2330 269 106 740

Series 0
D1 51.8 228 5995 376 108 370

D2 52.3 460 5995 376 108 668

D3 52.3 690 5710 379 10 772

04 52.3 920 5710 379 107 1022

D5 52.3 1151 5600 398 109. 1082

06 51.8 1368 5600 398 109 1220

Series E
EU -- 0 3960 365 92 560

El 52.3 230 4150 350 97 780

E2 52.3 460 4030 355 94 872

E3 52.3 } 690 4065 375 96 960

E4 53.2 936 4040 405 96 1150

E5 50.5 1111 4115 365 98 1200

E6 52.3 1381 4050 365 94 1250

Notes: (1) Concrete compressive strength at time of test, measured on 6 x 12-in.

cylinders.

(2) Concrete splitting tensile strength at time of test, measured on

6 x 12-in. cylinders.
V
(3) v = u , with = 1.0.
u 0 Acr

(4) Air dried density at time of test.

26
 Table 6. (cont.). Data concerning test specimens
Concrete( 4 ) Ultimate
Specimen Stirrup Reinforcement f^ (1) f ct (2) Dry Shear Stress
No. Yield Point Parameter Density (3)
(psi) (Psi) v (Psi)
f. (ksi) pf (psi) (lb./cu. ft)

Series F
Fl 53.2 234 4150 350 97 450

F2 52.3 460 4030 355 94 530

F2A 50.9 448 3970 355 94 620

F3 52.3 690 4065 375 96 734

F3A 51.4 678 3970 355 94 702

F4 50.9 896 4040 405 96 870

F5 51.8 1140 4115 365 98 920

F6 53.2 1404 4050 365 94 982

Series G
GO -- 0 4030 420 98 530

G1 52.3 230 4145 395 98 820

G2 50.5 444 3880 378 96 846

G3 51.8 684 4100 371 96 1060

G4 53.2 936 4420 406 97 1150
G5 51.8 1140 4005 395 99 1140

G6 51.8 1368 4005 395 99 1190

Series H

Hl 49.8 219 4145 395 98 400

H2 51.8 456 3880 378 96 620
H3 51.8 684 4100 371 96 866
H4 51.8 912 4420 406 97 940
H5 50.5 1111 3950 395 99 990
H6 49.8 1315 4080 385 98 1042

Series M

MO -- 0 3935 375 145 590

M1. 50.9 224 4180 390 145 760
M2 52.7 464 3900 365 145 980
M3 52.3 690 3995 395 148 1110
M4 50.9 896 4150 400 145 1140
M5 52.7 1160 3935 350 146 1280
M6 52.7 1392 4120 395 145 1320

Series N

N1 50.9 224 4180 390 145 460

N2 52.7 464 3900 365 145 780

N3 52.3 690 3995 395 148 960

N4 50.9 896 4150 400 745 7150
N5 50.9 1120 3935 350 146 1175
N6 50.0 1320 4120 395 145 1190

Notes: (1) Concrete compressive strength at time of test, measured on 6 x 12-in.

cylinders.

(2) Concrete splitting tensile strength at time of test, measured on

6 x 12-in, cylinders.
v
(3) vu = - A , with = 1.0.
cr

(4) Air dried density at time of test.

PCI JOURNAL/January-February 1976ҟ 27

 Table 7. Average concrete properties at time of test
(by series).

Compressive Splitting f f Air

Strength, f. Tensile ct ct Dry
Series (psi) Strength, f Density
(psi) c (lb./cu. ft.)

Sanded Lightweight Concretes

Series A 4040 365 .090 5.74 109

Series B 3890 351 .090 5.63 109

Series C 2230 252 .113 5.34 104

Series D 5770 384 .067 5.06 108

All—Lightweight Concretes

Series E 4060 370 .091 5.81 95

Series F 4050 365 .090 5.74 95

Series G 4085 395 .097 6.18 97

Series H 4095 390 .095 6.09 97

Sand and Gravel Concrete

Series M 4020 380 .095 5.99 146

Series N 4045 385 .095 6.05 146

Type I portland cement was used for The rollers ensured that separation
the lightweight concrete mixes and of the two halves of the push-off speci-
Type III portland cement was used for men was not restrained by the testing
the sand and gravel concrete mixes. machine. Both slip along the shear
The deformed bar reinforcement plane and separation across it were
used conformed to ASTM Specification measured continuously using linear dif-
A615. The #3 bar used for the shear ferential transformers attached to refer-
transfer reinforcement had a yield ence points embedded in the specimen,
point of approximately 50 ksi. The #6 as may be seen in Fig. 5.
bar longitudinal reinforcement had a Mast3 pointed out the need to con-
yield point of approximately 60 ksi. sider the case of a crack existing in the
Reinforcement details and concrete shear plane before shear acts. There-
properties are shown in Tables 6 and 7. fore prior to test, some of the specimens
were cracked along the shear plane by
Testing arrangements and applying line loads to their front and
procedures rear faces.
The push-off specimens were tested These loads were applied through
using a Baldwin hydraulic testing ma- steel wedges with the specimen in a
chine to load the specimen along the horizontal position. The dilation of the
shear plane as indicated in Fig. 1. Typi- specimen across the shear plane was
cal arrangements for test are shown in measured during the cracking opera-
Fig. 5. The specimen stood on the low- tion, using dial gages mounted on a
er platen of the testing machine and reference frame. The width of the crack
was loaded through the spherically produced was approximately 0.01 in..
seated upper platen of the testing ma- The specimens were subjected to a
chine, a load cell and a set of parallel continuously increasing load until fail-
plates and rollers. ure occurred, with short pauses as nec-

28
 essary to mark any cracks which may
have occurred. The average length of
time taken for a test was about 15 min-
utes. The ultimate load was defined as
the maximum load that could be carried
by the specimen.

Specimen behavior
The general behavior of all the initially
uncracked specimens was similar. No
slip along the shear plane, nor separa-
tion across the shear plane occurred in
these specimens until the formation of
diagonal tension cracks in the region of
the shear plane at shear stresses of from
400 to 700 psi.
These cracks were initially 2 or 3 in.
long and inclined at from 20 to 45 deg
to the shear plane. As the load in-
creased, some of the cracks lengthened
and additional cracks formed, so that
at failure there were in general a larger
number of diagonal tension cracks in Fig. 5. Arrangements for test of push-
off specimen.
the more heavily reinforced specimens
than in the more lightly reinforced
specimens. shown in Fig. 6.
At failure some of the cracks propa- The general behavior of all the in-
gated parallel to the shear plane, link- itially cracked specimens was similar.
ing with others, and extensive compres- Slip occurred along the preformed
sion spalling occurred in the inclined crack in the shear plane from the com-
concrete struts formed by the diagonal mencement of loading, and at a pro-
tension cracks. gressively increasing rate. In the case
In the initially uncracked specimens of the heavily reinforced sand and grav-
there was no slip along the shear plane el concrete specimens, a few diagonal
in the true sense of the word. Relative tension cracks occurred across the shear
motion of the two halves of the speci- plane at high loads.
men occurred as a result of the rotation This behavior had been observed in
and compression of the inclined con- previous tests of sand and gravel con-
crete struts as the reinforcement cross- crete push-off specimens. 5 However, no
ing the shear plane stretched. such diagonal tension cracks were ob-
The component of this relative mo- served in either the sanded-Iightweight
tion parallel to the shear plane is re- or the aII-lightweight concrete speci-
ferred to as slip when discussing be- mens.
havior, for the sake of brevity. The In all the initially cracked specimens,
component of the relative motion nor- the slip increased at a rapid rate at fail-
mal to the shear plane is referred to as ure and a small amount of compression
separation. The slip and separation data spalling of the concrete occurred ad-
have been reported in detail else- jacent to the shear plane crack. Typical
where. 4 Typical shear-slip curves for shear-slip curves for initially cracked
initialIy uncracked specimens are specimens are shown in Fig. 7.

PCI JOURNAL/January-February 1976ҟ 29

 100
90 Series G Initially uncracked
All-lightweight, (crushed aggregate,) concrete.
80
. 70 /G5
V 60 // ZG4
/G3
0 50 ! 02
40
v
DC 30
a
a
< 20
1107 0.05 in.

0
Slip (in.)
Fig. 6. Typical shear-slip curves for initially uncracked specimens (Series G).

100
90 Series H Initially cracked
All-lightweight,(crushed aggregate,) concrete.
80
Q 70 H6
Y H5
60
H4
a 50 H3
H2
40
v H1
30 /
a
a
' 20

[0 0 5 in.,

7
Slip (in.)
Fig. 7. Typical shear-slip curves for initially cracked specimens (Series H).

It can be seen from the shear-slip the value at which maximum shear re-
curves that the deformation behavior sistance was developed.
was relatively brittle in all cases. The The initially uncracked specimens
maximum shear resistance was not behaved in a more brittle fashion than
maintained as the slip increased beyond the corresponding initially cracked spe-

30
 1600
Sanded lightweight, (coated aggregate,) concrete.
1400 fc = 4000 psi ; fy = 50 ksi

1 200 Series A - p/
Initially uncracked

1000 / q ^
Vu O /
(psi)
800
D /

600 \Series B -
/ / Initially cracked
400

200

0 200 400 600 800 1000 1200 1400

Qty (psi)
Fig. 8. Effect of a crack in the shear plane on shear transfer
strength of sanded lightweight concrete.

cimens, in that the shear resistance after Ultimate shear transfer strength
ultimate decreased more rapidly as the The values of ultimate shear transfer
slip increased. However, because the strength obtained in the tests are shown
ultimate strength of the initially un- in Table 6, together with other perti-
cracked specimens was greater than nent data concerning the test speci-
that of the corresponding initially mens. The ultimate shear transfer
cracked specimens, the residual strength is expressed as a nominal ulti-
strengths of both types of specimens mate shear stress:
were about the same for slips of 0.05
in. or more. _ Ultimate shear
Deformation behavior was found to Vu Area of shear plane
become more brittle as the compressive
strength of the concrete increased; and The ultimate shear is defined as the
also to be more brittle in the case of maximum shear carried by the speci-
the all-lightweight concretes than in men during the test.
the cases of the sanded lightweight Effect of a crack in the shear plane—
concrete and the sand and gravel con- In Fig. 8 a comparison is made of the
crete. ultimate shear transfer strengths of the

PCI JOURNAL/January-February 1976 ҟ 31

 1600
Initially cracked concrete
1400 fc = 4000 psi ; fy = 50 ksi

Sand and gravel concrete, ҟ•

1200 this study ,—^^ •
previous ҟ•ҟ -
♦ •ҟ

1000
Vu
'^0'0
(psi)
800
a^ҟ 8
Sanded lightweight
concrete
600

400 All-lightweight concrete

200

ҟ
0 200 400 600 800 1000 1200 1400

Pfy (psi)
Fig. 9. Effect of aggregate type on shear transfer strength of
initially cracked concrete having f, — 4000 psi.

sanded lightweight concrete specimens sand and gravel concrete specimens, the
of Series A and B. The corresponding difference in strength between the in-
specimens of these two series were itially uncracked specimens and the in-
made as nearly identical as possible, itially cracked specimens decreased
except that the specimens of Series A continually as pf, increased.
were tested in an initially uncracked In previous studies5 of shear transfer
condition, while those of Series B were in sand and gravel concrete, similar
cracked along the shear plane before behavior was observed. However in
being tested. that case (for concrete with f' — 4000
It can be seen that for all values of psi), the strength of the initially cracked
the reinforcement parameter pf, consid- concrete became equal to that of the
ered, the existence of a crack in the initially uncracked concrete when p f
shear plane reduces the shear transfer was about 1350 psi.
strength, for a given value of pf, by an It was postulated that under the
almost constant amount. clamping force provided by this large
Similar behavior also occurred in the amount of reinforcement, the crack
case of the all-lightweight concrete spe- "locked up" and the concrete behaved
cimens. However, in the case of the as if it were initially uncracked. In such

32
 a case diagonal tension cracks formed all-lightweight concrete.
across the shear plane and the failure The differences in shear transfer
had the characteristics of a shear trans- strength do not correlate with the dif-
fer failure in initially untracked con- ferences in concrete splitting tensile
crete. strength. The shear transfer strength of
In the present tests diagonal tension lightweight concrete in design should
cracks also occurred for this heavy de- not therefore be related to the splitting
gree of reinforcement, but the strength tensile strength of the concrete.
of the initially cracked concrete fell a
The difference in the shear stresses
little short of that of the initially un-
which can be carried by lightweight
cracked concrete. It is posssible that the
and normal weight concretes of the
difference in behavior noted in the two
same compressive strength is usually
studies is only apparent and was caused
attributed to differences in the tensile
by experimental scatter in the strength.
strength of the concretes. However this
data.
cannot be the case in this instance,
In the previous study5 it was postu-
since the average splitting tensile
lated that the upper limit to the shear
strengths of the 4000 psi concretes were
transfer strength of initially cracked
concrete of a particular compressive 6.02V7 for the sanded and gravel
strength resulted from this "locking up" concrete, 5.69/ f', for the sanded
of the crack and subsequent behavior lightweight concrete, 5.78 ,/ for the
as if initially uncracked. However, this coated all-lightweight aggregate con-
does not appear to be the case for light- crete and 6. 14 V7 for the crushed
weight concrete in which no evidence lightweight aggregate concrete. (see
of such "locking up" behavior was ob- Table 7).
served.
The difference in shear transfer
No diagonal tension cracks occurred
strength between lightweight concretes
in any of the initially cracked light-
and sand and gravel concretes of the
weight concrete and the shear transfer
same compressive strength is probably
strength of the initially cracked light-
due to differences in roughness of the
weight concrete did not approach that
crack faces in contact. Some of the
of the initially uncracked lightweight
push-off specimens were cut open after
concrete at high values of pfy,. The ceil-
test and it could be seen that the crack
ing value of shear transfer strength for
faces in the sand and gravel concrete
Iightweight concrete must therefore re-
were rougher than those in the light-
sult from some other aspect of shear
weight concrete.
transfer behavior as yet unidentified.
Effect of aggregate type on shear In sand and gravel concrete, the
transfer strength—In Fig. 9 a compari- bond strength between the mortar and
son is made of the shear transfer the aggregate particles is smaller than
strength of initially cracked concrete of the tensile strength of the aggregate
three kinds, all having compressive particles. Cracks therefore generally
strengths of close to 4000 psi. propagate around the aggregate par-
It can be seen that the shear transfer ticles, producing a rough surface.
strength of the sand and gravel con- In lightweight concrete the bond
crete is consistently greater than that strength between the mortar and the
of the lightweight concretes, for the aggregate particles is apparently great-
same amount of reinforcement, and that er than the tensile strength of the ag-
the sanded lightweight concrete has gregate particles. In this case therefore,
a greater shear transfer strength than cracks propagate through the aggregate

PCI JOURNAL/January-February 1976 33

 1600
Initially cracked, sanded lightweight
(coated aggregate) concrete.
1400 f 4000 psi — q ; 6000psi — •
fc= 2500psi — A (fy = 50ksi)
1200 1200psi
/^^----
Vu = Pfyp( 3 Q + 0.5)ҟ u
Pfy
1000 but * 0.25fc nor 1200psi. \, . q
VU (y= 1.4) i q
i/
(psi)
Shear Friction % q
800
vu=PfyN
/
— 0^800psi p
but 0.2fc\^
600 nor 800psi.
(Y = 4) 0
400 q 500psi
4 (0.2ff for ff= 2500psi)
Shear Friction
200 ' using N = 0.85(1.4)
= 1.19

0 200 400 600 800 1000 1200 1400

Pfyҟ (psi)
Fig. 10. Comparison of shear transfer strength of initially
cracked, sanded lightweight concrete, with strength
predicted by current design equations.

and a smoother crack face results. This feet is only minor. However it is con-
difference in cracking behavior was sidered that it could influence shear
readily apparent in both the push-off transfer behavior, since the slips and
specimens that were cut open after test, separations at ultimate measured in the
and in the appearance of the two halves tests are only of the order of the size of
of cylinders subject to the splitting ten- the sand grains.
sile strength test. It can also be seen in Fig. 9 that
The difference in shear transfer be- while the rate of increase of shear trans-
tween the sanded lightweight concretes fer strength with increase in pf y is
and the all-lightweight concretes of about the same for all three concretes
similar compressive strength was prob- for moderate values of p f ,, the shear
ably due to similar causes. In the sand- transfer strength of the lightweight con-
ed lightweight concrete the cracks cretes commences to increase at a less-
er rate at a lower value of p f y than
propagated round the sand particles,
does the sand and gravel concrete.
producing a rougher surface than in the
It apears that for a given strength,
all-lightweight concrete.
the absolute maximum shear transfer
The roughness produced by this of- strength obtainable is less in the case of

34
lightweight concretes than in the case shear-friction equations can continue to
of sand and gravel concrete. This may be used provided that the coefficient M,
be due to the resistance to abrasion of is multiplied by
the lightweight concrete crack faces
300
being less than that of the sand and + 0.5
gravel concrete crack faces. pf„
"Skid" marks could be seen on the That is, Eq. (2) then becomes:
crack faces of the more heavily rein-
forced lightweight concrete push-off
specimens cut open after test. Similar
v u = M!,µ1 pfy +0.51 (3)

damage to the crack faces was not ob- Eq. (3) will be referred to as the PCI
served in the case of sand and gravel equation. Initially, no upper limit was
specimens. specified for v calculated using the
This may be the reason why the PCI equation, but subsequently an up-
"Iocking up effect noted in heavily per limit of the lesser of 0.25f', or
reinforced sand and gravel concrete 1200 psi was proposed for v,u.
push-off specimens did not occur in In Figs. 10 and 11 comparisons are
heavily reinforced lightweight concrete made between measured shear transfer
specimens. Failure of the lightweight strengths in initially cracked light-
concrete crack faces through extensive weight concretes and the strengths pre-
shearing off of local roughness appar- dicted by Eq. (2) and (3).
ently occurs before the crack faces can (Note that in making these compari-
become locked together. sons, the value of the capacity reduc-
tion factor cA is taken as 1.0, since the
Calculation of shear transfer strength material strengths and specimen dimen-
in design—Currently, the calculation of sions are accurately known.)
shear transfer strength in design is us-
It is seen that the PCI equation is
ually based on the provisions of Section
unconservative for both sanded light-
11.15—Shear-friction, of ACI 318-71.1
weight and aII-lightweight concretes
According to these provisions, shear
and should therefore not be used in the
transfer strength is given by:
design of connections between mem-
vu = 0 A vffv p (1) bers of lightweight concrete.
In Fig. 10 it can be seen that the
which may also be expressed as an ulti-
shear-friction equation, using pu= 1.4,
mate shear stress:
becomes unconservative for sanded
Vu lightweight concrete when pf, exceeds
v u = Air = µPfv (2) about 450 psi. It is therefore proposed
that the multiplying factor for sanded
Section 11.15 also specifies that v,, lightweight concrete, 0.85, contained
shall not exceed 0.2f, nor 800 psi. in. Section 11.3.2 of ACI 318 -71 should
The value of the co n fident of fric- also be applied to the coefficients of
tion µ, is to be taken as 1.4 for concrete friction fit, contained in Section 11.15.4.
cast monolithically, 1.0 for concrete For a crack in monolithic sanded
placed against hardened concrete, and lightweight concrete, fie, then becomes
0.7 for concrete placed against as-rolled 1.19. A line corresponding to use of this
structural steel. value of Ft in shear-friction calculations
In Section 6.1.9 of the PCI Design has been drawn in Fig. 10 and it is
Handbook,2 it is proposed that for val- seen to be reasonably conservative for
ues of pfy greater than 600 psi, the sanded lightweight concrete.

PCI JOURNAL/January-February 1976 ҟ 35

 1600
Initially cracked, all-lightweight concrete.
fc = 4000psi ; fy = 50ksi
1400
Coated aggregate . – 0
Crushed. aggregate – 0
1200psi
1200 Vu = pfyp i300
f + 0.5)^ .-----

///
but ' 0.25ff nor 1200psiҟ
1000
VU 0
(psi) Shear Friction ; '° 0
800
V = p fy P "./ 800 psi
but ) 0.2f ҟ JP
600 nor 800 psi 04
(r = 1.4)ҟ O
O
400 n Shear Friction
using p = 0.75(1.4)
'ii= 1.05
200

ҟ
0 200 400 600 800 1000 1200 1400
pfy (psi)
Fig. 11. Comparison of shear transfer strength of initially
cracked, all-lightweight concrete, with strength predicted
by current design equations.

It can be seen in Fig. 11 that the seen to be reasonably conservative for

shear-friction equation, using li= 1.4 all-lightweight concrete.
for a crack in monolithic concrete (as It can be seen from Figs. 10 and 11
specified in Section 11.15.4 of ACI 318- that the limiting value for vu contained
71), can be considerably unconservative in Section 11.15.3 of ACI 318-71 (0.2
for all-lightweight concrete. f, but not more than 800 psi) is also
It is therefore proposed that the mul- appropriate for both sanded lightweight
tiplying factor for all-lightweight con- and all-lightweight concretes, provided
crete, 0.75, contained in Section 11.3.2 that the reduced values of fit, proposed
of ACI 318-71 should also be applied to above are used.
the coefficients in friction I,c contained Subsequent to the publication of
in Section 11.15.4. For a crack in mono- ACT 318-71 and the PCI Design Hand-
lithic all-lightweight concrete, /i then book, Mattock proposed6 the following
becomes 1.05. equation for the shear transfer strength
of sand and gravel concrete:
A line corresponding to use of this
value of p, in shear-friction calculations vu = 0.8 p f y + 400 psi (4)
has been drawn on Fig. 11 and it is but not more than 0.3 f .

36
 1600
Initially cracked, sanded lightweight
(coated aggregate) concrete.
1400 ff= 4000psi - q ; ff= 6000psi -
ff= 2500psi - A (fy= 50ksi)
1200
L1
vu = 0.8pfy + 400psi
1000
VU but * 0.3fcҟ0"^1000psi
(psi) q
800
800psi , (0.2fc
for fc = 4000psi)
^Ɂa
600
/^ a Oҟ a
q ^^ 500psi, (0.2fc for ff- 2500psi)
400

vu = 0.8pfy + 250psi
200 but 3- 0.2f, nor I000psi
min. pfy a 200psi

200 400 600 800 1000 1200 1400

0ҟ
Pfy (psi)
Fig. 12. Comparison of shear transfer strength of initially
cracked, sanded lightweight concrete, with strength
predicted by Eqs. (4) and (5).

Also, pf y was to be not less than strength be used for lightweight con-
200 psi. crete:
In Figs. 12 and 13, the shear transfer (a) For sanded lightweight concrete:
strength predicted by Eq. (4) is com- v,,=0.8pfy + 250psi (5)
pared with the measured shear transfer
strength of sanded lightweight and all- but not more than 0.2f', nor 1000 psi.
lightweight concrete, respectively. (b) For all-lightweight concrete:
It can be seen that Eq. (4) is non- v,, = 0.8 pf, + 200 psi (6)
conservative in both cases and there-
fore it should not be used in the design but not more than 0.2 f', nor 800 psi.
of connections between members made Lines representing Eqs. (5) and (6)
of lightweight concrete. However, it have been drawn on Figs. 12 and 13,
can also be seen that for moderate respectively, and it can be seen that
values of pf ,,, v, increases with pf, at these equations provide a reasonably
the same rate as predicted by Eq. (4). conservative estimate of the shear trans-
It is therefore proposed that the fol- fer strength of the two types of light-
lowing equations for shear transfer weight concrete.

PCI JOURNAL/January-February 1976ҟ 37

 1600
Initially cracked, all-lightweight concrete.
f^ = 4000psi ; fy = 50ksi
1400
Coated aggregate - 0
Crushed aggregate - A
1200 psi
1200
(0.3ft for
vu= 0.8pfy + 400 psiҟf 4000psi)
1000. but t 0.3ff
vu 0
(psi) pҟ
O
800
-800psi, (0.2ff
for fc=4000psi)
600

vU = 0.8pfy+ 200 psi

400
but 3- 0.2fc nor 800psi

200
min. efy= 200 psi

0 200 400 600 800 1000 1200 1400

pfyҟ
(psi)
Fig. 13. Comparison of shear transfer strength of initially
cracked, all-lightweight concrete with strengths predicted
by Eqs. (4) and (6).

3. The shear-friction provisions of

Conclusions for Design Section 11.15 of ACI 318-71 may be
used to calculate shear transfer strength
On the basis of the study reported here, in the design of connections in light-
the following conclusions are drawn weight concrete, providing the value of
concerning shear transfer in lightweight the coefficients of friction µ, contained
aggregate concrete: in Section 11.15.4, are multiplied by
1. The shear transfer strength of the following factors:
lightweight concrete is less than that of
sand and gravel concrete of the same (a) For all-lightweight concrete hav-
compressive strength. ing a unit weight not less than
2. The shear transfer strength of 92 lb per cu ft, multiply [4 by
lightweight concrete is not significantly 0.75.
affected by the type of lightweight ag- (b) For sanded lightweight concrete
gregate, i.e., by whether the aggregate having a unit weight not less than
is a "coated aggregate" or "crushed ag- 105 lb per cu ft, multiply It by
gregate." 0.85,

38
 4. The provisions of Section 6.1.9 of
the PCI Design Handbook are not ap- References
plicable to lightweight concrete.
1. ACI Committee 318, "Building
5. The shear transfer strength of
Code Requirements for Reinforced
sanded lightweight concrete having a Concrete (ACI 318-71)," American
unit weight not less than 105 lb per cu
Concrete Institute, Detroit, 1971.
ft can be calculated using the following
2. PCI Design Handbook, Prestressed
equation:
Concrete Institute, 1971.
v.=0.8pfr+250psi 3. Mast, R. F., "Auxiliary Reinforce-
ment in Concrete Connections,"
but not more than 0.2f', nor 1000 psi Journal of the Structural Division,
and pf, to be not less than 200 psi. American Society of Civil Engineers,
V. 94, ST 6, June 1968, pp. 1485-
6. The shear transfer strength of
1504.
all-lightweight concrete having a unit
4. Mattock, A. H., "Effect of Aggregate
weight not less than 92 lb per cu ft
Type on Single Direction Shear
can be calculated using the following
Transfer Strength in Monolithic
equation:
Concrete," Report SM74-2, Depart-
v,,, 0.8pfy+200psi ment of Civil Engineering, Univer-
sity of Washington, Seattle, Wash-
but not more than 0.2f', nor 800 psi
ington, August 1974.
and pf,5 to be not Iess than 200 psi.
5. Mattock, A. H., and Hawkins, N. M.,
"Research on Shear Transfer in
Reinforced Concrete," PCI JOUR-
Notation NAL, V. 17, No. 2, March-April,
1972, pp. 55-75.
A0,. = area of shear plane 6. Mattock, A. H., "Shear Transfer in
A„1 = total area of shear transfer re- Concrete Having Reinforcement at
inforcement an Angle to the Shear Plane," Pub-
= compressive strength of con- lication SP-42, Shear in Reinforced
crete measured on 6 x 12- Concrete, American Concrete Insti-
in. cylinders. tute, 1974, pp. 17-42.
fat = splitting tensile strength of
concrete measured on 6 X
12-in, cylinders
Acknowledgment
vu = nominal ultimate shear stress This study was carried out in the Struc-
[v = Vu/( 4 A00)] tural Research Laboratory of the Uni-
Vu = ultimate shear force versity of Washington. It was supported
ji = coefficient of friction used in by the National Science Foundation,
the shear-friction hypothesis through Grant No. GK-33842X. Light-
p = A,f/ A,, weight aggregate was supplied by
0 = capacity reduction factor members of the Expanded Shale, Clay
ard Slate Institute.

Discussion of this paper is invited.

Please forward your discussion to
PCI Headquarters by June 1, 1976.

PCI JOURNAL/January-February 1976ҟ 39