372

LOAD TRANSFER IN COMPOSITE MASONRY

STEPHEN A. MATTY Structural Engineer, Greiner, Inc.

Timonj.um, Maryland, USA JAMES COLVILLE

Professor and Chairman, Dept. of Civil Engineering University of Maryland, College Park, Maryland 20742, USA

AMDE M. WOLDE-TINSAE Associate Professor, Dept. of Civil Engineering

University of Maryland, College Park, Maryland 20742, USA

ABSTRACT

The results of an experimental investigation into the load transfer mechanism in composite masonry construction are presented. Results include bed joint strain readings in 46 composite masonry prisms from which an estimate of the distribution of shear force in the collar joint may be obtained.

INTRODUCTION

Although composite masonry construction, consisting of clay brick and

concrete masonry wythes bonded together by a collar joint of mortar or

grout, is widely used in the construction industry, the structural

behavior of composite walls, including the shear capacity of the collar

joint, load transfer mechanism between the wythes, and the effects of

restrained differential movement between the wythes is still not fully

understood.

It is believed that the key to understanding composite masonry lies

in defining the behavior of the interconnection between the wythes. The

primary objective of this pape r is to present the results of a series of

tests to investigate the load transfer mechanism in clay brick and

concrete masonry composite walls. A more comprehensive report on this

subject is presented in Reference [1]. In addition, Reference [2]

presents a synopsis of results of mortared collar joint shear capacities.

373

HATERIALS

Two types Df brick units were used: a Jefferson unit manufactured by

Glen-Gery Brick Corporation, Herndon, Virginia, and a Continental 375

unit manufactured by Potomac VaLley Brick and Supply Corporation, Lorton,

Virginia.

Tests were conducted on these units in accordance with the methods

and procedures described in Standard Methods of Sampling and Testing

Brick and Structural Clay Tiles, ASTM C67-80a. The properties Df the

units were as follows:

Jefferson brick (Note: coefficient of variation in % in parentheses).

Size

Weight

Absorption

Saturation coefficient

rnitial rate of absorption

2 1/4" x 3 7/16" x 7 7/8"

1672 gr (0.1 %)

3.2% (22.0%)

0.62 (7.7%)

15.9 gr/min/30 in2 (5.5%)

Compressive strength 12,500 psi (2.3%)

Continental brick (Note: coefficient Df variation in % in

parentheses)

Size

Weight

Absorption

Saturation Coefficient

rnitial rate of absorption

Compressive strength

2 1/4" x 3 9/16" x 8"

1820 gr (6%)

6.8% (36.2%)

0.80 (8.1%)

35.7 gr/min/30 in2 (21.8%)

12,700 psi (7.2%)

Because Df the relatively high initial rate Df absorption Df the

continental brick units, these bricks were wetted and allowed to surface

dry for approximately 45 minutes before construction Df the composite

test specimen.

The concrete masonry units used were hollow load-bearing units with a

nominal dimension of 6" x 8" x 8", manufactured by Ernest Maier Block

Company, Bladensburg, Maryland. Average properties of the units were as

follows (Note: coefficient of variation in % in parenthesis).

Size 5 5/8" x 7 7/8"

Gross area 44.3 in2

Weight 11.82 lbs (1.35%)

Absorption 11.5% (6.2%)

Compressive Strength

(Gross area)

374

1395 psi (8.7%)

AlI mortars were mixed in accordance with the Standard Specifications

for Mortar for Unit Masonry, ASTM C270-80a. A portland cement-lime

mortar was used for both Type N and Type S mortars.

The average compressive strength of the mortars was determined

according t o ASTM C780-80 and C109-80. For the Jefferson bricks, the

mortar strengths were 520 psi (CV = 20.6%) and 970 psi (CV = 8.9%) for

Type N and Type S mortar respectively . Comparable data for the specimen

made using the Continental 375 units was 520 psi (CV = 36.3%) for Type N

mortar, and 1110 psi (CV = 27 . 7%) for Type S mortar.

For the taller composite prisma, the Type N mortar compressive

strength was 772 psi (14.5%) for the Jefferson units and 878 psi (43.%)

for the Continental units.

Horizontal ladder type joint reinforcement was provided by DUR-O­

WALL, Inc. Because of the r ela tively short specimen length, each collar

joint in the reinforced specimen was crossed by only a single #9 wire

(Note that the joint reinforcement was placed in the joint between

concre t e masonry units) .

The area of the joint reinforcement was 0.0171 in2 • Tests of three

tension specimens indicated that the yield strength of the s t eel was 88.3

ksi with an ultimate strength of 94.7 ksi.

Clay masonry prisms were tested for each combination of unit type and

mortar type. Three specimens were tested for each of the four parametric

combinations. Compressive strengths corrected for h/t ratio in

acco rdance with requirements of the Brick Institute of America are

presented below:

Jefferson units - Type N mort ar (JN) 2155 psi (CV=9.5%)

Jefferson units - Type S mortar (JS) 3050 psi (CV=1.6%)

Continental units - Type N mortar (CN) 1530 psi (CV=11.6%)

Continental units - Type S mortar (CS) 3310 psi (CV=3.9%)

FABRICATION

An experienced mason was used to fabricate 46 composite prisms. Forty

specimens were 16" high by 8" wide symmetrical prisms with two 3/8"

collar joints as shown in Fig la . The other six specimens were 32" high

x indicate s ~jir gage locations

I

(20 gugcs totul - eRch spccimen)

1

2

J

4

5

:1) Short spe~imC'1l 11) Tall Spce c:ímen

igllrc L - Composí tl' Tl'st Spl'cimen

w --...l VI

376

by 8" wide specimens with 3/8" collar joints as shown in Fig. lb. The

nomenclature used to identify the specimens is as follows: J;Jefferson

brick; C;Continental 375 brick; N; Type N mortar; S ; Type S mortar; R;

reinforcement; NR;no reinforcement, and T ; tall wall specimen (32"

high) •

For the 40 short composite prisms there were 8 combinations of 5

prisms fabricated. The combinations were as follows: J-N-NR; J-N-R; J-S­

NR; J-S-R; C-N-NR; C-N-R; C-S- NR ; and C-S-R. For the six tal ler

composite prisms two combinations of 3 specimens each were constructed :

JT- N-NR and CT-N-NR. After fabrication, the specimens were covered with

polyethylene and allowed to cure for at least 28 days.

MEASURING DEVICES

Load cells were used to measure the load in each brick wythe and

specially fabricated and calibrated clip gages were used to measure

deEormations across bed joints of the brick wythe and across the collar

joints. A MEGADAC SYSTEM 100 data acquisition system was used to record

alI data.

TESTING DF CDHPDSITE PRISMS

For the short composite prisms, two clip gages were placed across each

brick bed joint. For the taller camposite prisms, two clip gages were

placed across the first, third, sixth, ninth, and eleventh brick bed

joints. Therefore, for both types oE specime ns, a total of 20 gages were

used. (Figures la and lb indicate the location of the clip gages .)

Readings were taken continuously throughout each experimento Since

the specimen is symmetrical the 2 readings obtained on the left hrick

wythe were averaged with the corresponding values obtained on the right

brick wythe.

TEST RESULTS

Typical test results for each specimen consist of bed joint strain

readings versus applied compressive load. A representative plot of this

data for alI C-S-NR prisms is shown in Figure 2. In this figure, it may

be noted that the r e sponse is relatively linear f ollowing an initial

1,0;lU

pl' r Wythc

(klps)

377

o .000 5

Strilln Cln/in)

f' igllfl' 2 - Ikcl Joint Stra Lns - C- S- NR Spl'C Lrnl'n

. 00 10

378

"settling" load of 1000-2000 pounds. The five lines represent the strain

measurements on each of the 5 bed joints (in the short specimens) . Data

obtained in the lowest bed joint was compared to data recorded in simple

prism tests for the same magnitude of loading . Reasonably good agreement

(10% differences in strain) was obtained in this comparison and as a

result it is believed that the individual strain readings ou each bed

joint are representative of actual strains occurring in the composite

specimens.

Also, it should be noted that there was no discernable difEerence in

the load transfer results for either of the two clay masonry units used.

Therefore, results for both types of unit have been combined in this

paper.

A summary of the test results for all specimen groups is presented in

Tables 1, 2 and 3. These tables show the percentage of applied load

transferred through each instrumented bed joint. From this data, the

load being transferred through the collar joints of the specimens can be

estimated as the differences in load transferred through each bed joint.

This latter information is presented in Tables 1, 2 and 3 as load in the

collar joint.

Top of Specimen

Joint 1 Joint 3 Joint 6 Joint 9 Joint 11

Bottom of specimen

TABLE Load Distribution - Tall Prisms

Load/bed joint

.1SP

.56P

.81P

.82P 1.OOP

Collar joint load

.1SP

.41P (2 units)

.25P (3 units)

.01P (3 units)

.18P (2 units)

379

TABLE 2 Load Distribution - Short Prisms, Type N mor t a r

No Reioforeement \Hth Reinfo r ce me nt Top oE Load/bed Colla r Load / bed Colla r Speeimen Joint Jo in t Load Joint Joint load

Joint 1 .45P . 45P .14P .14P J oin t 2 . S3P . 08 P .43P .29p Joint 3 .69P . 16P .43P O Joint 4 . S3P .14P .68P .25P Join t 5 1.00P .1 7P I. OOP . 32P

Bot t om oE Speeimen

TABLE 3 Load Distribution - Short Prisms, Type S mortar

No Reinforeement With Reinforeement Top oE Load/ bed Collar Lo ad/bed Collar Speeime n joiot joint load join t joint load

Joint I .1 6P .1 6P .OlP .0lP Joint 2 .1?P . 01P .30P .23P Joint 3 . 32P .15P . 46P .1 6P Joint 4 .58P .26P . 77P .31P Joint 5 1.00P .42P 1. 00P .23P

Bottom of Specime n

ANALYSIS OF RESULTS

In revie wing the data presented in Table 1, it may be noted that

approximately 56% of the load is trans Ee rred Erom the loaded bloek wythe

through the eollar joint to the briek wythes in the top 8" of the

speeimen, and app roxima t ely 80% oE the load is transfe rred io the upper

halE of the specimen. Thus the eollar joint shear s tress is not uniform

along the height Df the speeimen a nd in fae t the eollar joint shear

stress is maximum in the vicinity Df the ap plied load. These

experimental r es ults are in general agreement with nume r ieal s tudies by

Anand [3,4].

Results for the short specimens are more varied. For the t ype N

mortar prisms with no joint reiofo reement, a eonsiderable portioo (45%)

of the applied load is transferred at the top of the speeimen , with 69%

of the load transEerred in the top half of the spec imeo. There is a lso a

fairly uniform distribution oE load through joints 3, 4 and 5. This

380

relatively rapid load transfer is similar to that obtained in the tal 1

prisms. Thus, for type N mortar without joint reinEoreement both the

tal1 and short speeimens are somewhat similar with respeet to load

transfer near the top of the speeimen. On the other hand, the

unreinforced type S mortar speeimens show a markedly different

distribution of load with a signlfleant fraetion of 10ad transEerred near

the bottom of the speeimens.

ANALYTICAL MODEL

Due to the eompexity of the experimental results a simple analytieal

model of the short composite prisms was deveIoped to qualitatively

investigate the effeet of the stiffness of the bed joints and eollar

joints on the load transfer rneehanism. This model, whieh eonsiders eaeh

masonry unit as a rigid mass and represents the mortar joints as linearly

elastie springs, is shown in Figure 3. Three springs were used to

represent the axial bed joint stiffnesses in the block and briek wythes,

respeetively, and the eollar joint shear stiffness. Results of two

combinations of the relative stiffnesses of these springs are presented

below in Table 4.

TABL.E 4 Load Distribution - Short Prisms

Top of Speeimen

Joint 1 Joint 2 Joint 3 Joint 4 Joint 5

Bottom of Speeimen

Load/bed Joint

.20P

.45P

.83P

.80P 1.00P

Model 1*

Collar Joint Ioad

.20P

.25P

.38P -.03P

.20P

*Model - KI ** Model 2 - Kl

70, K2 20, K3 20 70, K2 5, K3 20

Model 2**

Load/bed Joint

.20P

.42P

.66P

.80P 1.0OP

Collar Joint

.20P

.22P

.24P

.14P

.20P

These results indieate that the distribution of load transferred through

the eollar joint is not uniform. However, as the shear stiffness of the

eollar joint is redueed in relation to the axial bed joint stiffness of

the rnasonry wythes, there is a more uniform load transfer through the

eollar joint.

38 1

k 2

Bric k

kl k2

kJ

k l

k2

!,r i c k flJock

k2

k2

l' t g Llr é' J - Sim pl LEt e>d "'lo d e' l

382

CONCLUSIONS

The results of the load transfer data indicate that the type of mortar

and presence of joint reinforcement influence the load transfer across

the collar joint. For the taller composite prisms, the shearing stress

along the collar joint is nonuniform and the are a of high shear stress

occurs in the upper region of the specimen. AIso, approximately 80% of

the load is transferred through the collar joint to the brick wythe in

the upper half of the eomposite prism. The smaller composite prisms with

similar characteristics to the taller composite prisms, namely, Type N

mortar and no joint reinforcement, showed a similar load transfer

mechanism.

The smaller composite prisms with type S mortared collar joints, and

Type N mortared collar joints with joint reinforcement did not show the

same load transfer distribution as the prisms with Type N mortared collar

joints and no joint reinforcement.

Thus, for the Type S and Type N reinforced specimens, the area of

high shear stress occurs in the lower half of the collar joint although

the load distribution along the eollar joint seems to be more uniform.

Since the simple model analysis indicates a more uniform load transfer as

the stiffness of the collar joint is r educed in comparison to the bed

joint stiffness, the test data would only be consistent with this result

if the Type S mortar produces an increase in collar joint stiffness that

is relatively smaller than the corresponding increase in bed joint

stiffness as compared to Type N mortar. Also, it would appear that the

joint reinforcement has a more pronounced effect on the bed joint

stiffness than the shear joint stiffness.

The results presented clearly indicate that the interaction of

composite masonry through a mortared collar joint is complicated and that

a more accurate and comprehensive analytical model is needed to fully

comprehend the load transfer mechanism data obtained from the

experimental t ests. The major contribution of this paper is the

presentation of previously unavailable data on this interaetion which may

serve as a be nchmark in the evaluation of future analytical methods of

analysis of composite masonry construction.

383

ACKNOWLEDGEMENTS

The work reported in this paper was funded by the International Masonry

Institute. Any opinions expressed herein are solely the responsibility

of the authors and do not necessarily reflect the views of IMI.

REFERENCES

1. Matty, S.A., Colville, J., and Wolde-Tinsae, A.M., "An Experimental Study on the Shear Capacity of Collar Joints and the Load Tra ns fer Mechanism in Composite Masonry," Volumes I and 11, Final Report, submitted to the International Masonry Institute, July 1987.

2. Colville, J., Matty, S.A., and Wolde-Tinsae, A.M., "Shear Capacity of Mortared Collar Joints," Proceedings, Fourth North American Masonry Conference, Los Angeles, California, August 1987.

3. Anand, S.C., "Shear Stresses in Composite Masonry Walls," New Analysis Techniques for Structural Masonry, American So ciery-of Civil Engineers, New York, Sept. 1985, pp. 106-117.

4. Anand, S.C., and Yalamanchili, K., "Evaluation of Loads at Cracking in Collar Joints of Composite Masonry Walls," Department of Civil Engineering, Clemson University, Clemson, S.C.