AN INVESTIGATION OF FACTORS INFLUENCING THE BEHAVIOUR OF MASONRY INFILL IN STEEL FRAMES SUBJECTED TO IN-PLANE SHEAR

Dr J L Dawe Associate Professor T C Yong Graduate Student Department of Civil Engineering, University of New Brunswick, Fredericton, New Brunswick, Canada, E3B 5A3

ABSTRACT An investigation of various factors which may affect shear strength and behaviour of masonry infilled steel frames is presented. Findings of this study are believed to be equally applicable to panels constructed of burnt clay masonry units as well as to panels constructed of concrete masonry units. Factors such as openings in panels, bond and friction between panel and frame, provision of airspace between roof beam and panel, and column-to-panel ties are investigated using large-scale test specimens. A computer-aided analytical technique combining a frame stiffness analysis with a finite element panel analysis is used to determine stress distri­butions in the panels in the pre-cracking range.

1. INTRODUCTION The bene ficia l aspects of masonry in providing a load-sharing structural system in concrete and steel frame structures have become increasingly apparent since about the early fifties. Generally, much research has been done in the area of burnt clay masonry and concrete masonry units with regard to their structural strength, mechanical properties, and overall behaviour in engineered masonry structures. Many researchers have contributed widely to theoretical and experimental understanding of the general problem of shear behaviour of infilled frames. Some of this work relating directly to masonry infilled frames is briefly summarized elsewhere in these proceedings (1). In the case of masonry infilled steel frames, most engineers are aware that masonry infilled panels in steel frames provide a large reserve of resistance to in-plane shear loads and that, in return, the steel frameprovides considerable ductility to the system. The purpose of the investigation presented herein is to add, both analytically and experimentally, to the knowledge and understanding of the strength and behaviour of masonry infilled steel frames subjected to in-plane loading. In particular, effects of joint reinforcement, mortar strength, panel openings, and interface conditions between panel and frame, are investigated. The analytical procedure consists of a frame stiffness analysis combined with a finite element analysis of the masonry infill with simple 2-node elements at the interface to account for friction, ties, and compressive contact stresses. The results of twelve tests on large-scale masonry infilled steel frame specimens are presented, discussed, and compared with theoretical obser­vations where valid comparisons can be made.

2. ANALYTICAL TECHNIQUE

2.1 Introduction

Large-scale specimens considered in this investigation consist of simple rigid portal steel frames with infilled masonry panels. The analytical tech­nique, however, is generally applicable to multi-storey, multi-bay planar infilled frames and would be limited only by computer time and cost consider­ations. The analysis includes the effects of horizontal joint reinforcement, pinned or rigid frameconnections, panel-to-frame ties, friction between panel and frame, and provision of an airspace between roof beam and

803

panel. An iterative technique is used to determine the interaction between panel and frame with regard to friction and separation. Up to the occurrence of first major cracking in the large-scale specimens tested, experimental observations appear to confirm results predicted theoretically. It is evident that the analytical technique may be reliably used to evaluate stiffness as well as stress distributions in a panel and frame prior to initial major cracking.

2.2 Frame Analysis

The frame is analysed by a direct stiffness technique using frame elements of two nodes at which axial, transverse, and rotational displa cements are defined. As shown in figure l(a), frame element nodes are defined to correspond

10 11 12 13 14 15 16 17 H § .9 14 23 28 37 42 5 56 6:5

8 8 64 18

7 7 1.3 55 6~ ~F RAME

'i 6 62 20 f-- PA ~ !'ti 12 54 21

'>' ~ ~ 60

DETAIL 'A NEL

4 22 FRAME

3 11 53 59

3 58 23

2 2 I 10 15 24 29 38 43 52 57 24 DETAI L 'A'

I <: ~ 25 )(

;:...ft FIXED BASE 32 31 3 O 2 9 28 27 26

(o) FINITE ELEMENT DISCRETIZATION (b) INTERFACING ELEMENT

Figure 1. Finite Element Model of Frame and Panel.

to nodes on the panel periphery. Following the usual procedure, stiffness matrices for the frame elements are assembled directly into the frame stiffness matri x. Hinges may al so be inserted at the nodes and general in-plane member loads can be accommodated.

2.3 Panel Analysis

Figure l(a) shows a finite element mesh for a panel which is discretized into 8-node rectangular elements with two degrees of freedom at each node. The 16 x 16 stiffness matrix for each element is determined and assembled directly

804

into the panel stiffness matrix. Because of the size of stiffness matrix of each element (256 numbers per matrix) a mesh size analysis was carried out to determine the optimum number of elements to use for the size of large-scale specimens tested. It was determined that 16 rectangular elements gave accurate results as well as economy of computer costs for the iterative procedure requi red.

2.3.1 Reinforcement. The effects of horizontal reinforcement in the panel are also included in the analysis. It is assumed that reinforcement is inte­grally bonded to masonry units at the nodes of rectangular panel elements throughwhich the reinforcement passes. Two-node line elements are used to model the reinforcement between corresponding rectangular element nodes. The 2 x 2 stiffness matrix of a 2-node reinforcement element tangent to, or inter­secting a rectangular element at its nodes,is added directly to the rectangular element stiffness matrix. The resulting stiffness matrix is then assembled into the structure stiffness matrix.

2.4 Panel - Frame Interaction

At the time of assembly, the panel and frame stiffness matrices are uncoupled. The interaction between a frame node, i, and a panel node, j, is modelled by horizontal and vertical 2-node line elements each of an arbitrarily negligible length and of stiffness kn and ks, respectively,as shown in figure l(b). The horizontal element is used to model bond forces as well as the effects of ties between the masonry and steel at corresponding nodes, while the vertical element is used to model friction between the panel and frame or shear in ties.

2.5 Analytical Procedure

Although more general loads may be considered, loading for the analysis of specimens of the type described herein consists of a single horizontal force applied at node number 9 of the frame as shown in figure l(a). Interaction of the frame and panel are realized in the computer model by use of the inter­face elements described previously. During the first step of the analytical procedure, all interfacing friction elements are assumed to have zero stiffnesses while high nominal stiffnesses are assigned to all interfacing normal elements. The horizontal and vertical displacements of nodes i and j in figure l(b) are ui and Uj and vi and Vj ,respectively. The friction spring stiffness, ks , parallel to the interface,varies with the force in a normal element whose stiff­ness is kn. After the first iteration, all nodal displacements are known and a normal force between frame and panel nodes, i and j, is given by,

( 1 )

where a negative result indicates compression. If Fn is greater than the ten­sion bond force between the mortar and frame, kn is set equal to zero for subsequent iterations unless the panel and frame again come into bearing contact in which case kn is replaced by its original nominal value. If Fn is less than the tension bond force (or less than zero if the bond has been previously broken), kn is given its previous nominal value for the next iteration. If Fn is less than zero, the friction element is given a stiffness value of,

I(vj - vi)1 > O (2)

805

for the next iteration where ~ is the coefficient of friction between the masonry and the steel frame. In any iteration, if tension develops at a node or if I (Vj - Vi) I = O, ks is given a value of zero for the next iteration. If a tie exists at a node, ks and kn are given high nominal values and shear and tension forces at any iteration are checked against the breaking strength of the tie. If the tie breaks at any iteration the corresponding nodes are subsequently treated as untied. The iterative technique is continued until there are negligible differences between two successive iterations.

3. EXPERIMENTAL INVESTIGATION

3. 1 I ntroduct i on

Experimental results for 12 large-scale masonry infilled steel frame specimens are presented and discussed and compared with theoretical findings where applic­able. Overall dimensions for the masonry panels of all large-scale specimens were 3 600 mm long by 2 800 mm high. The panels consisted of 200 x 200 x 400 concrete masonry blocks placed in running bond within a surrounding rigid steel frame consisting of W250 x 58 columns and a W200 x 46 roof beam. Further over­all specimen characteristics as well as a description of testing apparatus, instrumentation,and testing procedure are available elsewhere (1,2).

3.2 Specimen Descriptions

Twelve large-scale specimens were tested during this investigation. Specimen characteristics were varied to determine the effects of mortar strength, panel­to-column ties, friction between panel and frame, panel openings, and airspace between roof beam and wall. The details required to study these effects are outlined in Table 1 for each specimen. All panels had horizontal truss-type

TABLE 1

Specimen Number

lA 2A,3A

4A 5A 6A lB

2B 3B & 4B

5B 6B

Specimen Details

Description

Mortar packed between column flanges and panel Mortar omitted between column flanges and panel 20 mm airspace between roof beam and panel Same as 4A but with flat bar panel-to-column ties Same as 2A & 3A but with flat bar panel-to-column ties Same as 2A & 3A but with polyethylene membrane between panel and frame to reduce friction and bond Same as lB but with a purposely very weak mortar Opening at center of panel (see Figure 2) Opening offset towards load (see Figure 2) Opening offset away from load (see Figure 2)

joint reinforcement at alternate courses. Details relating to joint rein­forcement and panel-to-column ties are available elsewhere in these proceedings (1). Panel openio_9 details are shown in figure 2.

806

DETAIL A WALLS 3B 8 4B

.?4 BONO BEAM

~BLOCKS ~OUT

I DODi" II I~ 2Xl5 M

WALLS 5B 868

ONLY 3 BONO BEAM BLOCKS

OEFORMEO BARS.1.2M LONG

~ X-SECTION AS ABOVE

OOORWAY

Figure 2. Panel Opening Details.

3.3 Experimental Results

j J I I I

I I

DETAIL B L.OAO

l I I I I I

I ~ :::E .~

1 I I I N N N

I I I I

11 I

LOCATION OF OPENING

SPEC . NO OIMENSION 'a' (M)

38846 1.4 5B 2 .0 6B 0 .8

Each specimen was mounted in a test frame and a horizontal load applied at roof level was gradually incremented in steps of 22.2 kN up to the ultimate capacity of the specimen. Recorded data included progressive cracking patterns, loads, deflections at various points along the frame, and slippage and separation between panel and frame. Loads at initial major cracking and ultimate as well as initial wall stiffness and mortar compressive strength are summarized in Table 2 for all specimens tested.

TA8LE 2 Summary of Test Results

Specimen Initial Major Ult imate In iti a 1 kN Mortar Compressive Number Crack load (kN) Load (kN) Stiffness (mm) Strength (MPa)

lA 245 449 71. 9 20.2 2A 413 538 47.1 19.5 3A 307 556 43.8 17.7 4A 169 209 26.9 18.9 5A 200 231 34.4 24.9 6A 311 423 72.5 38.6 18 270 420 40.9 24.4 28 155 310 45.5 12.4 38 55 285 34.2 17.5 48 66 335 34.2 58 90 245 38.5 16.5 68 67 365 50.0

3.3.1 Load-Deflection Studies. Load-deflection curves for all 12 specimens are compared in Figure 3. In each case the horizontal load at roof level is plotted on the vertical axis and the horizontal defl~ction of the spe~imen at the point of load is plotted on the horizontal.ax1~. Load-deflect1on curves for specimens lA, 2A, and 6A are compared 1n f1gure 3(a). The panel of Specimen 6A was mechanically tied to the columns whereas the panel of

807

specimen lA was tightly bonded by mortar between column flanges. The panel of specimen 2A on the other hand was free to rotate and slip relative to the frame. Consequently, as loading increased, it was able to assume optimum interaction with the deflected frame with regard to equivalent strut behaviour (3,4,5). Of the three, this panel had the highest initial major crack load and the highest ultimate load (table 2). (A major initial crack load was taken as the load at which a large visible crack,usually accompanied by a noticeable drop-off in load, occurred).

Specimen 4A and 5A each had an airspace along the underside of the roof beam. The panel of specimen 4A was not mechanically attached to the frame while specimen 5A had panel-to-column ties. Although the overall behaviour was similar for these specimens, as shown in figure 3(b), it appears that ties may have contributed to an increase in initial cracking load of about 18% but onlya slight increase in ultimate load is evident. Curves shown in figure3(c) are for Specimen 3A which had joint reinforcement and no ties, Specimen 18 which was similar but with a polyethylene membrane placed between the panel and frame, and Specimen 28 which was similar to 18 except that a purposely weak mortar was used. Comparing these curves, it appears that reducing friction and bond between panel and frame does not significantly reduce the cracking load although there is an indication that ultimate load may be significantly decreased (see table 2). Also it appears that a very weak mor­tar considerably reduces the stiffness beyond initial cracking as well as the ultimate strength.

Figure 3(d) shows the load-deflection curves for Specimens 38 to 68 which had openings as described in table 1 and figure 2. Specimens 38 and 48 each had a centrally-located door opening and did not experience a drop in load although first cracks appeared at 55 kN and 66 kN respectively. Specimens 58 and 68 had larger initial stiffnesses than panels with central openings. Specimen 58, which had an opening offset from center towards the loaded side, had the lowest ultimate load,while Specimen 68, which had the same size opening offset from center towards its unloaded side, had the highest ultimate load of the three. These observations are somewhat at variance with the findings of Mallick and Garg (6) who found that an opening remo te from the loaded side of such a specimen tends to decrease its capacity. Figure 3(e) shows the effects of cycling Specimen 6 four times to ultimate. On the first loading, the ultimate load was 365 kN and after three additional cycles approximately 85% of this load was still attainable.

3.3.2 Crack Studies. Progressive crack patterns for Specimens lA to 6A are shown in figure 4 while those for Specimens 18 to 68 are shown in figure 5. Comparing crack patterns of Specimens 2A and 3A, which had no mechanical panel­to-frame ties, with those of Specimen lA, in which the panel was mortared snugly between column flanges, and Specimen 6A which had flat bar ties, it appears that panel-to-frame ties induce random off-diagonal cracking. More tightly banded, compressional diagonal crack patterns occur for panels that are not tied to the columns. Such panels would be expected to more closely approximate an equivalent strut type of behaviour and to have overall better behaviour in load-reversal situations. The crack pattern shown in figure 5(a) for Specimen 18 which was similar to Specimens 2A and 3A except for a polyeth­lene membrane between panel and frameto reduce bond and friction stresses, also clearly displays equivalent strut type of behaviour. Specimen 28 which had a weak mortar has more extensive off-diagonal random cracking than Specimen 18 which was otherwise similar.

808

600 600 2A

500 500

6A .---. z

.---. 400 IA .!<:: 400 Z '--'

.!<:: ~ '--' 300 300 I

~ '"d I CIl

5A '"d o CIl 200 !-=1 200 4A o

!-=1

100 100

O 4 8 12 16 20 24 O 4 8 12 16 20 24

Oeflection - Mmm) Deflection - 6(mm) (a) (b)

600 3A 600

500 500 .---. Z

18 .!<:: '--' 400

.---. Z ~ .!<:: I '--' 28 '"d 300

CIl ~ o

I 20 !-=1

200 58 '"d CIl o

!-=1 100

O 4 8 12 16 20 24 O 4 8 12 16 20 24

Deflection - 6 (mm) Defl ecti on - 6(mm) (c) (d)

600

500 H >11 400

z .!<:: '--' 300 ~

I '"d CIl 200 o

!-=1

100

O 4 8 12 16 20 24 Deflection - 6 (mm) Location of H and 6

(e) (f)

Figure 3. Load-Deflection Curves For Test Specimens.

809

LOAD )

423KN illKNI . 1..~.C 4~JN ~ ~ r 1 445 KN I'\. 1 r 1 I···LT

DI 4123 K'" 44SKN 1 1 1 1 1 1 I 1 1 1 1

7' KN \. L24~I<N 1 1 1 I 1 1

1 L 1 1 1 T 1 T , 1 I 1 T T r -. I:~r::

1. 1 L 1 1 1 1 1 PIlIfN I 1 "l..

L , 2á3KN I 4.Hf14 I

Specimen lA (a)

1

LOAD~~ ______________________ --,

LOAD >

Specimen 3A (c)

I I 78'j(N 1 1 T ! 1.. 1. !.!7.I<N 1711KIII I I

_L ..... 7KNI 2GOKN 1 2ziKN 1 oi 22~KN 1

1 L I -. 12ÕQKN 1 I l T 1" , I

1 I L I 1 -. 1 1 1 I T ~ I I

1 1

1 I

I I

I I I 1 12231<N 1 I 1200lj:N , , 1 1

1 I 1 I

, 1 \.. 1 I 1 1 -. 1 1

1 I 1 "l 1 1 1 1 2123KiJ

1I)6KNI

Specimen 5A (e)

I I , I

r

LOAD •

LOAD >-

LOAD >

Figure 4. Crack Patterns for Series A Specimens.

810

111 KN .142311<N 1 , , 1 1 1 ..1 .~.J.414kN " , , f , I i !li4i("N ~ <1213 KN 1 1 1 1 I

I r -, T 1 .J

1 1

r 1

1 1 1 1 1 1 , -. , f , ] 1 ~ 1 1 1 -,

h414N"':'-i\..4IhN 1 T ~ 1

1 1 4IkKN1-.. 1 1 1 1'-"1. J

'1'. 'SI: 414 I<N!'.. """-I

IS!lI<N

Specimen 2A (b)

l'

]

T

J

I

"34~ 1 1 1 ISII<NI 1 I 78KN

1

1

,,,III'N I I I 1169""" J 20dKN 1 1 1 1 16911:N 1

1 1 I

I i 1

I I

20dKN 1 1 I 1 I 1 L 201lll<N1 1 , L , 1 I J , J

1 1 I T , T I 1 1 1 ""l ~

1 , , , , 1 T 1 I 260KN

I 1 1 I IMKN

1~41(M

Specimen 4A (d)

1 -.-1 I '

~""ICN , ..... " I 1 *'21<11 1 1 1 1 r

I 3615kN 1 1 I 1 1312..1N .1..., 136~IKN 1 1 , , 1312Kt.l ~-"*N~ 3U KN 1 I 1 J

1 1 1 36'I<N I J ..l 3li KN~ 4bIKNI\. 1 1 I 1 ., 1

I 1 .40iicao.. 1 13S~f(N I 1 4dlKN 1 .......r 1~36I11KN'1 32.lIKN 1 I 1

1 1

1 ... 1 1 1 1

" "n~KN 1 L "lo. 1 1

I Jl:386KN "IR 1(11 .2IIl7KN"

Specimen 6A (f)

f"o..

LOAD?

I 420j::N I I I .4afKN I I I '3~~KI'{ '\l400'JN ~ I I

1310KfoI 1400IKN I r r I I " 4pO K~ 420l<NJ I

I cn. I I

I I

I I I

I

I I I

I l 1 I I

I I I I I I I pOKpt I I l' [ I

I

I

I ~ I I 1 I T -x I

4~OKN13551to1 I I ~ I I T T 27ti<N'~NI ... \..

I I T 1 ""'-I I I 3101Gol'

Specimen 1B (a)

64KN 1 55KN :L I' KNl ".01tN I

B2KNt:3KN Á I II OtN I I

I I I I

I I I I

I I 1 I

I I I I

I I I I 55!<N

C

I T T I I I

I I I I I I

I I I I I I

I I I I I 1

I I I 1 I I

-' 5fiKN 192KN

" Specimen 3B

(c)

133 KN

I I

I I

I I

I I

I I

I 17a1KN I 190KIf 1218K~ 245. N I

19QK~

I I I

I I I

I I 1

I 1 1 9Q KN

., 245":N

I I r 1 1 78TN 1 I 1 1 I I

T 1 T lzlsK1I -r 1 -I I 1 1

I I 1 I 1 1

1 1 I I 1 1

I I 1 I I 1

90KN

Specimen 5B (e)

1 r 1

I 1 I

I I I

I l I

Figure 5. Crack Patterns for Series B Specimens.

811

I I I I I ~31O.g. I r I BIOKIl I I J I I

~ I J 21451ai1 I I I 1210~ ' I 1 I

24'100; 1IoKHI I I 1 1245p<N 1 :l I ~ I I

I I I I I 1 I 1

T 1 L 51""N I

I I I I I l I I I I I I I I L I

I _ L2~ I I I .. IItN I I

Specimen 2B (b)

133 KN

210Krl I I

1 1

I I I 1 166 K. i I I T266TN IIIIK~ I IlIlKfi 1

r I

I I

r I

I

1

c:.

166,q, A 26aKN'

I I I

I I 1 I I

1 I I 1

1 I

I I I seilH

13

Specimen 4B (d)

74 KN 67 KN

I 1133KN II!6KNI I !iJOKN 385iN I

I 8, KN~

I I I I I

I 1 I I I

I 1 I

I I J

8!11KN

74KH

l"" 94K~ IIOK;"

1\ t I r 67 ~N J 2221K I I ~1\. r I I l'\:" 1\ 36~KN I I ~l' I I

13«15 KN l'J"Sl I 1 I '" .J I

I I '-' "\l I I I 1 "'I

;3116 KIt '56~

J 1 I "" \.1\ ,

Specimen 6B (f)

I 1 1

T _L 1

I T I

I I I ~;q""" I I

The effects on cracking of a roof beam airspace along the top of a panel are shown in figure 4(d). The majority of cracks occur along horizontal joints with only a slight suggestion of diagonal cracking on the side remote from the load. The effects on cracking of using ties in such a wa ll are shown in Figure 4(e) for Specimen 5A. The ties appear to reduce horizontal cracking through the joints while increasing diagonal cracking.

Crack patterns for Specimens 38 and 48 which had centrally-located doorw~y openings are shown in figure 5(c) and 5(d) respectively. Cracks occurred primarily through horizontal joints and, as loading increased, very large crack widths developed at points A, 8, and C as noted. Figure 5(e) shows the crack pattern for Specimen 58 which had a doorway opening close to the loaded side of the specimen. Again, horizontal shear cracking through the mortar joints was quite evident although some diagonal cracking developed in the side remote from the load. In contrast, Specimen 68, which had an opening remo te from the load, developed extensive diagonal cracking in the panel between the opening and the load. This apparent diagonal strut action resulted in a much higher ultimate load than that obtained for Specimen 58.

3.4 Comparison with Theory

3.4.1 Load-Deflection Curves. Theoretical comparisons are valid only in the elastic pre-cracking load ranges. Theoretical load-deflection curves in figure 3 are compared with corresponding curves obtained experimentally. Generally, theoretical curves agree with test curves in the lower, pre-cracking load ranges. As the loads increase and cracking occurs,theoretical and experimental curves deviate markedly. The theoretical method gives a reasonably accurate estimate of initial stiffnesses in most cases.

3.4.2 Stress-Contours. The stress contours presented herein are valid only up to initial cracking of the masonry panel. However, there appears to be some correlation between theoretical stress contours and observed crack patterns. A complete study of stress contours and crack patterns for several of the specimens tested as well as for many hypothetical cases is available else-where (7). As examples, figure 6 shows the stress contours for major principal stresses and maximum shear stresses for Specimens 2A, 6A, and 38. The arrows in these figures show the directions normal to major principal stresses and parallel to maximum shear stresses,respectively. Specimens 2A and 6A were standard specimens with horizontal joint reinforcement and differed only in that Specimen 6A also had panel-to-column ties. The presence of these ties is evident by comparing stress contours for these specimens which were assumed to have a load of 220 kN applied at roof level. A comparison of the stress contours for Specimen 2A with the corresponding crack patterns shown in figure 4 reveals a correlation between high maximum shear stresses and horizontal cracking along the roof and floor beams as well as a correlation between diagonal cracks and maximum major principal stresses and their directions. For Specimen 6A the crack pattern is much more erratic due to the presence of panel-to-column ties. Again, there is reasonable correlation between the normal and shear stress contou r s and the cracking pattern observed. Diagonal cracks appear to have formed along directions normal to ma ximum major principal stress directions while horizontal shear cracks appear to have formed along maximum shear stress paths. The action of ties suggested by the major principal stress contours is also somewhat noticeable in the crack pattern. Major principal stress and maximum shear stress contours are also shown in figure 6 for Specimen 68 which had a centrally located doorway opening. A correlation between these contours and the experimentally determined crack pattern shown in figure 5(c) is also evident.

812

4.0 CONCLUSIONS

Twelve large-scale specimens consisting Df masonry infilled steel frames have been tested to determine their behaviour and strength when subjected to an in-plane horizontal shear load at roof level. The specimens were varied to evaluate effects Df panel-to-column ties, mortar strength, panel-to-frame f r iction and bond, airspace along roof beam, and openings in the panel. Initial stiffnesses and panel stress contours obtained from a proposed analy­tical technique were compared with experimental findings.

In view Df many random variables such as micro-cracking, variation in mor t ar strength in a panel, and stress concentrations, etc., which affect the strength and behaviour Df a masonry panel, co r relation between theory and test results was considered to be satisfactory. Column-to-panel ties result in extensive off-diagonal cracking which is felt to be undesirable in a load-reversal situation. (It is felt that a crack pattern tightly banded about the compres­sional diagonal results in a more desirable structural behaviour). Additionally, ties appear to increase the initial stiffness but do not appear to be beneficial in increasing either the initial major cracking load or the ultimate strength of masonry infilled steel frames. Provision Df an airspace along the roof beam at the top of a panel infill results in a decrease in initial major cracking load and ultimate load of about 50% and 60%,respectively, Df that observed for similar specimens with no airspace. The use of panel­to-column ties in these specimens does not appear to significantly improve their in-plane strength and behaviour. Comparing Specimens lB and 2B of table 2, it is indicated that a reduction in mortar strength of about 50% results in a reduction Df cracking strengthand ultimate strength Df about 40% and 25%,respectively. Comparing results for Specimens 2A and 3A with those of Specimen lB in table 2, there is some indication that a reduction in panel-to-framebond and friction tends to reduce the initial crack load and the ultimate strength.

Openings in a masonry infill panel markedly reduce the initial major cracking load while the ultimate load may not as significantly be affected. When a doorway opening is remo te from the applied load, the ultimate load is less affected than when the opening is placed towards the loaded side of the panel. It appears that if the opening must interrupt the compressional diagonal, it is preferable that it be placed 50 that as much panel material as possible is between it and the point of load. This gives a diagonal strut effect an oppo r tunity to develop in that part Df the panel. Since, in practical cases, loads will come from either side of a panel, it appears that the best location for a doorway opening is at the center of a panel.

5.0 ACKNOWLEDGEMENTS

Funding for this research was provided by the Canadian Masonry Research Foun­dation and the University of New Brunswick through its research grants programo This support is gratefully appreciated.

REFERENCES

(1) Dawe, J.L., and McBride, R.T., "Experimental Investigation of Shear Behaviour of Masonry Panels in Steel Frames", Proceedings, 7th Inter­national Brick Masonry Conference, Melbourne, Australia, February, 1985.

(2) McBride, R.T., Yong, T.C., Dawe, J.L., and Valsangkar, A.J., "Behaviour Df Masonry Infilled Steel Frames Subjected to Racking", Canadian Society of Civil Engineering, Proceedings, National Engineering Conference, Halifax Nova Scotia, May, 1984.

813

(3) Polyakov, S.V., "On the Interaction Between Masonry Filler Walls and Enclosing Frame when Loaded in the Plane of the Wall", Translations in Earthquake Engineering, pp. 36-42, Earthquake Engineering Research Institute, San Francisco, 1960.

(4) Holmes, M., "Steel Frames with Brickwork and Concrete Infilling", Proceedings of the Institution of Civil Engineers, pp. 473-478, vol. 19, August, 1961.

(5) Stafford-Smith, B., "Lateral Stiffness of Infilled Frames", Journal of the Structural Oivision, ASCE, vol. 88, ST6, pp. 183-199, 1962.

(6) Mallick, O.V., and Garg, R.P., "Effect of Openings on the Lateral Stiff­ness of Infilled Frames", Proceedings of the Institution of Civil Engineers, London, England, vol. 49, pp. 193-209, 1971.

(7) Oawe, J.L., and Yong, T.C., "In-Plane Behaviour of Masonry Panels in Steel Frames", Report submitted to the Canadian Masonry Research Council, c/o University of Calgary, Calgary, Alberta, April, 1984.

Specimen 2A Major Princ i pa 1 Stress CO i~tours

(a)

Specimen 6A Major Principal Stress Cont ours

(c)

Specimen 38 Major Principal Stress Contours

(e)

Specimen 2A Maximum Shear Stress Contours

(b)

Specimen 6A Ma ximum Shear Stress Contou r~

(d)

Specimen 38 Ma ximum Shear Stress Contours

(f)

figure 6. Maxlmulll Nonnal and Shea r Stres s Contours.

814