:IJI. t~III-I-ellt Itttst-tellsittlletl :llltl Itl-estressetl bl-icl(\\Tttli( :llltl cel-:llllics ill t;l-e:lt III-it:lill.

Kellllelh Thomas, Clay Producls Technical Bureau of Creal Brilain, Lld. , London, England

This chapter presents a sununary af some Teccnt work in Great Britain on post-stressed and pre-stressed brickwork and ceramics.

The investigation described each relate to beam construc­tion, and although the three projeets were sponsored by separate bodies, liaison existed through the Clay Products Teehnical Bureau of Creat Britain. As a direet result of the work described, a British Patent has been registered for a ceramic flooring system.

This chapter is in three parts. Section 1 describes work earried out by the author; sections 2 and 3 relate the results of work by Dr. J. M. Plowman and Mr. L. S. Ng respectively , the author providing lhe commentary.

History of Work in lhe Uniled Kingdom

Post-stressing and pre-stressing techniques are usually asso­ciated with cancrete construction, yet records show that post-stressed brickwork was successfully employed over 140 years ago for part of lhe Thames Tunnel Projeet 1.

In 1825 Marc Isambard Brune!'s plan for lunnelling under the River Thames was carried out, his scheme being based on the sinking of a caisson on either side of the Thames alld tunnelling between shafts. The caissons were, it is believed, not only lhe first attempts to reinforce briekwork but also the birth of the technique known as posl-s tressing.

Thc eaissons consisted of two vertical brick tubes 50 ft. diameter and 70 ft. high. The 30-in-thick walls were reinforced vertically wilh l-in-diameter wroughl iron bolts built into the brickwork and fastened to wooden curbs at the lop and bottom with nuts at thc threaded ends of lhe bolts. Horizontal

285

reinforcement consisted of iron hoops 9 in. wide and 1/ 2 in. lhick.

Building of the shaft was eompleted above ground with temporary piles supporting lhe strueture which ullimately weighed 910 tons. After completion of the brickwork and tensioning of the wrought iron rods. the pile:;: were graduaUy removed. Sinking of the shaft was engineered by loosening lhe earth within the brickwork enclosure and removing it by means of a windlass and buckets, the shaft descending under its own weighe During the operation of sinking the shaft , the whole structure was severely tested due to uneven strata ; an alarming surge took place and the tower suddenly dropped seven inches at one side and only three and a half inches 00

the opposite. Throughout the whole operalion, the struclure behaved as a composite unit, and work was completed without injury to the brickwork. The structure remains to this day , an outstanding example of reinforced and post-stressed brick­work.

Since the pioneer work of Brunel , developments in post­stressed and pre-stressed cera mies have been rather disap­pointing, no doubt due to the extensive use of reinforced and pre-stressed concrete. Current developrnents suggest a renewed interest in this form of construction.

Many types of stressed ceramic construction have been suggested , particularly in the form of beams and planks for prefabricated nooring systems, but perhaps the more inter­esting uses and suggestions relate to walls.

In Darlinglon , Counly Durham, l l-in-cavity walls have been constructed approxirnalely 24 feet high and 30 feet long with post-stressing rods in the cavity. This form of building permitted slender construction which would otherwise not have been possible without additional structural members ; Ihus the wall gave greater archilectural freedom without additional cos!. The building is described in more delail elsewhere 2 .

286 Designing, Engineering, and Constructing with Masonry Products

Tests have recently been carried oul 00 simulated cross-wall projections using reinforced concrete fIoar slabs compositely with reinforced brick walls. In Doe instance , post-stressing steel was used vertically to restrain cracking. Details af these tests are described elsewhere 3,4

Other work af interest in this field concerns pre-stressing af brickwork shear walls and walls 011 reactive soils5,6.

SECTION I

Testing

Post-Stressed Brickwork Beams- Belfast College of Tech· nology, Summer 1963

During the Summer of 1965, lhe author was responsible for testing tWQ post-stressed brickwork beams, to determine the possibilities af resting a suspended floor 011 beams constructed in this manner.

Figure 34-1. F/oor p/an (1st. project).

11 was appreciated by the architect and consulting enginee/ that this form af construction , using standard perforated wire-cut bricks, was unlikely to provide an econom­ically viable method of supporting a suspended noor, but as the client was a brick manufacturer inlerested in brickwork as a structural material , it was decided to prepare a preliminary design and initiate a smaJl testing programme.

The supporting structure for the suspended floor was intended to be a diagrid of 9-in. wide x 18-in-deep brick members with 9-io x 9-in x 18-in deep pre-cast concrete junction pieces (to facilitate post-stressing cables in two directions). See Figure 34-1. The diagrid system should be constructed on scaffolding and post-stressing cables threaded

·Architects: R. J. Mc Kinstry , Belfast , Consulting Engincc~ : Kirk , McClure and Morton , Bclfast.

through holes in the brickwork to anchor blocks, and then the whole framework post-stressed in situo AlI beams were to have been supported on 9-in. loadbearing panels of brickwork and the 3-in. pre-cast concrete floor slab units laid over each square of the diagrid system aner post-stressing. IntemaUy, it was proposed to express the coffered ceiling structure and to have the concrete squares lined in ao accoustically absorbent material. Unfortunately , the design was nol built , but the enthusiasm initiated io Belfast for this form of construction was nol ]ost; indeed, it gained momentum.

Structural Tests

Beam No. I . The beam was composed of perforated bricks and constructed on a timber plank, laying ali units as "soldiers" with mortar joints. Placing of bricks in this manner is a slow and tedious business but was considered preferable to building short lengths vertically and risking damage due to handling. To prevent mortar penetration in to the lower perforation a simple

,I

"pu li though" was devised by the bricklayer and consisted of rubber washer secured to a length of galvanized wire. This elementary device proved extremely effective. Bricks used fo the beam were wire-cut rustics having ao average compressive strength of not less than 4,000 Ib/ in2 and dimension approximately as indicated in Figure 34-2 . Mortar consisted o a 1:3 mix (by volume), ofmasonry cement/sand .

End aochorages were made up of a standard cast-iro distribution block, the reioforcement consisting of a simp~e spiral and light cage. Concrete was 1:2:4 m (cement-sand-aggregate by volume) , the cement belng "Fondu".

After constructing the brick soldier section beam in position on a nat bed , the brickwork was covered and kep permanently damp until testing of the beam took place. Afie 28 days the concrete anchor blocks were fítted to the be

Cu"ent Post-Tensioned and Preslressed Brickwork and Ceramics in Greal Bn'tain 287

ends. and a thin layer of mortar used to ensure that local high-point loading would not cause premature failure of the beam adjacent to the anchorage.

Six 0.276 (7 m.m.) high-tensile steel stressing wires were then threaded through the lower of the three holes in the perforated bricks and the anchorage block and a light stress applied to facilitate handling of the beam prior to testing. The stressing af wires was carried out using the P.S.C. mono wire system.t

The test rig and ancillary equipment consisted of the following:

I. Rolled steel joist overhead beam supported on brick piers.

2. Mild steel hanger flal bars with bossed holes aI 6 in.' centres.

3. Supporting steel rods with split pin fasteners . 4. Mild steel distribution saddles for points of support.

Figure 34-3. rest rig ful/y assembled supporting 8' 3" spal! beam (1st pro;ect).

S. I 5-ton hydraulic jack. 6. Load cell with gauge attached (20,000 Ib/in 2 capacity).

Figure 34-3 shows the test rig fully assembled, the beam in position and a final force of 15,000 Ib. applied to the cable. During the stressing processo it was necessary to apply a small force to each wire in lum rather Ihan lhe full force af 2,500 Ib., as ralalian of the cancrele anchor black due to full stress an the adjacent wire tended to disladge olher wires fram the distributian block .

When the test rig was initially considered. it was intended that each beam shauld be subjected to equal verticallaad at third paints af the span, but because anly one jack and load

tp.s.c. Mono \Vire Systcm supplied by P.S.C. Equipmcnt Limited, Ridgc Way. Ivcr, Bucks.

= •

I'

êi I ,.-l--+-1:IT

2. I •

2-': "oI---t-~~

2i

Figure 34-2. Cross section of brick for beams 1 al!d 2 (1st pro;ect).

cell was available at that time, it was decided to apply a central point load , jacking the brick beam against the overhead steel member.

Final st ressing of the beam and testing took place within a matter af haurs. Due to the experimental nature of the work it was decided not to graut the cable in position. Ideally , the hole for the post-stressing cable should have been within the "middle third" of lhe beam depth, thus ensuring that the whole af the bearn cross section was in campression. For this pilot investigation , the brick company's standard brick with a three-hole perfaration pattern was used, the centre af the bottom hole being 2-1/4 in. from the beams's neutral axis and thus outside the middle third.

After final stressing af the post-stressing wires , the beam was carefully positioned with supports at 8-fl. 3-in. centres and the hydraulic jack and load cell placed in the centre of the

288 Designing, Engineering, and Constructing with Masonry Products

beam. Load was gradually applied aI lhe mid poinl aI lhe span by lhe hydraulic jack. As no adverse effecls were apparenl aI a load of 4,045 lb. (1.75 lhe design load of lhe beam, lhe maximum estirnated design stress based 00 nominal dimension being 750 lb./in2) il was decided lo remove lhe cenlral load and increase lhe force in lhe posl-slressing cable lo 24 ,000 lb. On release af the central point load, the deflection returned approximalely lo zero. The load/defleclion graph (Figure 34-4) indicales Ihal after inilial sliffness lhe beam had a direcl elastic relationship.

Prior lo reapplying lhe cenlral poinl load, lhe beam was partially wrapped in polylhene sheeling as a safely precaulion. The load was applied for lhe second lime wilh lhe load deflection curve following an almosl idenlical palh lo Figure 34-4 bul when a load of 3,850 lb. was approached, lhe beam failed. Figures 34-5 and 34-6. This failure was considered lo be due to the principal tensile stress elose to suppart being greater Ihan lhe resislance of the fired clay. Beam No. 2. This beam was composed of8-3/4 in. x 4-3/16 in. x 2-7/8 in. bricks as for beam No. 1 bul was formed inlo a cross seclion nominally 18 in. deep by 9 in. wide. Figure 34-7. The lolal lenglh of lhe beam was 20 fI. The beam was constructed in a similar manner to its predecessor but provision made for Iwo posl-slressing cables. Figure 34-7. End anchorages were precast concrete blocks similar in design to the first beam, but having twio cast iron distribution blocks, see Figures 34-8 and 34-9. Beam No. 2 was conslrucled aI lhe same time as beam No. 1, and the same materiaIs used for each. Tesling again look place aI 28 days. Numerous difficulties were encountered due to each cable consisting af six slrands. Blockages occurred despile lhe "pul! Ihrough", alld lo Ihread lhe cables evenlually four courses of brickwork had lo be removed before lhe lesl could proceed.

Tms beam was slressed on lhe horizonlal plank bed, starting 011 cable 1 and stressing ooe wire then stressing one wire in cable 2, bul before lhe slressing was complele , cracking slarled in lhe beam immedialely behind lhe anchorage. (Figure 34-9). Failure was again considered lo be due lo lhe principal lensile slress adjacenl lo lhe supporl being grealer than lhe resislance of lhe fired c1ay.

Conclusions Both beams failed due to excessive tensile stresses. In beam

No. 1, the principal tensile stress at failure was estimated at 120 Ib/in .2 and beam No. 2, 90 Ib/in2 These values, however, assume the joints to have the same properties as the ceramic and are therefore somewhat conservative. If this form of construction were to be adopted, it is considered that the principal lensile slresses should be no grealer Ihan perhaps 60 Ib/in 2. To acmeve a smaller slress adjacenl lo lhe anchorages eilher (a) a larger cross-seclional area is necessary, (b) units having a mgher lensile slrength are required, or (c) reinforce­menl should be inlroduced inlo lhe bed joinls lo reduce lhe lateral strain of the mortar. This would increase the com­pressive slrength of lhe beams in lhe lenglh 7,8

Alternalively, lhe use of self-supporling slabs of posl­stressed ceramic would disperse the forces over a larger area. New unils would need lo be developed for Ihis.

SECTION 11

Tesling

Post-Stressed Brickwork Beams Sponsored By the Clay Pro­ducts Technical Bureau of Great Britain, Testing Carried Out By Or. J. M. Plowman·· - Spring 1965.

The Clay Producls Technical Bureau of Greal Brilain were consulted regarding lhe Belfasl projecl and as a direcl resul! decided to 'sponsor a similar research programme. They decided to investigate whether ceramics were suitable for posl-slressing and if Ihey could be used for flooes elc.

For Ihis pilol invesligalion, Ihere was no need lO go to lhe expense of designing and making special unils which might need modifying during lhe lesls. Therefore, slandard size bricks were adopled wilh holes aI lhe Ihird poinls. Although the brick size was standard, the hole positions were not standard and it was necessary to drill or punch solid bricks whilst still in the green state, in some instances.

For Ihis projecl, all beams were fabricaled by laying the bricks as "soldiers" wilh morlar joinls. By placing lhe slressing wires at the lower third point, the whole cross section of the beam was in compression whilst under axial stress. In view Df this arrangement, it was unnecessary to provide secondary sleel lo withstand handling slresses.

Laying of lhe bricks look a considerable lime because the basic unil was onJy 2-5/8 in long so Ihal Ihere were over forty joinls lo make a span of 10 ft. The beams did nol have lhe holes grouled.

Tests

Figure 34-10 shows a beam in lhe lesl rig . All beams lesled were idenlical in dimensions (4-1/8 in. wide x 8-5/8 in. deep x

**J. M. Plowman, Consulting Enginecr, Hatficld, Hcrtfordshirc.

a,OOo

a,ooo

',000 0

o

o 2

Figure 34-4. Loadl deflection graph [or beam 1 (1st project).

Cu"ent Post-Tensioned and Prestressed Bn"ckwork and Ceramics in Great Bn"tain 289

Figure 34-5. Fai/ure of 8 ' 3" span beam view adjacent to sllpport (1st project).

Figure 34-6. Failure of 8' 3" span beam view from end of beam (1st project).

$$$

t ~ ft- I e,"(rJo ~, -( ~ -4 .. CA8~ e

~ ~

$ $ $

Figure 34-7. Cross section of beam 2 (1st project).

,

Figure 34-8. 20' beam on supporting form prior to final stressing ( 1 st project).

- '" r > " -; ~

Figure 34-9. 20' beam showing end anchorage and method of fai/ure (1st project).

10 ft. long, centres of supports). The method of applying the central point load was similar lo the Belfast project, although a 20-ton proving ring was used in this series of experiments instead of a load cell for recording the applied load.

The large forces from the anchorages were concentrated on a smaU area due to the design of the patent cast-iron anchor p]ates, and a concrete distribution block was therefore provided at each end of the beams. A I: 1\2:2\2 concrete mix was adopted for the blocks using high aJumina cement with a 3/4 in. maximum sized aggregate, giving a 28.oay cube strength of7,300 Ib/in2

All beams were constructed on leveI supporting forms to ensure absolute straightness, the sequence being as fol1ows:

1. Bricks laid as soldiers using mortar as in Figure 34-10, each brick being dipped in water for approximately 1/2 minute before placing" Jointing was carried out concurrently with placing. Tables 34-1 , 34-2 and 34-3 give properties of the bricks and mortar used.

2. Mild steel bars 3/8 in. diameter were passed Ihrough the two lines of holes at the third points. Nuts on the ends were

I

290 Designing, Engineering, and Constructing with Masonry Products

Figl/re 34-10. Test rig jUl/y assembled supporting 10' span beam (2nd project).

Table 34-1 Crushing Strength of Bricks

Brick A 7,910 Ib./in 2 (Av.) Standard Deviation

1034

Brick Bb 4,830 Ib./in2 (Av.) Standard Deviation

692

Brick Ba 5,170 Ib./in2 (Av.) Standard Deviation

367

Brick C 3,840 Ib./in 2 (Av.) Standard Deviation

526

Table 34-2 Water Absorption Tests (24 hrs.)

Brick A Brick Ba Brick Bb Brick C Av. 11.4% Av. 11.45% Av. 9.72% Av. 9.7%

Density

1381b/ft3 1261b/ft3 I 128 Ib/ ft 3 122 Ib/ft3

Table 34-3 Mortar Cube Strengtns and Sieve Analysis of Sand

Mortar Mix (By Weight) 1 : 0·11 : 3 Portland Cement : Lime : Sand

Beam No. Average Cube Strength (4 in)

I 32701b'/in 2

2 3600 3 3520 4 3150 5 3150 6 2420 7 2700 8 3120 9 2690

10 2810 II 2690 12 2690 13 2810

Grading of Sand

Sieve Size 3/16;n.\ 7 ] 14 25 ] 52 % Passing 99 ] 96 I 93 81 I 27 ,

100 8~

... ~-..;"

Current Post-Tensioned and Prestressed Brickwork and Ceramics in Great Britain 291

Figure 34-11a. Cracking ar top flange. 10' span beam (2nd praject).

ligh tened on to washers bearing on lhe end bricks. Thus, the joints between the soldiers were kept in compression.

3. After 24 hours lhe nuls on the bars were lightened further.

4. AI 27 days lhe 3/8 in.-diameler bars were removed and lhe required number of 0.276 (7 m.m.) diameler high lensile wires Ihreaded Ihrough lhe lower of lhe series of holes.

5. The concrete distribution blocks were threaded on to the wires and bedded in high alumina morlar against the end soldiers.

6. Afler lhe mortar had sliffened the wires were slressed lo give an end force of 1,000 Ib.

7. AI 28 days lhe end force was increased lo 3,000 Ib. and lhe beam placed in lhe tesl rig. Figure 34-10.

8. After placing in the lest rig, lhe wires were furlher stressed lo the desired value. rable 34-4.

9. The beam was lesled lo faHure (Figures 34-1 la and 34-1 1 b) by loading aI lhe mid poin t, measuring defleclions aI suitable increments of load.

10. After failure in some instances , four soldiers mortared together were salvaged from lhe undamaged porlions of each bcam. These specimens were lesled in a 200-lon press applying lhe load Ihrough a baU joinl, lhe axis of which passed through lhe one-Ihird poinl of lhe specimen, Ihus, partly simulaling the aclion of lhe high lensile slressing wires. Failing loads of ,uch columns are given in Table 34-5.

Figure 34-11b. Spalling ar tap flange. 10' span beam (2nd project).

Beam No

I 2 3 4 5 6 7 8 9

!O 11 12 13

Results

Top

Table 34-4 Pre-stressing Stresses

Bottom Bottom Brick Str.

281b/in2 2151b/in2 0.043 35 267 0.033

56 430 0 .086 63 480 0.096 59 455 0.091

111 855 0.107 125 930 0.120 146 1120 0.140 48 370 0 .097 67 515 0.151

104 800 0.210

Bottom Calculated Bwk. Str.

0.12 0.16

0 .25 0.28 0.26 0.36 0.40 0.46 0.26 0.36 0.55

Due lo difficulties in oblaining bricks of the specified slrength for the tesls, lhe four batches did not become

292 Designing, Engineering, and Constructing with Masonry Producls

available in the sequence required. Therefore, tests could not follow 10gicaUy (without undue delay) from low stressed , low strength to high stressed , high strength bricks. The results are reported in chronological arder.

Beam I 8rick Ba. Tltis beam was used as a "'guinea pig" to develop lechniques for laying the soldiers and keeping lhe holes for the wire frec of mort<lT.

Beam 11 Brick Ba. Post-stressing force 4,000 Ih. At a load of 500 Ih. lhe centre joinl in lhe beam opened, followed hy lhe adjacenl one at 600 Ih. At 1,100 Ih. several joinls opened , lhe position of these joints being determined hy the friction of the brick and mortar on the wire. The deOection at this load was 1.95 in. At a load of 1,290 Ib. the deOection was 3.1 in. and the brickwork under the load distributing pIa te slarted to spall. No increase in load could be supported. Although the deOection was increased lo 6 in. little crushing of the brickwork occurred.

Beam 111 Brick Ba. Post-stressing force 5,000 Ih. AI a load of 650 Ih. several central joints opened. A load of 1,500 Ib. caused local crushing and a deOection of approximately 6 in. After removal of the load, the residual deOection was appraximalely 1/4 in. The loading was lhen moved c10se to a support to test lhe beam's shear strength. At a load of 2,100 Ih. joints opened as the beam benl. AI lwice lhis load, hrickwork to lhe midspan side of the load failed in Oexural compression. There was no shear failure.

Beam IV Brick Ba. At a post-stressing force of 5 ,000 Ib. one of the patcnl casl iron anchor plates collapsed resulling in lhe destruction of lhe beam.

Beam V Brick Ba. Posl-stressing force of 8 ,000 Ih. Failure occurred aI a load of 1,760 Ih. due to excessive Oexural compression af the brickwork.

Beam VI Brick Bb. Post-stressing force 9,000 Ih. Failure occurred aI a load of 1,720 Ih. due to excessive Oexural compression of the brickwork wilh a final dcOection of approximately 1-1/2 in.

Beam VII Brick Bh. Posl-stressing force 8,500 Ih. Failure occurred ai a load of 1,880 Ih. due lo fracture of the stressing wire close to Dne anchor gripo Maximum deflection approx­imately 4-1/4 in.

Beam VIII Brick A. Post-stressing force 16,000 Ib. When the stressing force reached 11 ,500 Ib. , the first brick developed a horizontal tensile crack along the axis of the wires. This crack spread thraugh the mortar joint into lhe distribution block outside the reinforcement cage.

On removal of lhe stressing force, the heam was found to be little damaged and by removing the first six soldiers at the damaged end it was possible to add a new distribulion block and make a shorter heam. Expanded aluminum mesh was inserted into the joints hetween blocks and soldiers at each

Table 34-5 Column Failing Loads and Brick/Pier Strength Ratio

Calculatcd Max Strcng!h Beam No. Column Failing Load Strcngth of Brick

2 3 51 ,000Ib. 0.6 4 5 125 ,000 1.46 6 100,000 1.25 7 80 ,000 1.00 8 9

10 80 ,000 0.61 I1 40,000 0.63 12 55 ,000 0.87 13 61 ,000 0.97

end. FaiJure af the beam was due to excessive flexural compression of the brickwork. The results have been corrected to aUow for the reduclion in span and can thus be compared directly in the graphs with aU the other beams, see Figures 34-12a , b. and c.

Beam IX Brick A. Post-stressing force 18,000 Ih. Failure occurred aI a load of 3 ,400 Ih. due to excessive Oexural compression of the brickwork. One of the cast iron anchor plates was cracked hy this load. Expanded metal reinforce­ments as in Beam VIII was used in the joints between brickwork and distribution blocks.

Beam X Brick A. Post-stressing force 21 ,000 Ib. This beam failed under a load of 3 ,650 Ib. due to excessive Oexural compression in the hrickwork at midspan. Expanded metal reinforcement was used in the joints as for Beam VIU.

Beam XI Brick C. Post-stressing force 6,900 Ib. This beam reached a point where the deOection increased greatly without any measurable increasc in the load of I ,400 Ib.

Beam xn Brick C. Post-stressing force 9,600 Ib. This heam IVas loaded lo 11 ,700 Ih. and then unloaded. The residual deOection was 0.013 in. after deOection under load of 0.184 in. Loading was then continued to 2,100 Ib., the deOeclion

being 1.358 in. On release the residual deOection was 0.125 m· In this beam, partial failure occurred at 1,650 Ib. by a horizontal crack forming at about 3 in. fram the top of th~ second "soldier" fram lhe load poinl. On increasing lhe loa to 21 ,000 Ih. , this crack extended to the third sold"r foUowing a direction inclined towards the central axis o~ th~ beam. The crack did not ex!end beyond these two "soldler~ and c10sed up almos! completely on removal of the loa . (Figure 34-1 la.)

Current Post-Tensioned and Prestressed Brickwork and Ceramics in Great Britain 293

-~ • ~ .. • • .. f

L I, ~ .. v

f---- -z !.---li

--- I.---- I AI 1\ I, • E N -OI .,/ V ~ V v ;

I 1/ ---rr / -!.--- --- I--5( '1 7 V

:; I-- I o. ••• • •• , L.1MI I ...

li' f--17

• • • L < C T , o • o .. I. I lo , 4-__ I ~ , I~CK S -

Figure 34-12a. Bending moment/deflection graph beams 3, 5, 6, and 7 (2nd project).

v '0 1 V V BI A ~ NC • I- 9 ~ 8 ~ ./ -~ ~ V V !.---~ V:: !.---V • ~ V ~ s

/ V o

---f u V 7 / ./ ~ ,

V " / 1 .. 1 1110 oi

r7 t-- !' o T 'N. ON , .. , 5

I-lia I

7 ,

rr D • , L • c T I o ..

lo .1:> I~ ,1+ l-i'- ,. , CK. ~.

Figure 34-12b. Bending moment/deflection graph beams 8, 9, and 10 (2nd project).

294 Designing, Engineering, and Constructing with Masonry Products

.( 9 8 E,4I M No • .. t V V

i---... 12

: V V V / ~ ---- /'

I 1/ .-/ V /

õ ./ / .. =V;í V - -lI ~ ./

- -~ /

It ./' ~ TU .. O ~ / - -}I ~/ .. ",IT

/1 - 6" ;7

I / o • • " < c- T I o .. • A ,,~ ,. lI/- I· . .~

Figure 34-12c. Bending moment/deflection graph beams lI, 12, and J3(2nd project).

Beam XIII Brick C. Post-5tressing force 15,000 Ib. This bearn faUed in lhe normal way by crushing the top I in. of a "soldier" adjacenl lo the load poinl. Failure occurred at a load of2,700 Ib. (Figure 34-llb).

The bending momenljdeflection graphs, (Figure 34-12a, 34-12b, 34-12c) are grouped according lo lhe basic slrength of the ceramic used. It will be seen that there is a discontinuity in each of the curves corresponding lo lhe load at which lhe beams cracked in the 10wer ar tension zone. ]n alI cases, these cracks took lhe form of a bond failure belween lhe mortar and the ceramic. Because of the necessary tolerance between the diameter of the holes in the ceramic and the size of the wires, the aclual posilion of the wires at faUure cannot be exactly known. When a bearn bends lhe wires, being free in lhe holes, wilI tend lo louch lhe upper boundary of the holes aI mid-span. This will reduce lhe eccentricily of lhe posl-slressing force, and Ihus ils efficiency in resisting the applied loads. Using these two limiting positions for lhe wires, (lhe cenlre of the holes and the top of the holes) lhe loads at which normal pre~stressing theory predicts zero tension have been deter­mined. Thus, for any one beam there is a narrow range of load over which tension may develop depending upon the actual position of the wires. These bands have been plotted on lhe graphs, Figures 34-12a, b, and c. lt will be noled thal mosl of the discontinuities in the curves occur at hlgher loads than marked for no tension, thus indicating that the mortar joints have some tensile strength.

The design load for posl-slressed ceramic unils should , aI this stage of knowledge, be that at which zero slress occurs aI the soffit of lhe beam. Basing calculalions on Ihis assumplion, the loads for the two possible posilions of the wires in each of the bearns have been oblained and Ihese are listed wilh the "factors of safety" in Table 34-6.

The tests on the four courses of ceramic in the form of a column eccentrically loaded were intended to determine the type of likely faUure rather than give accurate values. Results obtained are listed in Table 34-5. In ali cases the failure was by splitting parallel to the line of action of the load. Figures 34-13a and b show a typical case.

Author's Comments

1. Ceramics can be post-stressed satisfactorily. 2. Post-stressed ceramic beams behave in a manner similar

to concrete beams.

Table 34-6 Factors of Safety for the Two Positions of No Tension Load

Failing Load Factor of Safety = d

No Tension Loa

Lower No Factor of Upper No Factor af Beam No Tension Load Safety Tension Load Safety

2 190 Ib. 6.7 2301b. 5.5

3 280 5.5 230 4.6

4 5 530 3.3 620 2.9

6 610 2.8 720 2.4

7 580 3.2 670 2.8

8 1200 3.1 1380 2.7

9 1360 2.5 1580 2.0

10 1620 2.4 1860 2.2

11 440 3.2 520 2.7

12 670 3.2 780 2.7

13 1110 2.4 1280 2.1

Cu"ent Post-Tensioned and Prestressed Brickwork and Ceramics in Great Britain 295

Figure 34-/3a. Failure or brickwork column in 200 tOIl testing machine (211d project).

3. Shear failure is not a major concem. 4. Some shear reinforcement may be required adjacent to

the anchorages in some cases. 5. As with concrete , the recovery to detlected beams on

removal Df load is good. 6. Failure of ceramic beams in compression is by slow

crushing Df the material , whilsl supporling lhe load wilh increasing deflection.

7. Failure by sleel failure is as wilh concrele, sudden and complele.

8. Local lensile failure is always in lhe bond belween morta r and ceramic.

Discussion

There seems no technical reason why post-stressed ceramic unils should nol be used for beams and tloor slabs which will compete economicalJy with concrete or any other material.

Further tests are of course, required to develop suitably shaped unils. The major problems in lhe developmen I of such beams and slabs are:

I. Reduclion of number of joints by making longer ceramic units.

2. Joinling malerial of high lensile bond slrenglh easily applied.

3. Grouling wilh portland cemenl or olher adhesive Df wires to ceramics.

4. Long-Ierm behaviour. 5. Dcvelopmenl of suilable anchorage units. Wilh rcgard lo Item 4 il is Ihoughl Ihal creep and

shrinkage , both properties whkh are a cause of concem in Concrete Ooors , will be very much less with ceramics.

Figure 34-J3b. Fai/ure or brickwork column in 200 ton testing machine (2nd project).

SECTION 1Il

Laboratory Investigation into Extruded Clay Block Flooring System

This invesligation was carried oul by L. S. Ng, a fourth year degree sludenl in the Civil Engineering Departmenl of Sunderland Technical CoHege during lhe 1965-1966 academic year.

The objecls of lhe projecl were (a) lo delermine lhe physical properties of the extruded c1ay units , such as crushing slrenglh, modulus Df rupture and creep; (b) lo lest lhe bond strenglh of lhe jointing material and (c) lo find lhe u1timale load on a pre-stressed ceramic beam and compare it with the Iheoretical ultimate load. HoHow c1ay blocks used for aU the lesls were as illuslrated in Figure 34-14b.

As hollow c1ay blocks are cheap and Iight compared wilh other building materiais a new cera mie flooring system might result from Ihis invesligalion. The sludent's employerstt were also in leresled in such a system and suggested the projec!.

As difficulties had been experienced by lhe aulhor and Dr. Plowman in their investigations using traditionaJ mortars , it was dccided to bond the ceramic units with thin joints using epoxy resin as lhe bonding agen!. By so doing, il was hoped lha I lhe morlar would exhibit a much smaUer lateral strain and, hence, delay tensile splitting of lhe ceramic, thus permitting much larger forces to be applied along the beam axis.

It is not possible to record in Ihis chapler the fuH details of aO preliminary testing. The results are summarised as follows:

ttMCSSnl. Stecn Sehestcd and Partncrs, later Peter Heath and Partners.

296 Designing, Engineering, and Constrncting with Masonry Products

f I i I I I I

11'

'x'í IlE~M Pé!

11 I

• x'l,.. , .

~. 11' 11' lat~ la}'

Figure 34-14a. Arrangement of beams PI, n, and P3 (3rd project).

Sec.rION "X- X"

Figure 34-14b. Cross section of b/ock (3rd project).

Compressive Slrength (Individual Blocks Crushed)

Balch No. I Average compressive slrength of 5 blocks tested bctween 3/8 in. thickness plywood (3 ply) = 6,425 Ib./in2.

Batch No. 2 Average compressive strength of 4 blocks tesled between 1/8 in. lhickness plywood (3 ply) = 6,750Ib./in2

Young's Modulus of Elasticity (E) (Individual Blocks Tested Statically - Dircet Load)

The average value of E (five blocks) was found to equal 6.32 x 106 Ib./in2 measured over a 6 in. gauge length within lhe slress range O lo 1,830 Ib./in2

Shear Slrength (Small Beams 10 in. between supports - made up of 3 Blocks) 2

The average shear slrenglh based on 3 lesls was 137 Ib./in

Modulus of Rupture (Small Beams)

The modulus of ruplure varied from 341 Ib./in 2 lo 677 Ib./in2 over 12 lesls on beams of 16-in. and 30-in . span.

Two of lhe beams lesled failed in shear aI 63 and 80 Ib./ in2

wltich is considerably lower lhan lhe values obtained in lhe previous tests.

Creep Tesl (Individual Block)

Results were recorded over a 50 day period wilh an applied load of 10.000 Ib. Creep calculalions are based on a 6-in. gauge length ~ compensation for temperature and moisture movement being measured on ao unrestrained block. Creep at ao age af 50 days was 7.42 x 10-5 in/in.

Loading Tesls on Pre-Stressed Bcams

Loading tests were carried oul to examine lhe behaviour af three pre-stressed ceramic beams spanning 10 n. with concen­Iraled loading ai lhird poinls. The beams were formed by joinling exlruded hollow day blocks together wilh epoxy resin . Each beam was pre-stressed with tWQ O.20-in .-diameter pre-slressing wires and grouled wilh I: I cemenl/sand.

F abricalion

The hollow day blocks used in lhis series of lesls were of lhree different lenglhs, i.e. 8-1/2 in., 12-3/4 in . and 17 in. The arrangement of lhe blocks was as shown in Figure 34-14a.

Assembly of lhe beams was carried out on a sloping jig, the mix proporlions of the morlar (by weighl) being 100 parIs epoxy resin: 60 paris hardener: 40 parIs powdered china day passing B.S. sieve No. 120.

Each beam was pre·stressed with two O.20-in.-diameter bright plain wires with an ultimate tensile strength between 100 lo 110 Ib./in2 The lolal pre-slressing force for beams PI and P2 was 8,000 Ib. and for beam P3 , 10,000 Ib. Anchor plales were 3/4-in-thick and Iheir delail as shown in Figure 34-15. An epoxy resin joinl was used between lhe anchor plale and end block lo lake up the uneven surface of the block. AI the dead end of each wire, a sleel collar was filled wilh an eleclrical slrain gauge placed belween lhe anchor plale and the permanent anchor sleeve to measure the change in tcnsion in lhe wire, see Figures 34-16a and b. There were also four electrical strain gauges sluck aI the midspan of each beam to measure lhe slrain dislribulion. Figure 34-17. Afler pre-slressing, alI beams were grouled wilh I: I cemenl/sand.

Melhod of Pre·stressing

The wires were pre-slressed using Gifford Udall's sySlem*** of slressing shorl wires, each wire being jacked individually. Irtilially wires were jacked to the re'luired value and then

••• Cifford Udall's Systcm - This sys tcm is no longcr available in thC U.K.

Curren! Post·Tensioned and Prestressed Bn'ckwork and Ceramics in Great Britain 297

a. 2ã

Figure 34-15. Anchor plates for beams Pi , n, and P3 (3rd project).

Figure 34-16a. End anchorage at jacking end (3rd project).

Figure 34-16b. End anchorage at dead end (3rd project).

Figure 34-17. Electrical strain gauges at midspan (3rd project).

released lo ensure Ihal lhe wedges had achieved a firm grip on the wire, Each wire was then rejacked until there was about 1/8-in. gap belween lhe anchor sleeve and lhe anchor plale. A horseshoe washer, 1/8-in. Ihick, was Ihen dropped in lhe gap before lhe load was finally released.

Melhod of Tesl

Ali beams were lesled lo failure wilh a four poinl loading syslem, Figure 34-18, lhe Iwo applied poinl loads being ai lhe Ihird poinls.

Loading was apptied and increased unlil approximalely Iwice lhe design load was reached, lhe loading Ihen being released to determine the recovery, Beams were then reloaded to failure. Dial gauges were placed at midspan to measure defleclions and readings laken during lesting. AI lhe end of each test, the actual position of the pre-stressing wires, near the midspan were measured.

Resulls

Beam PI . Beam PI was pre-slressed and grouled six days afler assembly and jointing. Tesling look place lhe following day. The IOlal pre-slressing force was 8,000 Ib.

AI a lolal applied load of approximalely 1,200 Ib. (2.3 limes lhe original design load) , a visible crack was observed ai lhe resin joinl near lo midspan. AI Ihis poinl lhe defleclion was 0.73 in. The load/deflection curve (Figure 34-19) shows a large increase in deflection ai approximalely 1,080 Ib. The aclual crack probably occurred ai Ihis load. Afler reaching 1;357 Ib. lhe load was reduced lo 562 Ib. and lhe recovery was 93.4 percen!. Load was reapplied , bul lhe beam deflecled very rapidly wilhoul gaining much load. This was probably due lo lhe beam no longer being capable of laking lension afler cracking. Due lo excessive defleclion , lhe bollom flange of lhe upper sleel beam was louching lhe lop face of lhe clay beam

298 Designing, Engineering, and Constructing with Masonry Products

Figure 34-18. Tes/ rig assemb/y with third poin/ /oading (3rd project).

, ....

ID. - -V/SI/JJ. c.<Ua

V .. V ... I lo",,' /"

I ~ V --q ---, o

:..,;: 11. EC o ER Y .9a,4' %

I o lO l- "'

.., , .. DEFL4 TION

I , ,

~ .. •• 46D ... , .. " ... T/OU DUE. r" AOJlJ TlrD .... K.,I' 01'1 'ti RlNt; I!

/

-:::::=

'" ROA< D ,

", .. L C. F()~

Figure 34-19. Load/deflection CUlve for beam P1 (3rd project).

/' V ..-/

~ ~ ;..--

,. & lUz;. Nt;$. :: ~/NI. AL t;'AU<

, , I I..,) ,- "",,, -' OrAL A.::: "(JAL PE "LEdION' /(J~3/NS.

Current Post-Tensioned and Prestressed Brickwork and Ceramics in Great Britain 299

FigUre 34-20. Failure of beam P3 showing crack in bottom of beam (3rd project).

FigUre 34-21. Failure of beam P3 showing ,palling of beam top (3rd project).

and, therefore, loading was discontinued at approximately 1,300 Ib. At this load, the opening in the resin joint was about 1/2 in. and also some crushing of the ceramic occurred in the compression zone. The pattern of faUure was as shown in Figure 34-17. Maximum total load applied to the beam was 1,357 Ib. After the test, the beam was loaded to collapse by applying a central point load. Collapse of the beam was due to partia! crushing of the ceramic. When the load was removed completely, the beam recovered quite considerably due to the tension in the pre-stressing wire and, on inspection, it was noted that the grout in the duct was stil! soft and the pre-stressing wires had lifted to the top of the duct due to the deflection of the beam_ Thus, this beam could be considered as ungrouted.

At an applied load of 1,080 Ib. the deflection at midspan inereased rapidly though no craeking was observed. From the strain diagram at this load, assuming an E value of 6.32 x 106

lb./in2, the maximum compressive stress at the top fibre was 1,360 Ib./in2 and the maximum tensile stress at the botlom fibre Was 815 Ib./in2. These values do not take into aeeount creep or moisture and temperature movements.

Beam P2. Beam P2 was pre-stressed and grouted three days after assembly and jointing. Testing took plaee after a further seven days. The total pre-stressing force was 8,000 Ib.

Loading was applied in approximately 100-lb. inerements to 878 Ib. and then redueed to 604 Ib ., reeovery being 94.6

percen!. Loading was lhen increased to collapse. At the end of the test, the actual position of pre-stressing wires at mjdspan was measured and their eccentricity found to be approximately 0.775 in. instead of the intended eccentricity of 1.085 in. This was because part of the resin joint had squeezed into the duct during gluing of the beam_ The safe allowable applied load based on an eccentrieity of 0.775 in. was 380 Ib.

The first crack was observed at a totalload of I ,144Ib . and occurred at the resin joint. As the load increased to approximately 1,307 Ib. the second crack was seen and the load dropped to I ,162Ib. This crack developed into horizontal cracks in the compression zone and crack three followed. The beam failed by excessive cracking along the joints and finally collapsed by crushing of the ceramic at a total applied load of 1,639 Ib. A Load Factor of 4.3 was calculated based on an aUowable applied load of 380 Ib. the denection just prior to collapse being approximately 4 in.

The maximum load before cracking was taken as 1,062 Ib. After this load, deflection began to increase rapidly. From the strain diagram at this load, and assuming E = 6.32 x 106

Ib./in2, the maximum compressive stress was 905 Ib./in2 and the maximum tensile stress was 732 Ib./in 2.

Average losses of pre-stress at transfer and working loads was 2.5 percent and 10 percent respectively. The losses were taken from the percentage change of strain gauge at the end of

300 Designing, Engineering, and Constmcting with Masonry Products

the wires, and did not take ioto aecount the frictional force between the graut and the wires.

Beam P3 . Beam P3 was pre-slressed and grouted seven days after assembly and jointing. Tesling look place after a furlher three days. The pre-stressing force was 10,000 Ib.

Loading was applied up to 970 Ib. and then reduced lo 604 lb. recovery being 88 percent. Loading was again increased until coUapse occurred at 1,579 Ib. The average eccentricily of pre-stressing wires at midspan was measured after the test and found to be 0.825 in. A safe allowable applied load based on this eccenlricily was 5361b. giving a load factor of 2.95.

At a load of 1,185 Ib. a visible crack was observed ai the bottom of the block, Figure 34-20. As the load increased, lhe crack opened and the top of the block began to spaU away. AI 1,428 Ib. the continuous defleclion made it difficult to maintain a steady load. At 1,447 Ib. another crack appeared along the resin join!. The maximum load reached was 1,579 Ib. then the beam collapsed suddenly with the top portion of the block complelely disinlegraling, see Figure 34-21.

Dne would have expecled lhe beam to caITy a higher load than Beams PI and P2, since il had a higher pre-stressing force. However, its collapse load and cracking loads were approximately the same as for Beam P2. Also Ihis was the only beam which cracked at the block instead of at lhe join!. Therefore , the cause Df failure may have been due to a particular weak block bul furlher testing is obviously necessary to verify this. The relalionship belween applied load and defleclion was recorded and up to 1,000 Ib. lhe relationship was linear but just before any visible crack was observed the deflection increased rapidly. At lhe allowable applied load of 536 Ib. the deflection was approximalely 0.49 in., i.e. span/2 ,400 and at Iwice lhe allowable applied load of 1,072 Ib. lhe defleclion was 0.13 in. , i.e. span/425.

The maximum load before cracking was taken as 970 Ib. From the slrain diagram ai Ihis load and assuming E = 6.32 x 106 lb./in 2 , the maximum compressive stress at the top fibre was 1,450 Ib./in 2 and lhe maximum tensile slress at lhe bottom was 753 Ib./in2

The average loss of pre-stress at traosfer was 2.05 percent which was less than beam PI, 4.9 percent and P2, 2.5 percen!. This decreasing 10ss af pre-stress at transfer was probably due to improvement in the techoique of pre-stressing. It is not possible to record the 10ss of pre-stress at the working 10ad due to a failure in lhe eleclricity supply at Ihat time.

Author's Comments

I. The average crushing slrenglh of lhe exlruded clay blocks was 6,425 Ib. / in2 for balch No. I and 6,750 Ib.fin2

for batch No. 2. 2. The stress/strain curve during loading and unloading is

not the same , there is a hysteresis. If the average values of 10ading and unloading curves are taken , then the stress/strain relalionship is linear up lo a slress of 1,835 Ib./in 2 (No lest was carried out above this value.) The modulus of elasticity E = 6.32 x 106 Ib./in 2

3. Due lo lhe brittleness of lhe material the failure in compression is sudden and without warning.

4. From the shear tests , the shear stress at failure was 137 Ib./in2 This value is quite low and should be considered carefully during design.

5. The modulus of rupture from the flexural tests varies from 341 lb./in2 to 677 Ib./in2 Variation is mainly due lo lhe location of the resin joint wilich is weaker than the parent material.

6. The creep observed at an age of 50 days was 7.42 x 10-$ in./in.

7. In the majorily of beams lesled , failure occurred along the resin joints, indicating that the resin joint is not as strong as the pareot materia1. lt is considered that the joint weakncss may be due lo it being toa thick.

8. To increase the joint strength, it is recommended that the ends of blocks should be ground.

9. Since the ends of the blocks are not square, the beam formed by jointing Ihese blocks logelher wil! not be straight. This difficulty can a1so be overcome by grinding the ends.

10. The recovery of pre-slressed beams was found lo be very good. When lhe load had been reduced after reaching aboul twice lhe safe applied load, the recovery was approximately 90 percent. The average Load Factor at collapse was approx­imalely 3.5.

11. From strain gauge measurements on the pre-stressed beams, it was noted Ihal lhe day blocks were able to take a lensile strain of approximately 12 x 10-5 in./in. before cracking. Assuming an E value of 6.32 x 106 Ib./in 2 the lensile stress is approximately 763 Ib./in2

Discussion

Extruded ceramic blocks can be satisfactori1y pre-stressed. There seems no major technical reason why pre-stressed extruded ceramic blocks should not provide a satisfactory and economical flooring system, though there are still some problems:

I. The ends of the blocks may have to be ground smoolh to obtain a strong resin joint. The cost involved will undoubtedly be the deciding factor.

2. Long ceramic blocks will reduce the number of joinls, thus saving jointing material and labour and also reducing the tendency for failure where joints would otherwise have occurred. However, long extruded ceramic blocks are difficult to manufacture and tend to warp during firing.

3. Long-term behaviour is slill not fuUy understood. ]f a pre-stressed ceramic flooring system is to be developed,

further investigations are necessary and the following sugges-tions may be worthy of consideration.

a. To improve lhe strenglh of the joints, grinding the ends of the blocks may be necessary or alternatively another jointing material investigated.

b. The performance of various geometrical shapes should

be investigaled. c. Long-term tests should be carried out 00 loaded beams.

Cu"ent Post·Tensioned and Prestressed Brickwork and Ceramics in Great Britain 301

d. Since such beams are intended to be laid side by side and intcr·locked to act as a single flaor , behaviour in the lateral direction under load should also be investigated.

As a direct result of the investigation at Sunderland a British patent* has been taken out on a pre-stressed ceramic flooring system , the cross-section af which is illustrated in Figure 34-22.

The Future

As described, great interest has been shown in this form of construction and undoubtedly further developments will take place.

Pre-fabrication offers a challenge to the ceramic industry and one method of permanently assembling small components is to pre-stress.

In the realm of traditional brickwork, temporary externaI post-stressing techniques could be developed to facilitate crectian af large panels which, when in position , cauld be released from pre-compressian , thereafter functioning as tradi­tional panels.

H is not always appreciated by designers that some fired c1ay products can offer compressive strength far in excess of many concretes and with the advent of cheaper high·strength mortars and post or pre-stressed methads, yet another strue­tu ral material will be available.

References

I. " Reinforced Brick Masonry", Bulletin No. 5 , National Brick Manufacturers Research Foundation , February 1932.

*Patcnt Spccification No. Hd - 34864 - "lmprovcmcnts in Of Rclating to folooring and Like Structures"

Figure 34-22. Cross sectioll of Ilew ceramic block (3rd project).

2. Neill , J. A., "Post-tensioned Brickwork and its use in the construction of a factory at Darlington" , c.p T.B. Techllical Note, Vol. I , No. 9. May 1966.

3. Foster , D., "Reinforced Brickwork Box Beams" , S.C.P. 3 , Struetural Clay Products Limited , October 1966.

4. Plowman. J. M., Sutherland , R. J. M., and Couzens, M. L., "The Testing of Reinforced Brickwork and Concrete Slabs Forming Box Beams" , The Structural Engineer Vol. 45, No. lI , November 1967.

5. Henkley , A. T., "Test of one-storey Pre-stressed Briekwork Shear Walls", New Zealalld Engineerillg. 21 (6) June 1966.

6. Tasker , H. E., "Recommendations for the use of Pre-strcssed Brick or B10ek Walls on Reactive Soils", Technieal Record 52: 75:349, Department of Works, Commonwealth Experimental Building Seetion .

7. Prasan , S., Hendry , A. W., and Bradshaw, R. E. , "Crushing Tests of Storey-Height Panels 4-1/2 inches Thick", Proceedillgs of the British Ceramic Society, Issue No. 4, July 1965.

8. Thomas, K., "Laterally Reinforced Briekwork", (Some notes on current practiee in Eastern EUTOpe.) c.P. T.B. Techllical Note, Volume 2 , No. I. July 1967.