Report of the FIP Commission on

Prefabrication*

SYNOPSISThis report is a compilation of fourteen separate sections on items of inter-

est. They are: stressing together of precast units—elimination of mortar orcast concrete joints; self-stressing concrete; extrusion and vibro-stamping-principles and limitations; the electrical heating of high-tensile bars appliedto the development of mass production techniques; large units used underwater—principles and methods of jointing only; significance of early tensioncracks in the compression zone (including the effect of transporting); storageof precast elements, with a note of precautions necessary owing to relaxation,creep and deformation generally; columns—advantages of prestressing; aresume of existing knowledge on shear at the interface in composite construc-tion; precast piles over 30 m (98 ft.) long; an advisory note on the sudden re-lease of pre-tensioned reinforcement; precasting of prestressed bridge beamsby casting concrete in forms vibrated from below; poles for power lines; ac-celerated hardening. The report has been edited, and some portions deleted,to make it more suitable for PCI JOURNAL presentation..

INTRODUCTION

The Commission on Prefabrica-tion considered that the best waythat they could serve those con-cerned with precasting was to pre-sent reports and notes on items ofinterest on which information didnot appear to be readily available.

The Commission is continuing itsstudies of items which appear ofsufficient interest and in particularof:

(1) Stressing together of precastunits—elimination of mortar orcast concrete joints;

(2) Columns—advantages of pre-stressing;

(3) The electrical heating ofhigh-tensile bars applied to the

*Presented at the Fifth Congress, Federa-tion Internationale de la Precontrainte(FIP), Paris, France, June 1966.

development of mass produc-tion techniques;

(4) Self-stressing;(5) Accelerated hardening of

concrete—temperature andstresses.

Suggestions as to other items, theholding of symposia and ways inwhich the Commission's useful ser-vice could be increased would bemuch appreciated.

STRESSING TOGETHER OF PRECASTUNITS—ELIMINATION OF MORTAR OR

CAST CONCRETE JOINTS

In an endeavour to avoid the useof concrete or mortar in joints be-tween precast units stressed togeth-er, "dry" joints and "glued" jointshave been employed.

Non-bonded joints

The use of "dry" joints without a

October, 1967 41

filling requires the two faces of thejoint to fit together perfectly. Theedges of the joint faces are cham-fered in order to prevent local spall-ing. Elements are cast against eachmatching unit, so as to provide per-fect fit of adjoining faces.

Another method is applied by aBritish company which manufac-tures large-diameter piles from rela-tively short units (e.g. cylindricalunits 53 cm (21 in.) in diameterand 3 to 5 m (10 to 16 ft.) long) as-sembled together by means of pre-stressing. The units are cast vertical-ly in steel moulds, and haveaccurately formed ends which beardirectly on neoprene sheets in thejoints.

Adhesive -bonded jointsIn this system a layer of adhesive

ensures sound fitting of consecutiveunits. The strength of the joints canbe greater than that of the concretein tension, in compression and inshear, and the glue ensures water-tightness.

Epoxy resins of petrochemical ori-gin are preferred. They may containa mineral powder filler (cement hasbeen used in the USSR) and can bemixed with sand for use in thickjoints.

In general it is considered that,for good adhesion, the surfacesshould be well cleaned and not toodamp. Tests carried out with epoxyresins by the Centre Experimentaldu Batiment et des Travaux Publicshave shown the tensile strength ofan adhesive-bonded joint to be high-er than that of the hardened con-crete in 80% of the cases when thejoint faces were dry, and in 30%when the faces were damp. The ad-hesion of polyester resins is alsoweaker on damp concrete.

The surfaces of the joint must begiven very careful preparation. Tests

show that sandblasting and brushingin conjunction with the use of com-pressed air give good results on site,and that grease must be avoided onthe surfaces to be bonded.

When a "pure" adhesive is em-ployed, the thickness of the joint iskept down to a few tenths of a milli-meter for economy and to reducedeformation of the joint. The jointscan therefore be prestressed beforethe adhesive has set, causing the ad-hesive to ooze out at the edges.

Thicker joints, some millimetreswide, are used when the adhesivecontains a filler. Franz advises thatsuch joints should be 1 to 2 cm (0.4to 0.8 in.) thick because, in his opin-ion, it is more expensive to form pre-cision joint faces which will fit to-gether than to make a thick joint. Inthe USSR, a compound of epoxy res-in and cement was used for joints0.2 to 0.8 mm (8 to 32 mils) thick.

The mating surfaces must fit to-gether with greater precision whenthe joint is thin. To ensure a goodfit, the precast units which are sub-sequently to be joined together areoften cast one against the other.Sometimes the mating surfaces areground flat, but this may reduce ad-hesion.

When the precast units are as-sembled together by prestressing be-fore the adhesive has set, it is neces-sary to prevent relative sliding of theunits, for which purpose only it is ex-pedient to provide projections onone unit fitting into recesses on thenext, so as to key them together. Inthe USSR such recesses in the jointare designed to resist working shearforces.

Recent examples of executionIn southern and eastern parts of

the USA, post-tensioned piles up to160 ft. (49 m) long and 54 in. (137cm) in diameter have been used;

42 PCI Journal

these have been made up of centri-fugally spun 16 ft. lengths, pre-stressed together with epoxy joints.Mention must be made of concretepiles constructed of units assembledwith the aid of an Epikote resin-based adhesive compound. Adhesiveis placed in a collar fitted to the topof the pile installed in position. Next,the pile extension unit is lowered onto the adhesive, which is then steam-heated for 20 min., after which thejoint is cooled with water for 10 min.Pile driving can then be continued.This method was employed in theconstruction of the Cultural Centreat Melbourne.

Bridge piers in Malaysia

These piers, with an internal di-ameter of 70 cm (28 in.) and an ex-ternal diameter of 90 cm (35 in.),consist of 3 m (10 ft.) long spunconcrete units post-tensioned togeth-er. The method of manufacture pro-duced very flat ends, the amount ofplay being not more than about 1mm (0.04 in.). The units could thusquite easily be joined by means ofepoxy resin-based mortar.

Piers for quay in Indonesia

These are generally comparable tothe piers installed in Malaysia butdiffer in the method of manufacture,the units being 5 m (16 ft.) long andproduced by the vacuum process.This affected the jointing operations,since the mould was so designed thatit was not possible to obtain satis-factory flatness of the mating sur-faces. In this case the surfaces wereground to reduce the amount of playto 3 mm (0.12 in.).

Here again the units were con-nected by means of an epoxy-basedmortar.

Choisy le Roi Bridge over the Seine

The superstructure is formed of

four box-girders of constant depthbelow deck level. It was constructedby cantilevering out from the piersby the addition of precast units 2.50m (8.2 ft.) long. On the precastingsite, the units were concreted on abase having exactly the same shapeas the soffit of the bridge superstruc-ture and were cast in contact witheach other. To ensure exact locationwhen subsequently assembled, threelocating keys were provided, one inthe middle of the top slab and theother two in the webs of the box sec-tion. The precast units were thentaken to the bridge construction site.A telescopic guiding carriage wasinstalled in the preceeding box-gird-er and the next unit to be assembledwas placed on this carriage, whichallowed for movement in any direc-tion.

When the prestressing cables hadbeen threaded, the next unit to beassembled was fitted in position andprovisionally erected as a trial as-sembly. If the fit was acceptable, theunit was removed and the jointformed with adhesive. Preliminarytightening of the unit was affectedby means of screw-threaded rodsand finally the prestressing cablessecuring the unit were tensioned.The shear across the joint of about2 to 3 kg/cm2 (28 to 42 psi) wastaken up by the locating keys, whichprevented any relative displacement.

SELF-STRESSING CONCRETEThe first significant results in cre-

ating expanding admixtures wereobtained by the French scientistsLossier and Hendrique at the end ofthe 1930's.

Further developments by Mikhai-lov, Litver and Popov (at the Re-search Institute of Concrete and Re.inforced Concrete, Moscow) led tothe creation of stressing cement. Theexpansion of this cement takes place

October, 1967 43

when a strength up to 50 kg/cm2(700 psi) has been reached and self-stressing of concrete becomes pos-sible (the stress in the reinforcementcan reach 5000 kg/em 2 (70,000 psi)

Stressing cement is manufacturedby grinding together Portland ce-ment, aluminous cement and gyp-sum. The hardening and self-stress-ing of reinforced concrete made withstressing cement takes place by cur-ing in water at normal temperature.The type of stressing cement in-tended for precast factories requiresa preliminary curing in water at20°C (68°F) while the one intendedfor site use does not.

Since the waterproof properties ofthe stressed concrete are so high thatit is resistant to the penetration ofgas and petrol the stressing cementis widely used in the making of pres-sure pipes, when the longitudinal aswell as the coiled reinforcement be-comes prestressed.

The first plant to produce self-stressed pressure pipes has now beenconstructed in the USSR; centrifugaland vibro-extrusion methods will beused there.

A considerable programme of re-search is in hand in the USSR andincludes: a method of slowing uphardening by preliminary partial hy-dration of the cement; a study ofthe distribution of prestress; and thecontrol of the final prestress.

Since 1957, research on expandingcements and self-stressing concretehas been carried out in the USA atthe University of California. TheAmerican expanding cement, devel-oped by Klein and Troxell, was amechanical mixture (without com-bined grinding) of Portland cement,special sulpho-aluminate cement andslag, but the slag is not now used.

Some interesting investigationswith the American expanding ce-

ment, on slabs, shells and piles havebeen carried out by Lin. In 1963 in-vestigations were started in theResearch and Development Labora-tories of the Portland Cement Asso-ciation in Illinois, based on the com-ponents Portland cement, aluminouscement and gypsum, proposed bythe Research Institute of Concreteand Reinforced Concrete, USSR.

In contrast to the Soviet method,the Americans prepared the stress-ing cement by ordinary intermixingof the components instead of bycombined grinding; tests confirmedthe results of the Soviet investiga-tions on the properties of the stress-ing cement and further investiga-tions are in hand.

Part of a highway pavement wasconstructed in self-stressing concretein Hamden, Connecticut in 1963. Inthis case, measurements indicatedthat a steel stress of 5000 kg/cm2(70,000 psi) was reached. Experi-ments on filling of joints withstressing cement were also carriedout in the USA by Palmdel and Lo-dy.

Dzhun Way of the Peking Re-search Institute of Building Mate-rials has carried out some work onpressure pipes, using expanding ce-ment obtained from the USSR.

EXTRUSION AND VIBRO-STAMPING-PRINCIPLES AND LIMITATIONS

These are techniques wherebypressure and vibration are appliedsimultaneously to concrete. Whilstnormal vibration methods are ade-quate for most types of precast con-crete product, there are limitationson the efficient placing of high-strength concrete in intricate sec-tions. It has been found that withvibro-stamping techniques, it is pos-sible to use a mix with zero slumpefficiently, even with very compli-cated shapes. Under vibration, the

44 PCI Journal

stiffest mixes become fluid and theapplied pressure induces flow, thusensuring that the mould is complete-ly filled. The frequency of vibrationand intensity of pressure may bevaried to suit the workability of themix.

The raising of the vibro-stamperis aided by introducing compressedair under its bottom face to speedthe breaking of the bond with theconcrete and to eliminate damageto the surface of the unit.

Vibro-stamping had been widelyused in the USSR in the manufac-ture of staircases, tunnel linings,ribbed floor and roof slabs, channelsand I-beams; it has been used effec-tively in conjunction with preten-sioning and post-tensioning tech-niques. However, owing to certaindifficulties, the finish is often poorand its use is not recommended forthe manufacture of concrete unitswhere a high-quality surface is de-sirable.

Extrusion

The technique of extruding con-crete is similar in many respects tothat of vibro-stamping and involvesthe forcing of a plastic mix underpressure through a pre-formed die,and the use of a zero-slump mix un-der combined pressure and vibra-tion permits extrusion into compli-cated sections, with the concretehaving sufficient residual density toretain accurately its formed shapeafter the passing of a movable form.Since the extrusion process does notnormally suffer from the disadvan-tage of the poor finish experiencedwith vibro-stamping, it can be usedwith benefit in the manufacture ofhighly repetitive units which requirea very good surface finish.

Mono-purpose equipment

This type of equipment requires

separate extrusion units for eachdepth or width of member produced.A novel feature of the equipment isthat extrusion techniques are em-ployed in their true form in that themachine develops enough force inthe moulding chamber to extrudethe concrete sections and force themachine along its casting bed. Ex-trusion is achieved by augers whichcarry out the dual purpose of extru-sion and duct forming. It is claimedthat this method of internal pressureeliminates surface tension and re-sults in smooth top and bottom sur-faces.

THE ELECTRICAL HEATING OFHIGH TENSILE BARS APPLIED TO THEDEVELOPMENT OF MASS PRODUCTION

TECHNIQUES

The USSR has developed electro-thermal pre-tensioning of deformedbars for building units up to 24 m(79 ft.) Iong. The reinforcing barsare heated by an electric current andtransferred to their positions in theunits when the current has beenswitched off. The end supports pre-vent shortening, and so when thebars cool they are automatically pre-tensioned.

The following items are new de-velopments of the method:

(1) Tensioned bent-up reinforce-ment is now used. A straight rein-forcing bar is placed in the uppersupports and heated electrically sothat it sags and can then be se-cured to special anchorages lowerin the forms. During cooling thereinforcement is stressed uniform-ly in its sloping and straight por-tions.(2) In order to stress more close-ly in accordance with the bendingmoment diagram and relieve an-chorage forces at the ends, barscan be held along their centralportions only, so that when they

October, 1967 45

are stressed by cooling, their endportions remain unstressed.(3) Anchorage and jointing ofbars is achieved by pressing steelmuff pieces firmly into the defor-mation on the bars. This is a greatimprovement on the older meth-od of forming anchor caps byheating or welding and does notreduce the strength of the bar.Special release devices have been

developed both for the end and in-termediate anchorages.

Pretensioning by the use of awinding machine has been describedat previous Congresses. Electricalheating of the wire or strand rein-forcement greatly reduces the me-chanical effort required for winding.

SIGNIFICANCE OF EARLY CRACKS INTHE COMPRESSION ZONE (INCLUDING

THE EFFECT OF TRANSPORTING)

During the handling and trans-porting of prestressed precast units,it is economically advantageous toallow part of the unit to cantileverbeyond a lifting point or temporarysupport; this increases the tensilestresses at the top and the compres-sive stresses at the bottom of theprestressed section.

The allowable length of a canti-levering portion will therefore belimited by either of these stresses,but tensile cracks may sometimes betolerated. A large negative momentcapacity can be achieved by keep-ing the tensile stresses due to pre-stress at a low value. This may bedone in several ways, such as by de-flecting strands, by neutralizing thebond at the ends or by using pre-stressed top reinforcement.

The allowing of cracks, however,makes high negative moment capac-ity possible without these means,provided that the allowable com-pressive stress is not exceeded in thebottom fibres of the section.

Cracks in the compressive zonecan have disadvantages such as low-er resistance to corrosion of steel,excessive upward or differentialcamber and loss in positive crack-ing moment capacity.

By using an appropriate design,early cracks in the compressive zonewill be of only small width, governedby the amount, distribution andstress of the reinforcement coveringthe cracks. This reinforcement is nor-mally non-stressed and not more sen-sitive to corrosion than a parallelexample in normal reinforced con-crete. The cracks will close fully orpartially when the design load isapplied and, except for extremelycorrosive environments, the problemof corrosion is not significant.

The loss in stiffness with regardto negative moments may cause ex-cessive upward or differential cam-ber during storage, before the de-sign loading can be applied, andthese effects are then only partlyreversible. Units such as T-shapedslabs, where constant camber is im-portant, and elements which takenegative moments and are designedfor composite construction, shouldtherefore not be allowed to crack.For elements such as long beams,upward camber may not be a suffi-cient reason to prohibit cracking inthe compressive zone.

Russian investigators have noteda drop in positive cracking momentcapacity of prestressed beams hav-ing flexural cracks caused by nega-tive moments, deeper cracks givingrise to larger drops. Accordingly, theRussian specifications require a re-duction of the positive moment ca-pacity by 10%, and of the rigidity by15% when the compressive zone isassumed to crack.

The development of cracks willcause a redistribution of the corn-

46 PCI Journal

pressive stresses at the sections in-volved, a fact that may affect creeplosses. When cracks close on loading,the stress distribution will be some-what different from the original andthis may influence the positive mo-ment cracking capacity. As the Rus-sian investigation refers to smallbeams of a specific type, it is feltthat there is a need for further studyof this problem with particular at-tention to the magnitude of com-pressive stresses, creep and shrink-age.

Although the effect of early ten-sion in the compressive zone on thepositive cracking moment capacityis not well understood, early cracksare recognized in some countries.

In the USSR early cracks are per-mitted except for structures under-water or exposed to corrosive envir-onments or fatigue.

Although not specified in anycode, early cracking has been recog-nized in Sweden for some 20 yearsfor large pretensioned beams withthe exception of those used inbridges. Cracking zones have beenadequately covered with normal re-inforcement and the safety factorwith regard to the positive crackingmoment has been kept at 1.2, theflexural tensile strength being as-sumed to be 0.1 of the nominal cubestrength.

The PCI Code of the USA doesnot state any limit for flexural tensilestresses in the case of auxiliary rein-forcement in the zone subjected totension, and specified values maythus be exceeded when not detri-mental to proper structural behavi-our.

The British Standard Code ofPractice allows the specified tensilestresses to be exceeded during han-dling and construction for short peri-ods not exceeding 48 hr., providedthat the engineer is satisfied that the

increase will not lead to permanentdamage or cracking of the concreteor greater losses of prestress thanhad been foreseen.

According to the German specifi-cation, DIN 4227, certain flexuraltensile stresses should not be ex-ceeded during handling. Design ac-cording to this specification thusgives a small probability of cracksin the compressive zone.

French specifications by ASP re-quire tensile stresses during the han-dling phases to remain lower thanthe effective tensile strength at trans-fer.

STORAGE OF PRECAST ELEMENTS,WITH A NOTE OF PRECAUTIONS

NECESSARY OWING TO RELAXATION,CREEP AND DEFORMATION GENERALLY

If deformation due to shrinkageand creep is to be a minimum, it isimportant that the concrete is wellmatured before precast units areerected. Similarly, prestressingshould be carried out as far as pos-sible in advance of erection. To ob-tain the maximum effect from stor-age the following rules are sug-gested:

(1) During the important timeimmediately after casting, the con-crete must be kept wet to ensureproper hydration. The equivalentof about 4 days at 15° C shouldbe considered a minimum.(2) For a cold winter climate,dry-curing should be made in-doors for a longer time than fora more temperate climate, other-wise very little shrinkage andcreep will be developed beforeerection.(3) The store for dry-curing mustbe covered by a roof and also haveside protection as necessaryagainst the weather. The capacityof concrete for absorbing waterdirectly is great, but the inter-

October, 1967 47

change of moisture between con-crete and air through absorptionor diffusion is slow. Storing in thisway will also protect the unitsfrom warping and curvature dueto unilateral sunshine.Columns, piles and poles are often

stored in several layers. The pointsof support should be chosen so thatthere is a minimum of deflexion dueto bending. When more than twosupports are needed to prevent cur-vature, then accurately placed, un-yielding supports are essential.

Short beams are often stored inseveral layers and in this respectshould be supported as for columns.Their own weight has little influ-ence on stresses during storage andthe beams easily obtain camber dueto creep. If this camber is to be con-stant it is essential that the beamsshould be given the same generalcuring conditions and that theyshould be sheltered from sunshine.If the bottom layer is heavily loadedat the ends, horizontal friction forcesat the supports could influence thedevelopment of camber.

Long beams with small lateralstiffness should be propped or re-strained at the centre of the spanand at both supports to avoid warp-ing and lateral curvature.

Slabs are very susceptible to vari-ations of the value and position ofthe prestressing force, of the con-crete quality and the general curingconditions, and to yielding of sup-ports. If slabs are stored with anoverhang, for which prestressing isnot specially arranged, the creep de-formations appear to be much great-er than would be given by normalcalculations, and it is essential thatthis method of storing should beavoided.

Owing to flaws in the manufac-turing process or in storing, undue

deformations such as excessive cam-ber, lateral curvature or warpingmay sometimes occur. By using theeffect of irreversible creep from suit-able loading of the units, it is possi-ble to make corrections. Correctivemeasures are usually expensive toapply and, even if perfect whencompleted, the result is not lasting.Such measures should therefore onlybe used in isolated cases when thereis a possibility of controlling furthermovement after erection.

COLUMNS—ADVANTAGES OFPRESTRESSING

It is often held that there are ad-vantages to be gained from the pre-stressing of long columns, particular-ly as regards their handling anderection. Research and developmentwork carried out by a number of en-gineers has produced figures whichgive support to this view.

Ozell and Jernigan have tested 47columns, of which 41 were preten-sioned, with a 20 cm (8 in.) squareor 15 cm (6 in.) square section anda length/depth ratio from 10 to 32.The load was, except for one case,applied axially and the proportionof pretensioned reinforcement (0.53to 2.5%) and use of stirrups werevaried. The test results led the in-vestigators to the following conclu-sions:

(1) Stirrups should be usedthroughout.(2) The optimum prestress withregard to the ultimate columnstrength was about 0.25 of the ulti-mate cylinder strength, for length/depth ratios between 20 and 30.(3) For length/depth ratios lessthan 20, axially loaded prestressedcolumns had lower ultimatestrengths than the reinforced col-umns, whereas ultimate strengthswere almost equal for ratios above30.

48 PCI Journal

Zung-Teh Zia has approached theproblem of the prestressed columntheoretically, and found that hisanalysis compares favourably withthe experimental results of Ozell andJernigan. He concludes that pre-stressing penalizes the strength forlength/depth ratios of less than 25and that optimum prestressing forlength/depth ratios between 25 and35 corresponds to a ratio of pre-stressing reinforcement of roughly1%, using strand reinforcement andconcrete with a compressive strengthof 6000 lb./in. 2 (430 kg /em2).

Hall reports a series of tests on165 cm (65 in.) long columns, 5 X

7.5 cm (2 X 3 in.) in section and re-inforced with four 5 mm (0.197 in.)diameter wires, with varying preten-sion. The columns were loaded ec-centrically, the eccentricity varyingfrom 1/20 to twice the depth of thecolumns. Hall concludes that pre-stressing reduces load-carrying ca-pacity for axial loading, but for ec-centricities exceeding 1/10 of thecolumn depth, prestressing gives amarked increase. For very large ec-centricities, flexure and not bucklingdominates the failure. Since eccen-tricities less than 1/10 of the depthof the section are rarely encounteredin practice, the results of the teststend to the conclusion that prestress-ing is generally advantageous forslender columns.

Lorentsen has conducted an inves-tigation on pretensioned and rein-forced columns with a 20 cm (8 in.)square section and length/depth ra-tios varying between 15 and 39. Thecolumns were eccentrically loaded,the eccentricity being 1/300 of thecolumn length. Results indicate thatprestressing is favourable for length/depth ratios exceeding 27.

Lin and Itaya deduce theoretical-ly that stresses and deflexions may

be calculated with reasonable ac-curacy according to the elastictheory as long as the column re-mains untracked. Above the crack-ing load, the plastic behaviour ofconcrete as well as of steel must betaken into account within thecracked region.

More research is undoubtedlynecessary to clarify the influence ofslenderness, the eccentricity of loadand amount of prestress on the ulti-mate load of columns. An importantfactor is the influence of creep onstability but no work on prestressedcolumns has yet been done on this.It is hoped that further investigationswill clarify to what extent prestressis beneficial when creep is taken intoaccount. There is undoubtedly a con-siderable field in which prestressingcan add stability to structures con-taining slender columns, and furtherresearch and development on thismatter should be well worth while.

A RESUME OF EXISTING KNOWLEDGEON SHEAR AT INTERFACE INCOMPOSITE CONSTRUCTION

The methods normally employedto improve bond are as follows:

(1) Roughening of the surface ofthe precast concrete by cross-tamping, water-jetting, hacking orother means, but having first en-sured that there is full consolida-tion up to the surface.(2) The use of some form of cas-tellation as an alternative toroughening.(3) The use of resin-bonded gluesto promote bond has been advo-cated but not widely used.(4) Connexion by bars protrud-ing from the precast work, in sucha manner that they can be incor-porated in the work cast in situ.More than one method may be em-

ployed at the same time.

October, 1967 49

Experimental workExperimental work has shown that

where two methods of bond im-provement are used together, theshear strength developed at the inter-face is less than the sum of the twowhen used separately.

Hanson, who has carried out nu-merous tests, considers that thestructure can no longer be consid-ered monolithic when the displace-ment at the interface exceeds 127microns (5 mils). For castellations6.3 cm (2.5 in.) high and 12.5 cm(4.9 in.) long with smooth unbond-ed surfaces and a concrete of cylin-der strength 350 kg/cm2 (5000 psi)for precast concrete and 210 kg/cm2 (3000 psi) for in situ concrete,he found a limiting shear stress of25 kg/cm2 (350 psi), which is abouthalf the value that would be calcu-lated by considering the criterion tobe a failure in compression on thevertical faces.

The Russian Standard Specifica-tion, SNIP II—B.1-62, gives recom-mendations for the design of castel-lations.

Hanson found that for similar con-crete the shear strength developedby a rough surface was 35 kg/cm2(490 psi) which accords well withC & CA findings.

Further tests on rough surfaces,with adhesion eliminated and separ-ation of the surfaces prevented bystirrups, gave a shear strength of 17kg/cm2 (240 psi) after the directstirrup effect had been deducted.A smooth surface gave 13 kg/cm?(180 psi) at the first slip and 7 kg/cm2 (98 psi) thereafter.

Hanson found that there was anincrease in shear strength of 12.25kg/cm2 (170 psi) for each addi-tional 1% of reinforcement passingacross the interface, which is broad-ly confirmed by C & CA.

Other tests have been made byZelger and Dashner on beams witha smooth horizontal joint and onsmall elements with rough orsmoother interfaces. The results,lower than Hanson's, are:

(1) For small ele- 7-13 kg/cm2ments, smooth (98-182 psi)surface withoutstirrups

(2) For small ele- 13-19 kg/cm2ments smooth (182-266 psi)surface with stir-rups (0.9% of theinterface)

(3) For small ele- 8-19 kg/cm2ments, rough sur- (112-266 psi)face without stir-rups

(4) For small ele- 10-14 kg/cm2ments rough sur- (140-196 psi)face with stirrups(0.9% of the inter-face)

Tests on beams have shown that avertical pressure increases shear re-sistance up to 20 or 35 kg/cm2 (280or 490 psi).

There is obviously a case for im-proving adhesion by the use ofglues. Here mention may be made ofa composite bridge comprising aconcrete slab on steel girders withbond improved by gluing.

Most of the composite prestressedconcrete bridges constructed to datehave very low stresses at the inter-face and the projecting bars andtransverse prestress are easily ableto ensure resistance to sliding atthese faces.

PRECAST PILES OVER 30 M(100 FT.) LONG

For piles up to 30 m (100 ft.),the phenomenon of reversed tensilestresses can normally be overcome

50 PCI Journal

by reducing the drop during earlydriving and increasing the weightof the hammer, and information isavailable within this range. Beyondthe range, little is known and itis obviously impossible to increasethe hammer weight indefinitely.More attention must therefore bepaid, in the light of known groundconditions, to the value of the pre-stress, the height of the drop andthe head packing. Computer pro-grams are available to give thestresses set up under defined drivingand ground conditions but have notbeen widely used. Field researchunder favourable conditions to givecomparisons with computer resultscould perhaps lead to the confidentprediction of performance for longpiles.

In the meantime there seems tobe an immediate need for any de-tails available of long piles alreadydriven, and the notes given in TableI are therefore presented.

ADVISORY NOTE ON THE SUDDENRELEASE OF PRETENSIONED

REINFORCEMENT

It is generally accepted that re-lease of prestress by screws is im-practicable and therefore direct cut-ting of the wires must be resortedto if jacks cannot be used.

A number of firms now releaseby cutting the tendons rather thanby jacking. Sometimes this is doneto a carefully worked-out patternand by heating a 6-9 in. (15-23 cm)length of tendon so that the break isnot sudden. In many cases, however,the cutting is sudden and many thou-sands of building units in servicehave been released in this way, butsome disturbance to the bond musthave occurred and this should havebeen considered in the design.

An article on the influence of con-crete strength on strand transfer

length describes some experimentswhich indicate that sudden releaseof strand increases the transferIengths from between 20% and 30%and for strand over 1.25 cm (0.5 in.)diameter can destroy bond lengthaltogethcr for up to 30 cm (12 in.)from the end, and this could havean appreciable effect on shearstrength.

It should be recognized that thesudden release of any tendon de-creases bond and if this methodof release is to be used in the factoryit must be known to the designerand its effect taken into accountin his calculations. For some de-signs he will have to ban the methodaltogether.

In all cases, the exact procedureto be used as regards the patternand manner of flamecutting mustbe carefully worked out and agreedby the parties concerned. It is mostimportant that the operation shouldbe continuously supervised by a re-sponsible and experienced man.

PRECASTING OF PRESTRESSED BRIDGEBEAMS BY CASTING CONCRETE IN

FORMS VIBRATED FROM BELOW

This is a useful method which isnot widely employed except by theFrench. The process is derived fromtechniques of casting on vibratingtables and its main use has beenfor double-T beams for bridge work.

Steel forms are used, carried onelastic blocks set in concrete sup-ports. Vibration is applied throughpower vibrators (7 kw at 50 c/s)attached to the form bottoms. Thevibrators give rise to considerablecentrifugal forces (up to 8000 kg(8.8 tons) which impart vibrationwaves of large amplitude to theforms.

The workability of the concreteis such that it does not flow easilyunder this vibration, but forms a

October, 1967 51

TABLE I

Length Hammer/pile PrestressDesignation Size (ft.) Hammer ratio (psi) Remarks

Wells-Fargo Building, 18 in. square 100-138 McKiernan 0.4-0.3 750 Piles placed inSan Francisco Terry pre-drilled holes,

S-14 and only driven forlast 4-8 ft. intobedrock. Steelpointed.

Ala Moana Building, 18 in. octagonal 170 Vulcan 0.54-0.18 900 In 3 sections, withHonolulu 0.10 prestress through

joints. Long, harddriving throughalternating layersof sand, coral andclay.

Martinez-Benicia 20 in. square 126 McKiernan 0.3 737 Easy driving atBridge, California, (11 in. diameter Terry 10-20 blows/ft.

core) A-14 through silt to de-composed sand-stone where blowsper ft. increasedto 50-100.

Oil Terminal, 22 in. octagonal 115 6.5 tons 0.5 Some splicing wasSlagen, Norway (12 in. diameter done, with eaten-

core) sion pieces ofnormal pile section.Piles were steelpointed for drivinginto rock. Somepiles raked 1 in 3.5.

Lake Maracaibo 36 in. diameter, up to Steam 800Bridge, Venezuela wall thickness 215 hammer (pre-

5 in. with rams tensioned)lip to40,000 lb.

Port of Baton 36 in. diameter 160 McKiernan 800Rouge, Louisiana and 54 in. Terry

diameter, wall S-16thickness 5 in.

Wharves in 29 in. octagonal 125 10 tons 0.43 700 Some splicing wasLisbon (18 in. diameter done.

core)

Marine Works in 30 in. diameter up to 12 tons 0.5 750 Final sets 2 in. perFiji (21 in. diameter 140 blow, in silty clay

core) which tightenedup considerablyafter completion ofdriving.

Ilikai Building, 161/2 in. 120 McKiernan 1,000 In 2 sections, withHonolulu octagonal Terry no prestress

S-10 through joints.

Marine Terminal, 27.5 in. up to B.S.P. 10 0.55 1,200Milford Haven diameter, wall 145 tons semi-

thickness 3 14 in. automaticat center and steam4 t/2 in. at ends

Bridge over Napa 54 in. diameter up to Vulcan 0.35 844 Water-jetted. Driv-River, Vallejo, (44 in. diameter 123 200C en for last 20 ft.California core) only. Monolithic

units made onlong-line method.

Bridge over Napa 24 in. square up to Vulcan 0.55 760River, Vallejo 127 200CCalifornia

Wharf in San 18 in. octagonal up to Vulcan 0.38 800Francisco 155 140C

vertical

Wharf in San 20 in. square up to Vulcan 0.25 100Francisco 138 140C

raking2 in 5

52 PCI Journal

slope at an angle of about 45°.Placing is begun at one end of theform, the slope being progressivelyfed until the other end is reached.The vibrators are gradually movedforward so that they are generallybeneath the area of deposition ofthe concrete.

An elaborate system of trackwaysand easy-release and re-clampingdevices have been worked out forchanging the position of the vibra-tors.

POLES FOR POWER LINES

For certain mass-produced ele-ments, such as poles for power lines,tests are used as a basis for thedesign. Such tests show that theactual resistance of concrete to ten-sile stresses is very large. Tensilestresses up to 0.15 times the cubestrength should be allowed in thedesign. 80 kg/em2 (1100 psi) hasbeen successfully adopted for box-section poles for power lines, inFrance and North Africa.

Discussion of this report is invited. Please forward your discussion to PCI Headquartersbefore January 1 to permit publication in the April 1968 issue of the PCI JOURNAL.

October, 1967 53