Presented at the Sixth Congress of the FederationInternationale de la Precontrainte, Prague, Czechoslovakia,

June 1970

WEDGE ANCHORAGESYSTEM FOR STRANDPOST-TENSIONING

Reports on design and development of a 3-strand wedge anchorand anchorage components. This system has been applied indevelopment of tendons with working forces from 225 to 1300kips (100 to 600 t). Static and dynamic test results arereviewed, and use of the system in U.S. bridge and buildingconstruction is described.

Edward SchechterStressteel CorporationWilkes-Barre, Pennsylvania

Henry C. Boecker, P.E.Stressteel CorporationWilkes-Barre, Pennsylvania

The need for a new high capacitystrand post-tensioning system in theUnited States, was apparent by late1967. The demand for long span,cast-in-place post-tensioned struc-tures in the United States seemedcertain to expand rapidly, and com-parative costs of tendon materialsindicated that strand would econom-ically meet this growing need. Pull-through tendon systems, which re-quire neither fixed tendon lengthsnor factory-attached anchorages,have considerable cost advantagesfor long span cast-in-place bridges,buildings, and nuclear reactor con-tainment structures. Existing Euro-pean pull-through tendon systemswere well represented in the UnitedStates and licenses were unavailable.

In response to these factors the auth-ors began development of the S/HWedge Anchorage System.

SYSTEM CRITERIA

Specific criteria for this system, es-tablished prior to the start of hard-ware development, were:

1. Tendons must develop the min-imum guaranteed ultimate ten-sile strength of the assembledstrand.

2. Anchorages for multi-strandtendons should be so con-structed that, at ultimate loadin the unbonded state, at least2 percent elongation of the ten-don takes place before failureof any strand.

3. Anchorages should be designed

July-August 1971 49

Fig. 1. Three-piece wedge anchor (left) seats in wedge plate (right), each wedgeholding three post-tensioning strands

to meet the ACI Building Coderequirements for unbondedtendons for dynamic loading.

4. Component parts should bedesigned, in size and shape, tobe fabricated on existing ma-chine tools.

5. Field placing and tensioningsystems should be developed

, -CT/OA/ A

7 w/REsrf 'A,vo HOL,

Fig. 2. Symmetrically cut wedge an-chor

/DENT/FY//J&O TC

Fig. 3. Asymmetrically cut wedge an-chor

for maximum economy.6. Tendons should be primarily

designed for use in bondedstructures. Therefore, simpleand efficient grouting methodswere needed.

7. The anchorage devices shouldbe originally designed to use1/2-in. (13 mm) dia. 270K (19,000kg/em2) ASTM A-416 strand,but should be adaptable to im-provements made in this mate-rial and to larger size strands.

SYSTEM COMPONENTS

The S/H wedge is a 3-pieceslotted cone (Fig. 1) which gripsthree ½-in. (13 mm) dia. 270K (19,-000 kg/cm2) strands in its serratedteeth and seats in a conical hole in awedge plate. The strands pass fromthe tendon duct, through a transitioncone and a splay plate, into thewedge plate where they are anch-ored by the wedge. By incorporatinga number of these wedges in a multi-hole wedge plate, this system pro-vides tendons in sizes ranging from225 to 1300 kips (102 to 5901)*working force (Table 1).

°1t = 1000 kg

K

7 /A//.Fsr .qA/z

50 PCI Journal

Three-piece wedge. On the basis of15 years experience with wedgeanchorages for post-tensioning bars,it was decided that the strandwedges should be no larger than thelargest bar wedges. A primary de-sign objective was to achieve a ratiobetween the area of the anchor plateand the number of strands anchoredwhich would be lower than that ofexisting multi-strand anchorage sys-tems. Preliminary analysis indicated

that the combination which wouldbest achieve the most favorable ra-tio was the use of three strands ineach wedge component. This con-cept was new to post-tensioning sys-tems in the United States.

Several types of prototype anchorswere initially fabricated. Under longterm loading conditions, it was ob-served that 3-piece anchors cut sym-metrically as shown in Fig. 2 have atendency to crack across Section

Table 1. Strand tendon properties

Number oflh-in., 270Kstrands per

j tendon

Guaranteedultimatetensilestrength

kips

Load, kips Rigidsheath

diametern.

Dimensionsin inches

0.8 fs 0.7 f' 0.6 f' A B C9 371.7 297.0 260.0 223.0 25/s 11 7 18

12 495.6 397.0 347.0 297.0 25/s 11 7 1818 743.4 595.0 520.0 446.0 33/4 151/4 9 24

24 991.2 793.0 694.0 595.0 33/4 151/4 9 24

54 2230.0 1784.0 1561.0 1338.0 51/2 20 13 36

OO DOODO O

6foa7 /v.°vT

S/.HO 9-.S 4//12-5

SPLAY O/57/,S7T/oA1,LATE ^^° .•'SLATE

-0v WEDGE

A

lNEOGE TR UM/'ET

ANCHOR qGE ASSEMB/_Y

S/H04 --B

OOOOOOO

(0(5000(3000

S/N5 -S

July-August 1971 91

A-B. Should the crack continuethrough the wedge, Section A tendsto break free and, lacking positivesupport, creates an incipient failurecondition. It became obvious thatpositive center support was manda-tory for the new wedge. This re-quirement led to the asymmetricconfiguration shown in Fig. 3. Shoulda full crack occur across Section C-Din the key wedge piece, center sup-port is maintained against each ofthe three strands with all wedgecomponents, especially the smallcenter piece of the key, remainingin compression with no loss of anch-oring capability.

During machining of the wedge, agroove is cut around the outer body.This groove accommodates a neo-prene O-ring, which maintains align-ment of the wedge sections duringshipment and field placement. Anidentifying notch (Fig. 3) is alsomachined into the wedge, betweenthe right and left side sections, sothat field crews can correctly reas-semble the sections should they be-come separated. Wedge sections aremachined to close tolerance, mak-ing comparable sections inter-changeable. Each wedge is shippedas a complete unit, held together byits 0-ring.

Wedge plate. Observation and studyindicated that the most efficient andeconomical method of grouting wasthrough the center of the wedgeplate. Therefore, no separate groutinput or grout vent pipes are re-quired at the anchorage. Preliminaryplate layouts were made with thewedge holes placed in a circlearound a center grout hole. It wasfound, however, that the smallestsize transition cone could be ob-tained only with a square configura-tion of wedge holes.

Wedge plates are fabricated fromlongitudinal standard mill flat barsor plate. Plate cutting, hole drilling,and hole reaming are done on thesame manufacturing line with thesame equipment as that utilized toproduce anchorage plates for the barsystem. Experimental analysis deter-mined that the wedge plate materialshould have a higher yield pointthan that obtainable with rolledplate. To reduce plate deflection andprovide smooth wedge holes, allwedge plates are heat treated.

Prototype anchor plates tested un-der simulated field conditions indi-cated that wedges could be placed ata minimum spacing of only 1/a in.(6.4 mm) when seated. This closespacing permits easy field place-ment, does not affect the strengthcharacteristics of the tendon, and al-lows a close grouping of strands.

Splay plate. Since strand holes in thewedges are drilled with verticallymoving equipment, the holes in thewedge are straight and parallel to itscenter. Strands, however, must passfrom the tight area of the duct,through the transition cone, to thelarger area required by the holespacing in the wedge plate. Strandswould normally enter the wedgeplate at an angle not perpendicularto the plate and thus create the pos-sibility of nonlinear stress.

To insure that the strands enterthe wedge plate holes parallel toeach other and perpendicular to theplate, a thin steel splay plate isplaced between the bearing plateand wedge plate. It accommodatesthe strands from the transition areaby passing each strand through anindividual hole, which provides aprecise guide for the correct align-ment of all strands when tensioned.

52 PCI Journal

Fig. 4. Hydraulic ram, suspended by chainfall, simultaneously tensionsall 12 strands of a tendon

TENSIONING METHODS

Placing and tensioning proceduresdeveloped for the 9- and 12-strandseries embody no new concepts. Af-ter pulling strands through preform-ed ducts, the splay plate, wedgeplate, and wedges are positioned.The strands are then passed througha center hole hydraulic jack, anch-ored at a jacking plate by thewedges, and tensioned (Fig. 4).Anchor wedges are seated, afterelongation, by a secondary hydraulic

system, similar to that previously de-veloped for seating bar wedges.

After the experience gained intensioning 12-strand tendons, how-ever, field personnel became con-cerned with the problem of handlingthe heavy equipment necessary tostress tendons of the 18- and 24-strand series. Preliminary studies in-dicated that this, equipment, com-pletely assembled with jackingchairs and pulling devices, wouldweigh approximately 2 tons (1.8t).Equipment of this size could

July-August 1971 53

Gfl TES

r ^ --^rPUGL/.^/6d I i HEAO

I . 1-^ - I j^^PULL/NG` I / ---I WEDGE

IL TUBESAG/GN/NG TENS/ON/G Rf^^/

CARTR /OG,E

Fig. 5. Aligning cartridge used with tensioning ram

be efficiently handled only by small,long-reach cranes. Field experiencewith the 12-strand system indicatedthat 3-strand wedges were econom-ical to place and anchor in compari-son to monostrand systems. With the24-strand system, however, the costof crane time necessary to hold thejacking equipment in position whilefield crews placed stressing equip-ment, pulling heads and pullingwedges was so high that stressingwith these large tendons promisedto be uneconomical.

This problem was solved with anew approach in stressing the largertendons. A cartridge was designedin which the strands are pre-alignedbefore they enter the center hole ofthe ram, the hole having sufficientdiameter for the cartridge, pullinghead and pulling wedge to passthrough. The aligning cartridge con-sists of a series of parallel tubes(Fig. 5), each tube accepting a groupof three strands. This assures that

all groups of three strands in thewedge plate are precisely aligned forgripping by the pulling head. Thelightweight aluminum aligning car-tridge is readily placed by hand inproper position on the stressing ex-tension of the tendon, along withthe pulling head and pulling wedges.

The center hole ram is broughtover the cartridge, and gatesmounted on the ram are closedahead of the pulling head. The ramcylinder is then axially advanced,tensioning the tendon. After tension-ing load and elongation are reachedand anchor wedges seated, the ramcylinder is retracted and the gatesopened.

The ram is then slipped off thealigning cartridge and immediatelyplaced over the next cartridge,which has been readied for ram ac-ceptance. With this procedure thecrane no longer must hold the jackin place during the time needed foralignment of strands and the placing

54 PCI Journal

of the pulling head and the pullingwedges, as these operations are pro-ceeding in advance. Using thesemethods, a four man crew hasstressed as many as 38 tendons of 24strands each (912 strands) in a sin-gle working day. Without this spe-cial equipment, the maximum num-ber of 24-strand tendons that couldbe stressed by the same crew wouldprobably not be more than 12.

STATIC TESTS

Developmental tests on 3-strandanchorage. The asymmetrical natureof the 3-strand anchorage compo-nents created questions for the sys-tem developers as they sought toachieve the best possible results.Many hundreds of tests were con-ducted during the developmentalstage using combinations of the fol-lowing variables:

1. Wedge metallurgy2. Wedge angle3. Heat treatment4. Configurations of gripping

teeth5. Spacing of gripping teeth6. Length of wedgeAn important conclusion from this

testing was that only a very smallbut critical range of latitude existsin determining proper wedge angle.At one end of the range the wedgeangle can be reduced to the extentthat brittle shear fracture of thestrand wires occur. At the other end,the angle is increased so that grip-ping compression is decreased andslippage of the strands result. Whenthe final wedge angle was determ-ined, it was found to be the same asthat used for many years in thewedge anchors for high strength al-loy steel bars.

A heat treatment procedure wasresearched to insure that the throat

of the key wedge had sufficientdepth of hardness for the teeth togrip the strands and still remain suf-ficiently ductile to prevent crackingduring tensioning and anchoring.Varying the number of the grippingteeth in the wedge from 18 to 54threads per inch (per 21/2 cm) hadno discernible effect upon the per-formance of the wedge in ultimatestrength or fatigue characteristics.

After extensive testing, neverchanging more than one of the vari-ables from one test to another, thefinal S/H wedge was developedwhich now is guaranteed to providea minimum of 2 percent elongationat the ultimate strength of thestrand.

Static tests of 12 -and 24- strand ten-dons. The testing of 12-strand and24-strand tendons was grouped intofour phases:

1. Failure mode and/or perfor-mance of wedge plates at loadsabove the rated ultimatestrength of the strand.

2. Testing of full scale tendons toultimate load with parallelstrands.

3. Testing of full scale tendons toultimate load in a simulatedduct-transition cone condition.

4. Testing of full scale tendons ina simulated duct-transitioncone condition, using field ten-sioning procedures by tension-ing and anchoring tendon atnormal stressing load, then tak-ing tendon to ultimate load.

Phase 1 for the 12-strand tendonswas accomplished with four 11/s-in.(29 mm ) dia. special grade barsanchored by specially machinedwedges of the same outer configura-tion as the 3-strand wedge. This al-lowed testing of the wedge plate to1.2 times the ultimate load of the

July-August 1971 55

MACH/N

,BEAR/BL OCK

l^(/EOGEPLATE

BACK/.' PLATEfrV/T// 7" CENTE.eHOL E

Fig. 6. Phase 1 test of 24-strand wedge plate

3I0I T TT____ L. 0.40 --- _ r--^ - CELL --- 1---

RAMS TEST BEAM-\

^/.9G.e/NG EAJO ZEA.o EA./Fig. 7. Phase 3 test configuration for 24-strand tendon

r?AM A/o. / RA/V A/o- 2

Fig. 8. Phase 4 test configuration for 24-strand tendon

strand tendon. Dial indicators weremounted on the testing machine tomeasure plate deformation duringloading sequences. It was concludedthat a 2-in. (51 mm) thick plate,properly heat treated, will effective-ly contain 1.2 times ultimate load ofthe 12-strand tendon, when used inconjunction with a 5-in. (127 mm)dia. transition cone hole and havinga center hole to accommodate a 3/4-

in. (19 mm) dia. grout input pipe.

Phase 1 testing for the 24-strandtendon was accomplished by placinga wedge plate between two bearingblocks and using solid steel wedgesto apply 1.2 times the ultimate loadwith a 600-ton (544 t) hydraulic press(Fig. 6). It was concluded that a 21/4-

in. (57 mm) plate, properly heattreated, will effectively contain theabove force for a 24-strand tendonwhen used in conjunction with a T-in. (178 mm) dia. transition cone and

PCI Journal56

having a center hole to accommo-date a 1-in. (25 mm) dia. grout inputpipe.

Full scale testing for both tendonsizes was performed using a 14 x 4ft. (4.2 x 1.2 m) concrete test beamwith an 8-in. (203 mm) center hole.Tensioning was done with three 200-ton (180 t) rams placed parallel in atriangular arrangement, with a 1000-kip (454 t) load cell between thebeam and the rams. Over-all tendonlength, from anchor to anchor, was aminimum of 16 ft. (4.8 m).

Phase 2 tests indicated that wedgeplates tested in Phase 1, when usedwith the already developed 3-strandwedge, will effectively contain either12 or 24 strands at ultimate load, anddevelop at least 2 percent elonga-tion.

Phase 3 testing was conducted inthe same manner as Phase 2, exceptthat a splay plate and strand transi-tion device were added at the non-jacking end of the tendon (Fig. 7).Here, the testing was expanded tosimulate elongation of the strandgroups for 9 in. (23 cm) through thesplay plate at 0.8 times ultimate load.This test phase resulted in the con-clusions that a %-in. (9.5 mm) thickplate is sufficient for 12-strand ten-dons and a /z-in. (13 mm) thick plateis sufficient for 24-strand tendons.

For the final test phase of bothtendon sizes it was felt necessary toapproximate actual field stressingand wedge anchoring procedures asclosely as possible, and then to takethe anchored tendon to ultimateload (Fig. 8). This testing phase alsobrought the conclusion that thesizes and configurations of splayplates, wedge plates, and wedges aresuch that the rated ultimate strengthof the tendons and 2 percent mini-mum elongation can be reached.Static tests of 54-strand tendon. In

developing the 54-strand tendon andanchorage system, by extension ofwork previously done with 12- and24-strand tendons, the followingsteps were taken:

1. Testing for failure mode and/or performance of wedge platesat loads above ultimate load ofthe strand tendon.

2. Testing of full scale 54-strandtendon to ultimate load withparallel strands.

3. Testing of full scale 54-strandtendon to ultimate load in asimulated duct-transition conecondition.

4. Testing of full scale tendon un-der full field conditions of load-ing and configuration.

Before the start of this work, test-ing facilities had to be developedfor these unusually large and heavytendons. It was decided to constructa special test bed and static load fa-cility for the performance of thesetests (Fig. 9). This test bed consistsof two hexagonal reaction blocksmade of prestressed concrete. Themovable block is forced away by six225-ton, 12-in. stroke (204 t, 30 cm)hydraulic rams which develop an ul-timate force of 2700 kips (1220 t).The testing of the 54-strand tendonwith a minimum guaranteed ulti-mate strength of 2230 kips (1012 t)was well within the capacity of thismachine.

The first phase of the 54-strandtendon program tested prototypebearing plates and wedge plates to1.1 times ultimate load, or 2460 kips(1114 t). This test was performed us-ing high strength steel bars withspecial wedge anchors made to fitthe geometry of the wedge anchorplate. Measurements were taken ofboth elastic and plastic deformation.It was determined that 2-in.(51 mm) thick bearing plates and

July-August 1971 57

Fig. 9. Special test bed and static load facility for testing 54-strand tendon

3'/2-in. (89 mm) thick wedge platesmeet the criteria for the system withadequate safety.

In Phase 2, full scale ultimate loadtests were performed on prototypebearing plates, wedge plates andwedges. Minor modifications weremade, where necessary, in these ma-terials in order to improve the per-formance of the total 54-strand sys-tem.

In Phase 3, the entire tendon trainusing strand, duct-transition cone,bearing plate, splay plate, wedgeplate and wedges, which will beused in nuclear containment con-struction, was subjected to ultimateload in eight tests. These tests es-tablished that the 54-tendon systemwill consistently develop 100 percentof the minimum guaranteed ultimatestrength of the 54 strands, with anelongation at failure of not less than2 percent (see Fig. 10). A 5/s-in. (16mm) splay plate thickness was alsodetermined in this test phase.

Phase 4 testing is still underway.

DYNAMIC TESTS

Numerous dynamic tests have been

58

performed during the developmentof the 3-strand wedge. The largesttendon tested dynamically, with anultimate force of 3721 kips (168 t),is described in detail here. The pur-pose of this test was to determinewhether the proposed new anchor-age system would meet the require-ments for dynamic load capabilityas stated in ACI 423.°

Test requirements. The specificationrequires that the test tendon be ableto withstand, without failure, at least500,000 cycles of stress varied be-tween loads equal to 60 and 66 per-cent of the minimum tendon guaran-teed ultimate tensile strength. Thespecification also requires that thetest tendon withstand, without fail-ure, at least 50 cycles of stresschange according to the followingformula:

R=0.6 f2000^-

L+100

'"Tentative Recommendations for Con-crete Members Prestressed with Un-bonded Tendons," ACI-ASCE Committee423, ACI Journal, Feb. 1969, pp. 81-86.

PCI Journal

L is the length in feet of the shortesttendon to be used in the structure.A tendon of nine strands, anchoredwith 3-strand wedges, was chosenfor this test.

Test equipment and materials. Thetests were performed on a MaterialTesting System (MTS) UniversalTesting Machine, Model 901.83 (Fig.

11), 300,000 lb. (136 t) capacity, atthe University of California's Lawr-ence Radiation Laboratory, Berke-ley.

The strand material was ASTMA-416 ½-in. (13 mm) 270K (19,000kg/cm2) special grade. Anchoragecomponents were 7'/a-in. dia. x 3 1/2-in. thick (184 x 90 mm) wedge plates;71/4-in. dia. x 5/s-in. (184 x 16 mm)

AVERAGE LEAD

PER STRANDKIPS

1 B !!± La'1D 43K/PS

' £[AIg. 2.6

';1

I

{,

/JLOAD 12.4 nIPSELON6.2.329;

B8LOAD ¢L5 KIP

^ ELONG.2.3

s/45 ANCHORAGE

TESTS B8 _ 59 -511 9%ELONGAT/ON

, /.7 L J

Fig. 10. Tests on three 54-strand tendon systems show good repro-ducibility

July-August 1971 59

4L40

35

30

25

20

t5

(0

0

Fig. 11. Machine used for dynamictesting

splay plates; and 3-strand wedges.In order to duplicate in this test

the maximum splay angle of thestrand as it emerges from the duct,passes through the transition coneand then into the anchorage, the testassembly was constructed with onewedge plate having wedge holes lo-cated at a 1 1/4-in. (32 mm) radius andthe other wedge plate with holes lo-cated at a 21/4-in. (57 mm) radius(Fig. 12).

Procedure and results. The test unitwas preassembled with 3/4-in. (19mm) bars for spacing and position-ing. These were needed to insurethat each of the nine strands wasanchored by the wedges at the samelength in this short, 21-in. (53 cm)test specimen. The test unit wasmounted in the MTS machine andthe 3/4-in. (19 mm) spacing bars re-

moved. The machine was then ad-justed to apply and record the re-quired loads. The cycle rate wasthen raised to six cycles per secondand the test initiated. The 500,000-cycle test was run first. Upon com-pletion of this test, the MTS ma-chine was readjusted for load and,without removing the test specimenfrom the machine, the 50-cycle testwas run.

The test sample withstood the500,000-cycle and 50-cycle testswithout failure. The anchorage endswere checked before and after test.Wedge seating was normal as antici-pated. There was no distortion ofthe wedge anchor plate, no evidenceof wire slip, and no wire failure.

TYPICAL BRIDGE APPLICATIONS

Developed initially for cast-in-placepost-tensioned bridges in California,the system has spread to many otherstates. The S/H Wedge AnchorageSystem has been successfully placed,stressed and grouted in over 250bridge spans using more than 125,-000 individual 3-strand wedges. Allbridge tendons for the system arepulled through preformed ductsafter the concrete has been placedand cured.

Post-tensioning sequence. The fast,efficient post-tensioning sequence isas follows:

1. Bearing plates and transitioncones are attached to end bulk-heads.

2. Rigid galvanized duct is placedin conjunction with mild rein-forcing steel.

3. After concrete placement andcure, strand tendons are pulledinto ducts, and splay plates,wedge plates and anchors areplaced.

4. Tendons are tensioned to prop-er load and elongation.

60 PCI Journal

L F'4(7Y)

4.

N

x'3//('4 HOLEFG12 ^'q..• 6 G

2/LN /

5. After cutting off strand stress- construction. An excellent example

ing lengths, grout caps are is the 110 ft. (33 mm) square gymna-

placed and tendons are grouted. sium roof of Laney College, Oak-land, California (Fig. 18). Ducts

A number of these operations on a ' were placed in the forms in correct

typical bridge project are shown in position for the two-way grid ofFigs. 13 to 17. post-tensioning forces. After con-

crete had cured, the tendons were

TYPICAL BUILDING APPLICATIONS pulled through, anchorages were at-tached and the tendons tensioned

The S/H tendon system is also and grouted. Some of the typical op-

adaptable to cast-in-place building erations are shown in Figs. 19-23.

' i

II II

11 ' i

GOUPLE^

[NELGE7'/ 3/c')

\. JPLAY ^

3^ ^ BA.es Nor

N

Y I^x S^ ••

ce 9?

Fig. 12. Fatigue test set-up for 9-strand tendonJuly-August 1971 61

cartridge is placed on the tendon priorto attaching the jacking ram.

Fig. 17. Grout is pumped into the ducts at thefar end until it free flows from the vent, whichis then capped to build up to specified pres-

sure.

Fig. 13. In a two-span, multiple box girder bridge, rigid ducts are positioned in thewebs along with shear steel.

Fig. 14. Tendons are pulledthrough the ducts using a pulling

sock.

Fig. 15. Upper tendon is stressedand anchored with split wedges..The lower 24-strand tendon haswedge plate in place but is not

stressed.

62 PCI Journal

Fig. 18. Gymnasium roof clear-spans 110 ft. with atwo-way, post-tensioned concrete waffle roof.

Fig. 19. Semi-rigid ducts areplaced in the two-way ribs anddrape up over column supports

in the perimeter beams.

Fig. 20. After roof concrete wascast and cured, tendons arepulled into the ducts using spe-cial guides at the edge of the roof.

Fig. 21. Spray plates were placedon the tendons to separate the9-strand tendon into three groups

of 3 strands.

Fig. 22. Closup of a fullystressed and anchored 9-strandtendon. Jack in place willstress and anchor adjacenttendon. Excess of strands arecut off with an abrasive saw.

Fig. 23. For grouting, special grout caps and valvesare used. The second duct is being flushed with

water just prior to pumping in grout.

July-August 1971 63