Wedge Anchorage System for Strand Post-Tensioning

Home , Wedge

Presented at the Sixth Congress of the Federation Internationale de la Precontrainte, Prague, Czechoslovakia, June 1970 WEDGE ANCHORAGE SYSTEM FOR STRAND POST-TENSIONING

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

Edward Schechter Henry C. Boecker, P.E. Stressteel Corporation Stressteel Corporation Wilkes-Barre, Pennsylvania Wilkes-Barre, Pennsylvania

The need for a new high capacity In response to these factors the auth- strand post-tensioning system in the ors began development of the S/H United States, was apparent by late Wedge Anchorage System. 1967. The demand for long span, cast-in-place post-tensioned struc- SYSTEM CRITERIA tures in the United States seemed Specific criteria for this system, es- certain to expand rapidly, and com- tablished prior to the start of hard- parative costs of tendon materials ware development, were: indicated that strand would econom- 1. Tendons must develop the min- ically meet this growing need. Pull- imum guaranteed ultimate ten- through tendon systems, which re- sile strength of the assembled quire neither fixed tendon lengths strand. nor factory-attached anchorages, 2. Anchorages for multi-strand have considerable cost advantages tendons should be so con- for long span cast-in-place bridges, structed that, at ultimate load buildings, and nuclear reactor con- in the unbonded state, at least tainment structures. Existing Euro- 2 percent elongation of the ten- pean pull-through tendon systems don takes place before failure were well represented in the United of any strand. States and licenses were unavailable. 3. Anchorages should be designed

July-August 1971 49 Fig. 1. Three-piece wedge anchor (left) seats in wedge plate (right), each wedge holding three post-tensioning strands

to meet the ACI Building Code for maximum economy. requirements for unbonded 6. Tendons should be primarily tendons for dynamic loading. designed for use in bonded 4. Component parts should be structures. Therefore, simple designed, in size and shape, to and efficient grouting methods be fabricated on existing ma- were needed. chine tools. 7. The anchorage devices should 5. Field placing and tensioning be originally designed to use systems should be developed 1/2-in. (13 mm) dia. 270K (19,000 kg/em2) ASTM A-416 strand, but should be adaptable to im- provements made in this material and to larger size strands. -CT/OA/ , A SYSTEM COMPONENTS 7 w/RE The S/H wedge is a 3-piece srf A,vo HOL, slotted cone (Fig. 1) which grips three ½-in. (13 mm) dia. 270K (19,- Fig. 2. Symmetrically cut wedge an- 000 kg/cm2) strands in its serrated chor teeth and seats in a conical hole in a wedge plate. The strands pass from the tendon duct, through a transition cone and a splay plate, into the wedge plate where they are anch- K ored by the wedge. By incorporating a number of these wedges in a multi- 7 /A//.F hole wedge plate, this system pro- sr .qA/z vides tendons in sizes ranging from 225 to 1300 kips (102 to 5901)* /DENT/FY//J working force (Table 1). O TC Fig. 3. Asymmetrically cut wedge anchor °1t = 1000 kg

50 PCI Journal Three-piece wedge. On the basis of that the combination which would 15 years experience with wedge best achieve the most favorable ra- anchorages for post-tensioning bars, tio was the use of three strands in it was decided that the strand each wedge component. This con- wedges should be no larger than the cept was new to post-tensioning sys- largest bar wedges. A primary de- tems in the United States. sign objective was to achieve a ratio Several types of prototype anchors between the area of the anchor plate were initially fabricated. Under long and the number of strands anchored term loading conditions, it was ob- which would be lower than that of served that 3-piece anchors cut sym- existing multi-strand anchorage sys- metrically as shown in Fig. 2 have a tems. Preliminary analysis indicated tendency to crack across Section

Table 1. Strand tendon properties

Guaranteed Number of ultimate Load, kips Rigid Dimensions lh-in., 270K tensile sheath in inches strands per strength diameter j tendon kips 0.8 fs 0.7 f' 0.6 f' n. A B C 9 371.7 297.0 260.0 223.0 25/s 11 7 18 12 495.6 397.0 347.0 297.0 25/s 11 7 18 18 743.4 595.0 520.0 446.0 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

O O DOODO O

6foa7 /v.°vT S/.HO 9-.S 4//12-5 S/H04 -- SPLAY O/57/,S7T/oA1 B ,LATE ^^° .•SLATE

-0v WEDGE OOO OOOO A (0(5000(3 000 lNEOGE TR UM/ET

ANCHOR qGE ASSEMB/_Y S/N5 -S

July-August 1971 91 A-B. Should the crack continue Wedge plates are fabricated from through the wedge, Section A tends longitudinal standard mill flat bars to break free and, lacking positive or plate. Plate cutting, hole drilling, support, creates an incipient failure and hole reaming are done on the condition. It became obvious that same manufacturing line with the positive center support was manda- same equipment as that utilized to tory for the new wedge. This re- produce anchorage plates for the bar quirement led to the asymmetric system. Experimental analysis deter- configuration shown in Fig. 3. Should mined that the wedge plate material a full crack occur across Section C-D should have a higher yield point in the key wedge piece, center sup- than that obtainable with rolled port is maintained against each of plate. To reduce plate deflection and the three strands with all wedge provide smooth wedge holes, all components, especially the small wedge plates are heat treated. center piece of the key, remaining Prototype anchor plates tested un- in compression with no loss of anch- der simulated field conditions indi- oring capability. cated that wedges could be placed at During machining of the wedge, a a minimum spacing of only 1/a in. groove is cut around the outer body. (6.4 mm) when seated. This close This groove accommodates a neo- spacing permits easy field place- prene O-ring, which maintains alignment, does not affect the strength ment of the wedge sections during characteristics of the tendon, and al- shipment and field placement. An lows a close grouping of strands. identifying notch (Fig. 3) is also machined into the wedge, between Splay plate. Since strand holes in the the right and left side sections, so wedges are drilled with vertically that field crews can correctly reas- moving equipment, the holes in the semble the sections should they be- wedge are straight and parallel to its come separated. Wedge sections are center. Strands, however, must pass machined to close tolerance, mak- from the tight area of the duct, ing comparable sections inter- through the transition cone, to the changeable. Each wedge is shipped larger area required by the hole as a complete unit, held together by spacing in the wedge plate. Strands its 0-ring. would normally enter the wedge plate at an angle not perpendicular Wedge plate. Observation and study to the plate and thus create the pos- indicated that the most efficient and sibility of nonlinear stress. economical method of grouting was To insure that the strands enter through the center of the wedge the wedge plate holes parallel to plate. Therefore, no separate grout each other and perpendicular to the input or grout vent pipes are re- plate, a thin steel splay plate is quired at the anchorage. Preliminary placed between the bearing plate plate layouts were made with the and wedge plate. It accommodates wedge holes placed in a circle around a center grout hole. It was the strands from the transition area found, however, that the smallest by passing each strand through an size transition cone could be ob- individual hole, which provides a tained only with a square configura- precise guide for the correct align- tion of wedge holes. ment of all strands when tensioned.

52 PCI Journal Fig. 4. Hydraulic ram, suspended by chainfall, simultaneously tensions all 12 strands of a tendon

TENSIONING METHODS system, similar to that previously developed for seating bar wedges. Placing and tensioning procedures After the experience gained in developed for the 9- and 12-strand tensioning 12-strand tendons, how- series embody no new concepts. Af- ever, field personnel became con- ter pulling strands through preform- cerned with the problem of handling ed ducts, the splay plate, wedge the heavy equipment necessary to plate, and wedges are positioned. stress tendons of the 18- and 24- The strands are then passed through strand series. Preliminary studies in- a center hole hydraulic jack, anch- dicated that this, equipment, com- ored at a jacking plate by the pletely assembled with jacking wedges, and tensioned (Fig. 4). chairs and pulling devices, would Anchor wedges are seated, after weigh approximately 2 tons (1.8t). elongation, by a secondary hydraulic Equipment of this size could

July-August 1971 53 Gfl TES d I r ^ --^rPUGL/.^/6 i HEAO

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

IL TUBES AG/GN/NG TENS/ON/G Rf^^/ CARTR /OG,E

Fig. 5. Aligning cartridge used with tensioning ram

be efficiently handled only by small, all groups of three strands in the long-reach cranes. Field experience wedge plate are precisely aligned for with the 12-strand system indicated gripping by the pulling head. The that 3-strand wedges were econom- lightweight aluminum aligning car- ical to place and anchor in compari- tridge is readily placed by hand in son to monostrand systems. With the proper position on the stressing ex- 24-strand system, however, the cost tension of the tendon, along with of crane time necessary to hold the the pulling head and pulling wedges. jacking equipment in position while The center hole ram is brought field crews placed stressing equip- over the cartridge, and gates ment, pulling heads and pulling mounted on the ram are closed wedges was so high that stressing ahead of the pulling head. The ram with these large tendons promised cylinder is then axially advanced, to be uneconomical. tensioning the tendon. After tension- This problem was solved with a ing load and elongation are reached new approach in stressing the larger and anchor wedges seated, the ram tendons. A cartridge was designed cylinder is retracted and the gates in which the strands are pre-aligned opened. before they enter the center hole of The ram is then slipped off the the ram, the hole having sufficient aligning cartridge and immediately diameter for the cartridge, pulling placed over the next cartridge, head and pulling wedge to pass which has been readied for ram ac- through. The aligning cartridge con- ceptance. With this procedure the sists of a series of parallel tubes crane no longer must hold the jack (Fig. 5), each tube accepting a group in place during the time needed for of three strands. This assures that alignment of strands and the placing

54 PCI Journal of the pulling head and the pulling of the key wedge had sufficient wedges, as these operations are pro- depth of hardness for the teeth to ceeding in advance. Using these grip the strands and still remain suf- methods, a four man crew has ficiently ductile to prevent cracking stressed as many as 38 tendons of 24 during tensioning and anchoring. strands each (912 strands) in a sin- Varying the number of the gripping gle working day. Without this spe- teeth in the wedge from 18 to 54 cial equipment, the maximum num- threads per inch (per 21/2 cm) had ber of 24-strand tendons that could no discernible effect upon the per- be stressed by the same crew would formance of the wedge in ultimate probably not be more than 12. strength or fatigue characteristics. After extensive testing, never changing more than one of the vari- STATIC TESTS ables from one test to another, the Developmental tests on 3-strand final S/H wedge was developed anchorage. The asymmetrical nature which now is guaranteed to provide of the 3-strand anchorage compo- a minimum of 2 percent elongation nents created questions for the sys- at the ultimate strength of the tem developers as they sought to strand. achieve the best possible results. Many hundreds of tests were con- Static tests of 12 -and 24- strand ten- ducted during the developmental dons. The testing of 12-strand and stage using combinations of the fol- 24-strand tendons was grouped into lowing variables: four phases: 1. Wedge metallurgy 1. Failure mode and/or perfor- 2. Wedge angle mance of wedge plates at loads 3. Heat treatment above the rated ultimate 4. Configurations of gripping strength of the strand. teeth 2. Testing of full scale tendons to 5. Spacing of gripping teeth ultimate load with parallel 6. Length of wedge strands. An important conclusion from this 3. Testing of full scale tendons to testing was that only a very small ultimate load in a simulated but critical range of latitude exists duct-transition cone condition. in determining proper wedge angle. 4. Testing of full scale tendons in At one end of the range the wedge a simulated duct-transition angle can be reduced to the extent cone condition, using field ten- that brittle shear fracture of the sioning procedures by tension- strand wires occur. At the other end, ing and anchoring tendon at the angle is increased so that grip- normal stressing load, then tak- ping compression is decreased and ing tendon to ultimate load. slippage of the strands result. When Phase 1 for the 12-strand tendons the final wedge angle was determ- was accomplished with four 11/s-in. ined, it was found to be the same as (29 mm ) dia. special grade bars that used for many years in the anchored by specially machined wedge anchors for high strength al- wedges of the same outer configura- loy steel bars. tion as the 3-strand wedge. This al- A heat treatment procedure was lowed testing of the wedge plate to researched to insure that the throat 1.2 times the ultimate load of the July-August 1971 55 MACH/N l^(/EOGE PLATE

,BEAR/ BACK/. PLATE BL OCK frV/T// 7" CENTE.e HOL E

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

3 T I0I 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 were Phase 1 testing for the 24-strand mounted on the testing machine to tendon was accomplished by placing measure plate deformation during a wedge plate between two bearing loading sequences. It was concluded blocks and using solid steel wedges that a 2-in. (51 mm) thick plate, to apply 1.2 times the ultimate load properly heat treated, will effective- with a 600-ton (544 t) hydraulic press ly contain 1.2 times ultimate load of (Fig. 6). It was concluded that a 21/4- the 12-strand tendon, when used in in. (57 mm) plate, properly heat conjunction with a 5-in. (127 mm) treated, will effectively contain the dia. transition cone hole and having above force for a 24-strand tendon a center hole to accommodate a 3/4- when used in conjunction with a T- in. (19 mm) dia. grout input pipe. in. (178 mm) dia. transition cone and PCI Journal 56 having a center hole to accommo- developing the 54-strand tendon and date a 1-in. (25 mm) dia. grout input anchorage system, by extension of pipe. work previously done with 12- and Full scale testing for both tendon 24-strand tendons, the following sizes was performed using a 14 x 4 steps were taken: ft. (4.2 x 1.2 m) concrete test beam 1. Testing for failure mode and/ with an 8-in. (203 mm) center hole. or performance of wedge plates Tensioning was done with three 200- at loads above ultimate load of ton (180 t) rams placed parallel in a the strand tendon. triangular arrangement, with a 1000- 2. Testing of full scale 54-strand kip (454 t) load cell between the tendon to ultimate load with beam and the rams. Over-all tendon parallel strands. length, from anchor to anchor, was a 3. Testing of full scale 54-strand minimum of 16 ft. (4.8 m). tendon to ultimate load in a Phase 2 tests indicated that wedge simulated duct-transition cone plates tested in Phase 1, when used condition. with the already developed 3-strand 4. Testing of full scale tendon un- wedge, will effectively contain either der full field conditions of load- 12 or 24 strands at ultimate load, and ing and configuration. develop at least 2 percent elonga- Before the start of this work, test- tion. ing facilities had to be developed Phase 3 testing was conducted in for these unusually large and heavy the same manner as Phase 2, except tendons. It was decided to construct that a splay plate and strand transi- a special test bed and static load fa- tion device were added at the non- cility for the performance of these jacking end of the tendon (Fig. 7). tests (Fig. 9). This test bed consists Here, the testing was expanded to of two hexagonal reaction blocks simulate elongation of the strand made of prestressed concrete. The groups for 9 in. (23 cm) through the movable block is forced away by six splay plate at 0.8 times ultimate load. 225-ton, 12-in. stroke (204 t, 30 cm) This test phase resulted in the con- hydraulic rams which develop an ul- clusions that a %-in. (9.5 mm) thick timate force of 2700 kips (1220 t). plate is sufficient for 12-strand ten- The testing of the 54-strand tendon dons and a /z-in. (13 mm) thick plate with a minimum guaranteed ulti- is sufficient for 24-strand tendons. mate strength of 2230 kips (1012 t) For the final test phase of both was well within the capacity of this tendon sizes it was felt necessary to machine. approximate actual field stressing The first phase of the 54-strand and wedge anchoring procedures as tendon program tested prototype closely as possible, and then to take bearing plates and wedge plates to the anchored tendon to ultimate 1.1 times ultimate load, or 2460 kips load (Fig. 8). This testing phase also (1114 t). This test was performed us- brought the conclusion that the ing high strength steel bars with sizes and configurations of splay special wedge anchors made to fit plates, wedge plates, and wedges are the geometry of the wedge anchor such that the rated ultimate strength plate. Measurements were taken of of the tendons and 2 percent mini- both elastic and plastic deformation. mum elongation can be reached. It was determined that 2-in. Static tests of 54-strand tendon. 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 plates performed during the development meet the criteria for the system with of the 3-strand wedge. The largest adequate safety. tendon tested dynamically, with an In Phase 2, full scale ultimate load ultimate force of 3721 kips (168 t), tests were performed on prototype is described in detail here. The pur- bearing plates, wedge plates and pose of this test was to determine wedges. Minor modifications were whether the proposed new anchor- made, where necessary, in these ma- age system would meet the require- terials in order to improve the per- ments for dynamic load capability formance of the total 54-strand sys- as stated in ACI 423.° tem. In Phase 3, the entire tendon train Test requirements. The specification using strand, duct-transition cone, requires that the test tendon be able bearing plate, splay plate, wedge to withstand, without failure, at least plate and wedges, which will be 500,000 cycles of stress varied between loads equal to 60 and 66 per- used in nuclear containment construction, was subjected to ultimate cent of the minimum tendon guaran- load in eight tests. These tests es- teed ultimate tensile strength. The tablished that the 54-tendon system specification also requires that the will consistently develop 100 percent test tendon withstand, without failure, at least 50 cycles of stress of the minimum guaranteed ultimate strength of the 54 strands, with an change according to the following elongation at failure of not less than formula: 2 percent (see Fig. 10). A 5/s-in. (16 2000 R=0.6 f ^- mm) splay plate thickness was also L+100 determined in this test phase. Phase 4 testing is still underway. "Tentative Recommendations for Con- DYNAMIC TESTS crete Members Prestressed with Un- bonded Tendons," ACI-ASCE Committee Numerous dynamic tests have been 423, ACI Journal, Feb. 1969, pp. 81-86. 58 PCI Journal L is the length in feet of the shortest 11), 300,000 lb. (136 t) capacity, at tendon to be used in the structure. the University of California's Lawr- A tendon of nine strands, anchored ence Radiation Laboratory, Berke- with 3-strand wedges, was chosen ley. for this test. The strand material was ASTM A-416 ½-in. (13 mm) 270K (19,000 Test equipment and materials. The kg/cm2) special grade. Anchorage tests were performed on a Material components were 7'/a-in. dia. x 3 1/2- Testing System (MTS) Universal in. thick (184 x 90 mm) wedge plates; Testing Machine, Model 901.83 (Fig. 71/4-in. dia. x 5/s-in. (184 x 16 mm)

AVERAGE LEAD PER STRAND KIPS 4L 40 1 B !! ± La1D 43K/PS £[AIg. 2.6

35 ; I {, 1 LOAD 12.4 nIPS /J ELON6.2.329; 30 B8 LOAD ¢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 moved. The machine was then ad- justed to apply and record the required loads. The cycle rate was then raised to six cycles per second and the test initiated. The 500,000- cycle test was run first. Upon com- pletion of this test, the MTS machine was readjusted for load and, without removing the test specimen from the machine, the 50-cycle test was run. The test sample withstood the 500,000-cycle and 50-cycle tests without failure. The anchorage ends were checked before and after test. Wedge seating was normal as antici- pated. There was no distortion of the wedge anchor plate, no evidence of wire slip, and no wire failure.

TYPICAL BRIDGE APPLICATIONS Developed initially for cast-in-place Fig. 11. Machine used for dynamic post-tensioned bridges in California, testing the system has spread to many other states. The S/H Wedge Anchorage System has been successfully placed, stressed and grouted in over 250 splay plates; and 3-strand wedges. bridge spans using more than 125,- In order to duplicate in this test 000 individual 3-strand wedges. All the maximum splay angle of the bridge tendons for the system are strand as it emerges from the duct, pulled through preformed ducts passes through the transition cone after the concrete has been placed and then into the anchorage, the test and cured. assembly was constructed with one wedge plate having wedge holes lo- Post-tensioning sequence. The fast, cated at a 11/4-in. (32 mm) radius and efficient post-tensioning sequence is the other wedge plate with holes lo- as follows: cated at a 21/4-in. (57 mm) radius 1. Bearing plates and transition (Fig. 12). cones are attached to end bulk- heads. Procedure and results. The test unit 2. Rigid galvanized duct is placed was preassembled with 3/4-in. (19 in conjunction with mild rein- mm) bars for spacing and position- forcing steel. ing. These were needed to insure 3. After concrete placement and that each of the nine strands was cure, strand tendons are pulled anchored by the wedges at the same into ducts, and splay plates, length in this short, 21-in. (53 cm) wedge plates and anchors are test specimen. The test unit was placed. mounted in the MTS machine and 4. Tendons are tensioned to prop- the 3/4-in. (19 mm) spacing bars re- er load and elongation.

60 PCI Journal

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 of Figs. 13 to 17. post-tensioning forces. After concrete had cured, the tendons were pulled through, anchorages were at- TYPICAL BUILDING APPLICATIONS 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.

GOUPLE^

i 11 i

II II [NELGE 7/ 3/c)

\. JPLAY ^

3^ ^ BA.es Nor

L F4 x3//(4 HOLE (7Y) FG12 ^q..• 6 G

N2 /L/ 4. N

Y I^ x S^ ••

N ce 9?

Fig. 12. Fatigue test set-up for 9-strand tendon July-August 1971 61 Fig. 13. In a two-span, multiple box girder bridge, rigid ducts are positioned in the webs along with shear steel.

Fig. 14. Tendons are pulled Fig. 15. Upper tendon is stressed through the ducts using a pulling and anchored with split wedges.. sock. The lower 24-strand tendon has wedge plate in place but is not stressed.

Fig. 17. Grout is pumped into the ducts at the cartridge is placed on the tendon prior far end until it free flows from the vent, which to attaching the jacking ram. is then capped to build up to specified pres- sure. 62 PCI Journal Fig. 18. Gymnasium roof clear-spans 110 ft. with a two-way, post-tensioned concrete waffle roof.

Fig. 19. Semi-rigid ducts are placed in the two-way ribs and drape up over column supports in the perimeter beams.

Fig. 20. After roof concrete was cast and cured, tendons are pulled into the ducts using special guides at the edge of the roof.

Fig. 21. Spray plates were placed on the tendons to separate the 9-strand tendon into three groups of 3 strands.

Fig. 22. Closup of a fully stressed and anchored 9-strand tendon. Jack in place will stress and anchor adjacent tendon. Excess of strands are cut off with an abrasive saw.

Fig. 23. For grouting, special grout caps and valves are used. The second duct is being flushed with water just prior to pumping in grout. July-August 1971 63