DESIGN ULTIMATE LOADTEST OF 1/10-SCALEMICRO-CONCRETE MODEL OFNEW POTOMAC RIVERCROSSING, 1-266

W. G. Corley*ҟJ. E. CarpenterManager Senior Research EngineerStructural Research Section Structural Research Section

H. G Russellҟ N. W. HansonResearch Engineer Principal Research EngineerStructural Research Section Structural Research Section

A. E. CardenasҟT. HelgasonResearch Engineer Associate Research EngineerStructural Research Section Structural Research Section

J. M. Hansonҟ E. HognestadAssistant Manager DirectorStructural Research Section Engineering Research Department

HIGHLIGHTS

Under contract with the designersof the proposed bridge, Howard,Needles, Tammen and Bergendoff,Consulting Engineers of New YorkCity, a lho-scale micro-concrete mod-el of a prestressed concrete cantile-ver box girder bridge was con-structed and tested in the StructuralResearch Laboratory of the PortlandCement Association. An artist's ren-dering of the prototype, the pro-posed I-266 Potomac River crossingat Washington, D.C., is shown inFig. 1. When completed, the bridgewill be one of the largest cantilever

*All authors are with Portland Cement As-sociation, Skokie, Illinois.

prestressed concrete bridges in theworld.

The prototype bridge on whichthe model was based has a 750-ft.(229 m) main span as shown in Fig.2. Each side span will be 440 ft. (134m) and the roadway deck will be 110ft. (34 m) wide. Since the prototypeis symmetrical about the center ofthe main span, only one-half of thebridge was modeled as can be seenin Fig. 3. At 13io-scale, the model was81 ft. 6 in. (24.8 m) long, 11 ft. (3.4 m)wide, about 6 ft. (1.8 m) deep at thepier and 1 ft. 3 in. (0.38 m) deep atmidspan and at the abutment. Con-structed of materials having proper-ties similar to those of the prototype,the model represented the "direct"method of structural modeling as de-

70ҟ PCI Journal

Described are the details of making and testing a 1/10-scale modelof a proposed three-span bridge for Washington, D.C., that will beone of the world's longest prestressed concrete cantilever type bridgeswhen completed. Test results are given that show suitability of de-sign and conformance with specifications.

Fig. 1. Proposed prestressed concrete cantilever bridge for 1-266 Potomac Rivercrossing near Three Sisters Islands

scribed in detail elsewhere (1)

The webs and fascias of the proto-type box girder bridge will be 18 in.(0.46 m) thick and 60 ft. (18.3 m)deep at the pier, more slender thanany previously used in this type ofbridge. In addition, the fascias areheavily curved in cross-section. Bothside spans are curved in plan andtherefore subjected to considerabletorsion. Since the construction of thisbridge will set several precedents, itwas decided that structural modeltests should be used to assist in thedesign. The tests were carried out tostudy performance of the modelbridge under application of deadload and design live load. In addi-tion, behavior of the model under

extreme overload was to be deter-mined.

This report describes constructionof the ciao-scale prestressed concretemodel, the testing procedure, andthe results of both service load anddesign ultimate load tests. It isshown that the model supported thedesign service load without structur-al cracking and safely withstood thesevere overload of 1.5D + 2.5 (L +I)° required by Section 1.6.6—LoadFactors, AASHO Standard Specifica-tions for Highway Bridges(2).

°D = effect of dead loadL = effect of design live loadI = impact load

November-December 1971ҟ 71

Fig. 2. Plan and elevation of prototype bridge

4

MODEL CONSTRUCTION

Assembly of superstructure. Al-though the prototype is designed tobe cast in place in segments, themodel was constructed of precast 3-ft. (0.91 m) segments that were se-quentially grouted in position andpost-tensioned together to form thecomplete bridge. This use of precastsegments was strictly for conveni-ence in the laboratory. To simulatethe field construction, continuity ofreinforcement was maintained acrossall joints. Dimensions of a modelsegment near the pier are shown inFig. 4. Complete details of construc-tion and testing will be given else-where (3 .

The superstructure segment di-rectly over the pier was cast in aspecial form and set on a steel"rocker" supported on a reinforcedconcrete block representing the pier.Next, the adjacent first main-spanand side-span segments were posi-tioned and the joints connectingthem to the pier segment were cast.The second main-span and side-spansegments were then positionedagainst the completed 3-segmentportion of the bridge, the connecting

joints were cast, and the appropriatetendons were tensioned. This proce-dure was repeated until the modelbridge was constructed.

Post-tensioning of the prestressingtendons proceeded with erection ofeach new segment. In general, thenegative moment longitudinal ten-dons required to connect a givensegment to the completed portion ofthe bridge were tensioned one dayafter the joint was cast. Appropriatediagonal web tendons crossing thejoint and a similar number of fasciatendons and deck transverse tendonswere also tensioned soon after eachnew segment was in place. The re-maining tendons were tensioned lat-er, as stresses calculated for the erec-tion plan permitted. Other tendonslocated in the pier diaphragm, theabutment diaphragm, and the soffitpositive moment region near theabutment were also tensioned whencalculated stresses indicated that itwas appropriate(3>.

Using the cantilever method, themodel bridge was constructed sothat the superstructure was alwaysheavier on the abutment side of thepier. Overturning of the partiallycompleted bridge was prevented by

72ҟ PCI Journal

Fig. 3. Test setup for 81 1h-ft. long model

a temporary support initially locatednear the pier in the side span andlater moved to a position about two-thirds of the side span away from thepier. At all times after the first longi-tudinal tendons were stressed, a

3,000-lb. (1360 kg) weight represent-ing the 300,000-lb. (136 t) weight ofconstruction equipment on the pro-totype, was kept near each end ofthe model( 3 ). It is intended that thesame cantilever construction proce-

72'

t^

Fig. 4. Dimensions of precast model segment near pierNovember-December 1971ҟ 73

WEBS AND FASCIAS

CAST TOSEGMENT

PRECAST

Fig. 5. Model bridge segment construction

dure will be used for the prototype.

Construction of segments. Each pre-cast model bridge segment was con-structed in three operations. Road-way and soffit sections were precastseparately and then joined togetherby casting webs and fascias. Fig. 5shows this procedure schematically.

Soffit sections were cast on a con-tinuous platform built each side ofthe pier at the location where themodel bridge was later to be assem-bled and tested. The platform,shown in Fig. 6, was constructed tothe shape of the bottom surface ofthe soffit. It served initially as a basefor casting soffit sections and subse-quently as a working platform dur-ing erection of the bridge. Curvedside forms for fascia stubs, interiorforms for web stubs, and greasedrods to form prestressing ducts in thestubs were attached to the platform.Continuity of geometry was ensuredby casting the soffit sections in se-quence starting each side of the pier.

'14

Roadway sections were cast decksurface down on special adjustableplatforms as shown in Fig. 7. Sideforms for the fascia stubs, interiorforms for the web stubs, and greasedrods to form prestressing ducts in thestubs were attached to the platforms.Longitudinal and transverse pre-stressing ducts were formed bygreased steel rods encased in poly-vinyl chloride (PVC) tubing andplaced in the deck. Continuity of theroadway sections was ensured byaligning longitudinal ducts withmetal templates and, as can be seenin Fig. 7, by casting each new sec-tion against a previously cast sec-tion.

The precast roadway and soffitsections were placed in an assemblyframe that was adjusted to ensureIongitudinal prestressing duct conti-nuity and proper relative geometry.In the assembly frame, web and fas-cia prestressing ducts were formedby steel rods encased in PVC tubing.

PCI''Journal

Web duct continuity between seg-ments was established by use of ad-justable templates. After forms wereattached to the web and fascia stubson the deck and soffit sections, asshown in Fig. 8, the bridge segmentwas completed by casting the re-maining portions of webs and fas-cias.

Erection of segments. The erectionof each 3-ft. (0.91 m) section of thebridge model began with the hoist-ing of the segment onto an adjusta-ble temporary scaffolding and

clamping it to the completed portionof the bridge. The temporary scaf-folding and clamping were then ad-justed until observation by surveyinginstruments ( 4 ) indicated that the seg-ment was in its proper position rela-tive to the completed portion of thebridge.

Once a segment was properlyaligned, duct formers for prestress-ing tendons were placed across thejoint, and other required joint form-ing was done. Prior to placing thejoint concrete, a slow curing epoxy

Fig. 6. Soffit platform used for casting soffit slabsNovember-December 1971ҟ 75

Fig. 7. Precast deck slabs were cast on adjustable platforms

adhesive was placed on those por-tions of the joint surfaces that ap-proximately paralleled the planes ofprincipal stress and would be sub-jected to high shearing stresses dur-ing the test.

After curing, the joint wasstripped, duct formers were re-moved, tendons were threadedthrough the ducts, and anchor as-semblies were attached. To com-plete the erection cycle, tendonswere stressed and the temporarysupports were removed. Generally,each cycle was completed on thethird day after placing the joint con-crete.

Method of prestressing and compen-sation for losses. Prestressing ten-dons in the deck, soffit, webs, fasciasand diaphragms were anchored atthe dead end with a button head andat the stressing end with a frictionanchor. An adjustment device wasplaced between the friction anchorand the anchor plate to provide ameans for precise tensioning of eachtendon. Fig. 9 shows the anchorageassembly used for deck tendons.

Two operations were required tocomplete tensioning of a tendon..

76

First, a stressing jack gripped andpulled the wire while pushingagainst the end of the friction an-chor. When this jacking force wasreleased, the friction anchor grippedthe tendon. Loss in tendon forcecaused by slip within the anchor wasrecovered in the second operationwhen the jack pushed against thebearing plate instead of against theanchor. This lifted the grip from theadjustment device so that it could beextended to hold the anchor in itscorrect position.

All tendons were stressed initiallyto 20 percent above the final desiredvalue. This overstress was chosen tocompensate for losses calculated forsteel relaxation, concrete creep andshrinkage and tendon friction. A rep-resentative deck tendon was esti-mated to retain 81 percent of origi-nal prestress after losses of 8.4 per-cent from relaxation, 6.0 percentfrom creep, 0.4 percent from shrink-age and 4.2 percent from friction.Relaxation loss was calculated usingan equation developed at the Uni-versity of Illinois( ). Creep, shrink-age and friction losses were deter-mined from material tests and fric-

PCI Journal

Fig. 8. Precast segment with web and fascia forms in place

Lion tests carried out in conjunctionwith construction of the model.

Similitude of reinforcement. Grade60 deformed bar reinforcement( 6 ) inthe prototype was represented bygalvanized welded wire fabric hav-ing a yield stress of about 60 ksi(4220 kgf/cm2) and meeting re-quirements of ASTM Designations:A185-68 and A82-66( 7,8 ). Mesh sizesused were 2 x 2 in.-12/12 and 2 x 2in.-14/14 (5.1 x 5.1 cm-2.05/2.05mm and 5.1 x 5.1 cm-1.63/1.63mm). The size and amount of meshwas varied to provide about thesame percentage of reinforcement ateach section in the model as in theprototype.

The 32 mm (1¼ in.) prestressing

bars considered in the design of theprototype were represented by 5 mmor ¼ in. (6.35 mm) prestressing wirein the model. Final desired prestress-

Fig. 9. Anchorage assemblyNovember-December 1971ҟ 77

Table 1. Properties of model concrete at 28 days

Location Test of Average Standard Coefficientof 6 x 12-in. values deviation of

concrete cylinders* psi psi variation, %

Segments f' 5,870 520 8.9Segments E 3,350,000 192,000 5.7Joints f' 7,770 595 7.7

* F = compressive strengthE = modulus of elasticity1000 psi = 70.3 kgf/cm2

ing force at a nominal stress of about0.6 f, where f' is the strength of thewire, was 7.07 kips (3210 kg) for theY4-in, wire and 4.35 kips (1970 kg)for the 5-mm wire. The total pre-stressing force rather than the indi-vidual tendon force was modeled.Consequently, the '/4-in. and 5-mmprestressing wires in the model rep-resented approximately six and fourtendons, respectively, in the proto-type.

Longitudinal ducts for the Y/4-in.(6.35 mm) tendons in the roadwaywere placed in two layers with a t-in. (2.5 cm) transverse spacing. A to-tal of 170 tendons passed throughthe five segments in the vicinity ofthe pier. Approximately 10 percentof these were anchored at each endof the first five segments and a likeamount were anchored at each jointbeyond.

As seen in Fig. 7, ducts for trans-verse 5-mm roadway tendons wereplaced in a single layer between thedeck longitudinal tendons. Whererequired near the abutment in theside span, '/4-in. (6.35 mm) longitudi-nal soffit tendons were placed atmid depth of the soffit slab andspaced 31/2 in.' (8.9 cm) on centers.Web prestress was applied with Y4

in. or 5-mm tendons spaced at 63/4

in. (17.2 cm) or 9_in. (22.9 cm) as re-

quired to reproduce prototypestresses.

Properties of concrete and prestress-ing reinforcement. A micro-concretemix with proportions of 1 part TypeIII cement to about 5.25 parts Elginfine aggregate was used to cast thesegments. A water-cement ratio of0.7 gave a slump of about l in. (2.5cm).

The 28-day compressive strengthand elastic modulus measured bycompression tests of 6 x 12-in. (15"x30 cm) cylinders are summarized inTable 1.

Joints between the segments weremade from a low slump mortar of 1part Type III cement and 3 partsmasonry sand. The 28-day compres-sive strengths, measured by com-pression tests of 6 x 12 -in. (15 x 30cm) cylinders, are listed in Table 1.

Prestressing wire met the require-ments of ASTM Designation: A421-65( 9 ). Strengths obtained from testsof representative samples of wireused in the model are shown in Ta-ble 2.

TESTING PROCEDURE

Application of loads during con-struction. As construction of the

model proceeded, forces were ap-plied to simulate the dead load of

78ҟ PCI Journal

the prototype('). At 'ho-scale, the Table 2. Properties of model prestress-model required application of nine ing reinforcementtimes the self-weight of each seg-ment to reproduce prototype stress-es. A schematic drawing showingloading equipment provided at 3-ft.(0.91 m) intervals along the bridgeis shown in Fig. 10.

After a new segment had beenstressed, force was applied to thelower rods of loading equipment atthe completed joint. This was doneby means of temporary hydraulicrams located below the test floor".

Yield stress at Ultimate1% elongation strengthL f9 (ksi) fe (ksi)

264 292 218 254

1000 psi = 1 ksi = 70.3 kgf/cm2

Anchor nuts on the rods were thentightened against the lower side ofthe floor to hold the required force.Coil springs in the system permitted

MODELBEARING rSTEEL CROSSHEAD

II // 11,-LOADING ROD

TEST FLOOR f

.,Jj I!IHV

HOLES ( ••3'-0" SPACING

VARIABLE .

VARIABLE

3 —ANCHOR NUTRAM . LIVE LOAD

CROSSHEAD° Y ; DEAD LOAD

CROSSHEAD

Fig. 10.`Loading equipment details

November-December 1971 79

small movements of the bridge dur-ing construction while maintainingthe total load within a minimum of93 percent of that intended. Totalapplied load during construction ofthe model was determined both bymeasuring the spring lengths and bymeasuring the pier and support re-actions.

Temporary support was providedunder the side span of the modelduring erection. The first supportwas located 5.5 ft. (1.7 m) from thepier until ten main span segmentsand eleven side span segments hadbeen erected. Load was then trans-ferred to a temporary support lo-cated 29.3 ft. (8.9 m) from the pier.This support was removed when thebridge was seated on the abutment.Forces at the temporary support andat the pier were measured at eachstage of construction.

Instrumentation. Loads, reactions,reinforcement strains, concretestrains, web and fascia deflections,superstructure deflections, and rota-tions were measured during the testusing procedures described else-where (310)•

Eighteen calibrated load cellswere used to measure forces. Tencells under the pier and two at theside span abutment measured reac-tions. Applied loads were monitoredboth with pressure cells in each hy-draulic system and with six loadcells distributed through the deadload and live load systems.

Bonded electrical strain gageswere placed on eleven main decktendons above the pier and on foursoffit tendons at the calculated loca-tion of maximum positive moment.Tendon forces were also sensed witha load cell at each end of one Ionglongitudinal deck tendon and withone load cell each on a soffit tendon,

a web diagonal tendon and a fasciatendon.

Strain gages were placed on theconcrete surfaces at 360 locations.These included gages on nine heav-ily instrumented cross-sections, twodiaphragms, and on the soffit nearthe pier.

Lateral deflection at mid-depth ofeach web and fascia was measuredat a location midway between thepier diaphragm and the first inter-mediate diaphragm in both the mainspan and the side span. To accom-plish this, a vertical framework wasattached near the top and bottom ofeach web and fascia to support alinear variable differential transfor-mer displacement sensor at mid-height(10).

Deflections of the superstructurewere sensed at 26 locations includ-ing the quarter-points and mid-lengths of each span and the freeend of the main span cantilever. Alinear potentiometer connected tothe test floor directly below eachmeasuring point was used to sensethe vertical movements.

Rotation was measured over agage length including the first lon-gitudinal deck tendon cutoffs oneach side of the pier. This measure-ment combines the strain above andbelow the section into a single indi-cation of response to bending mo-ment. Strains were measured over a24-in. (61 cm) gage length using lin-ear variable differential transformersensors.

At each load increment, all itemsof information described above wererecorded. Loads, reactions, strainsand deflections were recorded onprinted and punched tape using ahigh speed VIDAR digital data ac-quisition system. Recording of 400channels of information with thisequipment required about 40 sec-

g0 PCI Journal

I.OD+1.0(L+I)

CalculatedLOADҟ Uncracked Section

I.OD+O.O(L+I)

0.95D+O.O(L+I)ҟ Measured

O.OD+O.O(L+I) 0

ҟ

ҟ0.2ҟ

0.4

CANTILEVER DEFLECTION in.

Fig. 11. Load vs. deflection of cantilever end for service loadtest

onds. Rotation and web lateral de-flection were continuously traced onoscillographic recorders. Selectedload vs. rotation, load vs. vertical de-flection, and load vs. web lateral de-flection outputs were displayed con-tinuously on X-Y recorders with theY-axis representing applied load.

Application of design loads. Hydrau-lic loading equipment below the testfloor was arranged to apply deadload and live load to the model usingtechniques described elsewhere(4).The two independent hydraulic sys-tems used to apply these loads areshown schematically in Fig. 10.

The first step in the test sequencewas to transfer the load from thesprings to the hydraulic system andhold 1.0 dead load of the prototype(1.0 D), that is, a load ten times the

November-December 1971

self-weight of the model. A set of ini-tial readings was then taken.

Design service load tests—Afterinitial readings were taken under1.0 D, 1.0 (L + I) representing bothlane loads and concentrated loadsfor HS20-44 loading( 2 ) was added infive equal increments. All electronicinstruments were read and the mod-el was visually inspected at each in-crement. This represented the designservice load condition of 1.0 D + 1.0(L + I). No cracking due to applica-tion of this load was observed. Thelive load was then reduced to zeroand all gages were again read.

Dead load tests—After completionof the design service load test, thedead load was increased from 1.0 Dto 1.3 D in six equal increments. Allelectronic gages were read and the

81

1.5D +2.5(L+I) -

Measured -11.5D+O.O(L+I) -

CrackObserved

CalculatedUncracked Section

I.OD+0.0(L+I)

O.0 D+O.O(L+ I)

ҟ

ҟ2ҟ3ҟ4ҟ5

CANTILEVER DEFLECTION in.

Fig. 12. Load vs. deflection of cantilever end for ultimate loadtest

LOAD

model was visually inspected at eachincrement. No cracking due to appli-cation of this load was observed.The load was then reduced to 1.0 Dand instruments were read to deter-mine residuals.

Design ultimate load test—Thedead load was first increased from1.0 D to 1.3 D in three equal incre-ments. Then four smaller, equal in-crements of load were added tobring the total up to 1.5 D. Finally,live load was applied in four incre-ments until a total of 1.5 D and 2.5(L + I) was carried by the model.All electronic gages were read andthe model was visually inspected af-ter each increment. Under applica-tion of the design ultimate load,cracks were observed both over thepier and near the abutment. Al-though some inelastic deformationwas evident, the model safely sup-

ported this extreme overload. Theload was then reduced to 1.0 D, andanother set of readings was taken.

TEST RESULTSPerformance during construction.Observed reactions, strains and de-flections during construction of thebridge model were all within antici-pated limits. The bridge model wasobserved to respond elastically aseach new segment was erected anddead load was applied. In addition,no cracks attributed to applied loadwere found.

Performance under design serviceload. Under the design service loadof 1.0 D + 1.0 (L + I), the micro-concrete model was observed to per-form as anticipated. No crackscaused by applied load were found.Strains and deflections measured atcritical locations were observed to

82ҟ PCI Journal

be proportional to the applied load,indicating that the structure re-mained essentially elastic. A com-parison of measured and calculatedload deflection curves for mid-span(end of cantilever) deflections isshown in Fig. 11. Calculated deflec-tions were based on an uncrackedsection and measured material prop-erties.

Behavior at service load was with-in the limits generally assumed indesign. These experimental resultsindicate that the serviceability re-quirements implied by the AASHOSpecifications(2) are met by the de-sign.

Performance under design ultimateload. After the service load tests, thedesign ultimate load of 1.5 D + 2.5(L + I) was applied to the micro-concrete model. Under this extremeoverload, the model was observed tosafely carry the applied loads. Someinelastic strains and deflections wereobserved, and cracks occurred bothin the negative moment region overthe pier and in the positive momentregion near the abutment. All of theinelastic behavior observed underapplication of design ultimate loadwas within ranges anticipated.

Fig. 12 shows calculated and mea-sured load vs. deflection relation-ships for mid-span (end of cantile-ver). Calculated deflections werebased on an uncracked section andmeasured material properties. Ascan be seen, the observed deflectionswere in satisfactory agreement withthose calculated.

After the overload was removedand the condition of 1.0 D had beenrestored, all cracks were observed tohave closed until they were barelyvisible to the naked eye. This behav-ior indicated, and measured strainsconfirmed, that the longitudinal pre-

November-December 1971

stressing tendons remained elasticunder application of the design ulti-mate load.

CONCLUDING REMARKS

Results of the structural tests car-ried out on this 'ho-scale model of thenew Potomac River crossing, I-266,show that under the application ofservice load, representing dead loadof the prototype and live load plusimpact under HS20-44 loading, nostructural cracking occurred and thebridge remained essentially elastic.Consequently, the design meets ser-viceability requirements implied bythe AASHO Specifications. In addi-tion, the results show that the modelsafely carried the extreme overloadof 1.5 D + 2.5 (L+ I) with only mi-nor structural cracking. Consequent-ly, it is concluded that the designalso meets the strength requirementsof Section 1.6.6—Load Factors of theAASHO Specifications.

Following completion of the de-sign ultimate load test, several spe-cial tests will be made to comparecomputed and observed behavior ofthe model. Finally, the model willbe tested to destruction.

ACKNOWLEDGMENTS

This work was done under con-tract with the designers, Howard,Needles, Tammen and Bergendoff,Consulting Engineers of New YorkCity. Mr. Gerard F. Fox is the part-ner-in-charge of design and Mr.Fred H. Sterbenz is the Project En-gineer. Mr. A. Gordon Lorimer isConsulting Architect on the project.

Mr. G. I. Sawyer, Assistant Direc-tor, Department of Highways andTraffic, District of Columbia, is theowner's respresentative.

The model was constructed andtested by the Cement and ConcreteResearch Institute, a Division of the

83

Portland Cement Association. Workwas carried out by personnel of theStructural Research Section with theassistance of personnel from thePaving and Transportation ResearchSection. Felix Barda, Associate Re-search Engineer, and A. P. Christen-sen, Senior Research Engineer, pro-vided valuable assistance in the de-sign and construction of the model.

Dr. Roy E. Rowe, Director of Re-search and Development, Cementand Concrete Association, GreatBritain, served as Models Consul-tant.

This work was carried out underFederal Aid Project No. DC-VA-I-266-2(102)1. The work was aided bystaff of the Federal Highway Admin-istration, Department of Transporta-tion.

REFERENCES

1. Mattock, A. H., "Structural ModelTesting—Theory and Applications,"Journal of the Portland Cement Associ-ation, Research and Development Lab-oratories, Vol. 4, No. 3, September1963, pp. 12-23; PCA DevelopmentBulletin D56.

2. "Standard Specifications for HighwayBridges," Tenth Edition, 1969, TheAmerican Association of State High-way Officials, Washington, D.C.

3. Corley, W. G., Carpenter, J. E., Rus-sell, H. G., Hanson, N. W., Cardenas,A. E., HeIgason, T., Hanson, J. M. andHognestad, E., "Construction andTesting of 1/10-Scale Micro-ConcreteModel of New Potomac River Crossing,I-266," Final Report, Cement and Con-crete Research Institute, a Division ofthe Portland Cement Association,

Skokie, Illinois (to be published).'4. Hognestad, E., Hanson, N. W., Kriz,

L. B. and Kurvits, 0. A., "Facilitiesand Test Methods of PCA StructuralLaboratory," Journal of the PortlandCement Association, Research and De-velopment Laboratories, Vol. 1, No. 1,January 1959, pp. 12-20 and 40-44;Vol. 1, No. 2, May 1959, pp. 30-37;Vol. 1, No. 3, September 1959, pp.35-41; PCA Development BulletinD33.

5. Magura, D. D., Sozen, M. A. and Siess,C. P., "A Study of Stress Relaxation inPrestressing Reinforcement," Journal ofthe Prestressed Concrete Institute, Vol.9, No. 2, April 1964, pp. 13-57.

6. ASTM Designation: A615-68, "Stand-ard Specification for Deformed Billet-Steel Bars for Concrete Reinforce-ment," American Society for Testingand Materials, Philadelphia, Pennsyl-vania.

7. ASTM Designation: A185-68, "Weld-ed Wire Fabric for Concrete Rein-forcement," American Society forTesting and Materials, Philadelphia,Pennsylvania.

8. ASTM Designation: A82-66, "ColdDrawn Steel Wire for Concrete Rein-.forcement," American Society forTesting and Materials, Philadelphia,Pennsylvania.

9. ASTM Designation: A421-65, "Un-coated Stress-Relieved Wire for Pre-stressed Concrete," American Societyfor Testing and Materials, Philadel-phia, Pennsylvania.

10. Hanson, N. W., Hsu, T. T. C., Kurvits,0. A. and Mattock, A. H., "Facilitiesand Test Methods of PCA StructuralLaboratory—Improvements 1960-65,"Journal of the Portland Cement A-soci-ation, Research and Development Lab-oratories, Vol. 3, No. 2, May 1961, pp.21-31; Vol. 7, No. 1, January 1965, pp.2-9; and Vol. 7, No. 2, May 1965, pp.24-38; PCA Development BulletinD91.

Discussion of this paper is invited. Please forward your comments to PCI Headquartersby March 1 to permit publication in the March-April 1972 issue of the PCI JOURNAL.

84ҟ PCI Journal