Summary Report
Applications ofHigh Strength Concrete
for Highway Bridges
Harold J. JobseDirector at Engineering ServicesConcrete Technology CorporationTacoma, Washington
Saad E. MoustafaChairman
Department of Civil EngineeringUniversity o1 New Orleans
New Orleans, Louisiana
The U.S. Department of Transporta-tion authorized Concrete Technol-
ogy Corporation to identify through an-alytical development and laboratory in-vestigation applications of high strengthconcrete for highway bridges. The ap-plications were to optimize the use of10,000-psi (70 MPa) concrete for ad-vancement in the development ofbridge members.
NOTE , This paper is based on a report prepared forthe OFfu;es of Research and Development, FederalHighway Administration, United States Depart-ment of Transportation. The full report (NumberFHWA 821096) is available at cost of duiilicationthrough the National Technical Information Ser-vice, Springfield, Virginia 22161.
Over the years, the practical designcompressive strength of portland ce-ment concrete has steadily increasedfrom 3 to 6 ksi (21 to 42 MPa). Now it isfeasible to produce and design concretewith compressive strengths of 10 ksi (70MPa) and higher. Concrete of thisstrength requires careful selection ofhigh quality mixing materials and morerigid quality control. High strength con-crete has been used successfully in thecolumns of a 76-story office building andin a 461-ft (140 m) liquid petroleum gasstorage vessel.
High strength concrete has beenmade and used by prestressed concreteproducers for many years. However, the
44
producers use the high early strength ofthe concrete so that the formwork maybe removed, the concrete members pre-stressed, and the forms reused every daywithout taking advantage of the even-tual 8 to 10 ksi (56 to 70 MPa) compres-sive strength in the structural design.
OBJECTIVEThe objective of the project was to
demonstrate the advantages of highstrength concrete in highway bridges byidentifying analytically those membersor systems most likely to benefit. Cer-tain analytical predictions would hetested and the results applied in theform of design recommendations andproposed code changes. The investiga-tion was limited to plant precast mem-bers since those products showed themost potential for effective use of highstrength concrete. Practical techniquesfor manufacturing and constructingmembers with high strength concretewere to be included in the report to il-lustrate the advantages in a way whichwould lead to immediate use.
MIXING AND PLACING HIGHSTRENGTH CONCRETE
Guidelines for mixing and placinghigh strength concrete are based on thefollowing parameters:''3
Constituents: Cement, aggregates,and admixtures that will produce con-crete of the required strength and insureconsistency of properties over a periodof time must be selected. The maximumsize of coarse aggregate must he small,usually not greater than 0.5 in. (13 mm).Superplasticizers may be used.
Water-cement ratio: The water-cement ratio must he low, usually be-tween 0.30 and 0.35, by weight.
Cement content: The cement contentmust be high, usually between 800 and950 lb per cu yd (475 and 565 kg/m3).
SynopsisThis report summarizes the results
of a study to determine the potentialfor high strength concrete in highwaybridges. Current data and applicationsare reviewed and related to AASHTOspecifications. Conceptual designsare developed to explore the potentialfor such possibilities as increasedspan lengths; wider girder spacings;reduced cross-sectional areas ofcompression members; and thinnerplates for segmental construction,arch bridge members and piers.
The capacity of thin concrete platesin buckling is studied. Comprehensiveinteraction diagrams developedthrough a computer program are usedto determine the controlling factor forsquare hollow piers made with thinplates of high strength concrete. Thediagrams were confirmed by labora-tory load tests under varying eccen-tricities.
Also presented are code proposalsand concepts for the more rapidadaptation of high strength concretefor use in highway bridges.
Consolidation: Good consolidation insturdy forms is necessary to produceadequately dense concrete.
Quality control: A well-managedquality control program, includingplanning and cooperation of all partiesinvolved, is essential to success.
PHYSICAL PROPERTIES OFHIGH STRENGTH CONCRETE
The properties of high strength con-crete, similar to the properties of normalstrength concrete, can be predicted byanyone familiar with concrete engi-neering. Ratios of tensile strength to
PCI JOURNALJMay-June 1984 45
15
14
13
12
11
10
STRESSKSI 9
a
7
5
4
3
2
1
O = NILSON i SLATE 25
0 = WANG, SHAH & NAAMAN 3B
_ = KAAR, HANSON & CAPELLIs
0
n
0 r i
0 0,001 0.002 0.003 0,004 0.005 0.006
STRAIN
Fig. 1. Stress-strain curves for high strength concrete.
square root compressive strength forhigh strength concrete are similar tothose obtained for normal strength con-crete.4 Stress-strain relationships forhigh strength concrete are shown in Fig.1 4'B The initial slope of the curves be-comes steeper with increasing compres-sive strength (modulus of elasticity in-creases), and the strain at maximum
stress tends to increase slightly. Also,the slope of the descending portion ofthe curves is steeper for the highstrength concretes than for normalstrength concretes, so the high strengthconcrete in effect is more brittle.
Data indicate that the modulus ofelasticity for 10-ksi (70 MPa) concrete isrepresented better by the formula 4.6,7
46
E e = 26 w x•S f (U.S. units)
or
E, = 0.0337 w' 5 f, (S.I. units)
than by the equation:
E, = 33 w' ,J fe (U.S. units)
or
E, = 0.0428 w'•5 f (S.I. units)
presently suggested by the AASHTOSpecifications4 and the ACT Code9 fornormal strength concrete. High strengthconcretes creep less than normalstrength concretes when loaded to agiven percentage of compressivestrength.4 Shrinkage of high strengthconcrete is about the same as that ofnormal strength concrete.4'7
ADVANTAGES ANDDISADVANTAGES OF HIGH
STRENGTH CONCRETEOne advantage of high strength con-
crete is its greater compressive strength,which can be evaluated in relation tounit cost, unit weight, and unit volume.High strength concrete, with its greatercompressive strength per unit cost, isthe least expensive means of carryingcompressive force. In addition, itsgreater compressive strength per unitweight and unit volume allows lighter,more slender bridge members.
Other advantages of high strengthconcrete include increased modulus ofelasticity and increased tensile strength.Increased stiffness is advantageouswhen deflections or stability govern thebridge design, and increased tensilestrength is advantageous in service loaddesign in prestressed concrete.
A disadvantage of high strength con-crete is that the mix has much less waterthan normal strength concrete. This re-sults in mixes that have reduced work-ability and handling time, making them
more difficult to place and properlycompact. In addition, high quality ag-gregates in necessary sizes and cementthat will consistently produce concreteof the required strength may not beavailable in some localities.
Structural design considerations maypreclude effective use of the increasedconcrete strength. Cross-sectional di-mensions often are governed by factorsother than stress, such as minimumcover, so that the full strength capabilityof the concrete is not used. Further, thetotal prestress force that can be gener-ated may not be sufficient to take ad-vantage of the high strength concrete.
Other disadvantages of high strengthconcrete are its additional cost and theadditional expenses of increased qualitycontrol. Finally, the AASHTO Specifi-cations" tend to discourage the use ofhigh strength concrete because thespecifications are based on the proper-ties of normal strength concrete.
APPLICATIONS OF HIGHSTRENGTH CONCRETE
Solid Section GirdersThe effect of using high strength con-
crete in four different solid section gird-ers was investigated. Cross-sectionaldimensions of the four girders—the bulbtee, the Washington State Departmentof Transportation (WSDOT) Standard,the AASHTO-PCI Standard Sections,and the Colorado Standard—are shownin Fig. 2. Two methods of attaching thedecks were considered: integrally castand cast in place. Decks were cast inplace on the completed girder withoutshoring, so that the entire dead load ofboth girder and deck was carried by thegirder section.
To determine span capabilities of thegirders, AASHTO HS20-44 loading wasused, with a lateral distribution factorS/5.5. Allowable stresses conformed tothe AASHTO Specifications with allow-
PCI JOURNALIMay-June 1984 47
2W
78'
iT_
-1
0
O
0
170 SERIES B0 k 109 SERIES
WSDOT GIRDERS BULB TEE
TYPE 2 d 4S TYPE E4
AASHTO - PCI GIRDERS 72" COLORADO GIRDER
Fig. 2. Solid section girders — Cross sections investigated.
48
GIRDER LENGTH (FT)
0 so 100 150 200
10 KSI
6 I I
10
j iS6
10a
10 4
BULB TEE
AASHTO-PCITYPE Vi
WSDOT120 SERIES
COLORADO
SPACING 8'-0"
Fig. 3. Span capabilities — Basic 72-in. (1830 mm) sections with integral decks.
GIRDER LENGTH (FT)0 50 100 130 200
BULB TEE
AASHTO- PCITYPE Vi
WSDOT120 SERIES
COLORADO
SPACING 6'-0"
Fig. 4. Span capabilities — Basic 72-in. (1830 mm) sections with cast-in-place decks.
able tension in the precompressed ten-sion zone equal to 3 , f (0.249 f ) andallowable compression equal to 0.4f. .
The effectiveness of high strengthconcrete in increasing the span capahil-
ities of a given cross section is shown inFig. 3. Here, for a beam spacing of 8 ft(2.4 m), the span capabilities of all thebasic 72-in. (1830 mm) deep integraldeck sections are shown for the different
PCI JOURNALIMay-June 1984 49
200U-
z
a 150aW
XQ 100
4 6 8 10 12
GIRDER SPACING (FT.)
Fig. 5. Depth variations — Bulb tees with integral deck.
200LL
z
< 150a0)
X'4 1002
BULB TEE'S
60" B.T. Bksi
72" BT Bksi
48- B.T 10 ksi
4 6 8 10 12
GIRDER SPACING (FT.)
Fig. 6. Depth variations — Bulb tees with cast-in-place deck.
concrete strengths. The maximum avail-able prestress force sets span length ca-pabilities at the same value for all con-crete strengths for the AASHTO-PCIand Colorado sections. However, for thebulb tee and WSDOT sections, the ben-efits of using high strength concrete areapparent.
The same information for the basic72-in. (1830 mm) deep sections with
cast-in-place decks is shown in Fig. 4,The benefits of high strength concretein increasing span capabilities are easilyidentified, since the available prestressfactor does not control.
A comparison of Figs. 3 and 4 showsthe advantage of precasting the deck asan integral part of the beam section be-fore prestressing. In addition to provid-ing a high strength concrete in the deck,
50
the prestress force is extended into thecompression flange resulting in greaterspan capabilities.
The potential for using shallowermembers with increasing concretestrength is shown in Figs. 5 and 6. Forthe integral deck, the potential for a re-duction in depth from 72 to 48 in. (1830to 1220 mm) is available for girderspacings up to 6 ft (1.8 m). The influenceof maximum available prestress forcelimits the potential reduction for widergirder spacings. For the cast-in-placedeck, the potential for reducing depth ofsection from 72 to 48 in. (1830 to 1220mm) with an increase in concretestrength from 6 to 10 ksi (42 to 70 MPa)is available for all girder spacings.
Making maximum use of the greaterload carrying capabilities of highstrength concrete girders requires dif-ferent designs for bridges. Two designsof a 150-ft (46 m) simple span bridge areshown in Fig. 7. On the left side, nine6-ksi (42 MPa) concrete girders wereused. On the right side, four 10-ksi (70MPa) concrete girders were used. Theadvantages of the high strength concreteare evident; only four girders areneeded for the high strength concretedesign, while nine girders are neededfor the normal strength concrete design.Despite the thicker cast-in-place deckneeded for the greater transverse spanbetween the four girders, the overalldead load is reduced, and therefore totalprestressing requirements are reduced.
Post-Tensioned Box GirdersMultiple span, cast-in-place, continu-
ous post-tensioned box girder bridges ofconstant depth were represented by thetwo-span continuous structure shown inFig. 8. Concrete strengths were 6, 8, and10 ksi (42, 56, and 70 MPa). Overallbeam depths were 4.5, 5.5, and 6.5 ft(1.4, 1.7, and 2 m). Allowable stresseswere the same as those used in the solidsection girder analysis, except that al-lowable tension in the precompressedtension zone was assumed to be 6 FT
(0.498 ,j7). Loading was three lanes ofAASHTO HS20-44, without lane reduc-tion.
Span capabilities for the differentgirder depths are shown in Fig. 9. Highstrength concrete for continuous boxgirders of 150 to 250 fl (46 to 76 m) spansincreased span capabilities, As with theintegral deck solid section girders, themaximum available prestress force lim-ited capabilities of the high strengthconcrete.
Segmentally Post-Tensioned BoxGirders
Segmentally post-tensioned boxgirder bridges of medium to long spanwere represented by the free cantileverShubenacadie Bridge (South Mainland,Nova Scotia) with a 700-ft (213 m) mainspan and 372-ft (113 m) side spans.10Overall dimensions of the bridge areshown in Fig. 10. The bridge was con-structed with 5-ksi (35 MPa) concreteand used 1.25-in. (31.75 mm) diameterthread bars for post-tensioning. No ten-sion was allowed in the precompressedtension zone. AASHTO HS20-44 load-ing was used.
The bridge was reanalyzed using10,000-psi (70 MPa) concrete to deter-mine how much the thickness of thelower flange could be reduced and whateffect this reduction would have on theoverall moments. As shown in Fig. 11,using high strength concrete reducedthe total flexural prestress force by morethan 10 percent as a result of the re-duced dead load. The optimum lowerflange thickness is 1.6 ft (0.5 m), ob-tained at about 8-ksi (56 MPa) strength.
For segmentally post-tensioned boxgirder bridges, high strength concrete isfeasible in regions, such as the lowerflange, where the design is controlledby stress. In regions such as in the deck,where the design is controlled by otherfactors, normal strength concrete can beused. The webs may be constructed ofeither high strength concrete or normalstrength concrete, depending on mini-
PCI JOURNAL/May-June 1984 51
11r
10 3/4"
'
4 •_ 0" 41 -0n 4-01 ,0 • 41_0", 10 3/4"
5 1 2"Deck.
6 1/2Deck
9- 6 KSI CONCRETE GIRDERS 4- 10 KSI CONCRETE GIRDERS
Fig. 7. Comparison of cross section designs of a 150-ft (45.7 m) simple span bridge made possible by using alternateconcrete strengths.
0C_0Cmz
v
7(D
c0mt SPAN SPAN
2 SPAN CONTINUOUS BRIDGE HU.
42'-R ( 3 TRAFFIC LANES) Z 20C
16 3/4 I Q
BEAM d la
a T ]DEPTH
MIDSPAN >iCENTER
TYP. SUPPORT 2
4'-0" 4 8 8-8•' = 34'-8" 40„ 100
SECTION
c- Span limited by preetreae level
B KSS
K5{t ^ c s b
BEAM DEPTH
Fig. 8. Two-span continuous post-tensioned box girder bridge. Fig. 9. Span capabilities — Two-span continuous box girder bridge.
cn
372-0"
35-b"----
r10•L L tio
2D '-p,. _^#
SECTION a-e
lB
La
SHIJBENACADIE BRIDGE
A372-O' 700-0"
ELEVATION
35'-6"
- N-'— 20'-0"— --#'- ----^-
SECTION A-A
Fig. 10. Shubenacadie free cantilever bridge.
mum thickness requirements, minimumshear reinforcement requirements, andthe contribution of concrete in the websto the shear carrying capacity.
Compression MembersSolid pier shafts and elements of Y-
piers (Fig. 12) are examples of corn pre s-sion members in bridges. Compressionmembers were expected to benefitgreatly from applications of highstrength concrete.
To study the effect of increasing thecompressive strength of concrete from 6to 10 ksi (42 to 70 MPa), interaction dia-grams were developed for the compres-sion strut shown in Fig. 13. The dia-grams were developed for a pin end
condition, assuming that the cross sec-tion possessed the bilinear moment-curvature relationship defined by straincompatibility and the elastic propertiesof the concrete and prestressing strand.The deflected shapes were determinedby integrating the curvatures along thelength of the member. The ultimatestrength design method was used withfactored loads.
Fig. 14 shows that capacities of thecompression strut increase with in-creases in concrete strength.
Results of the compression memberstudy are shown in Table 1. Three slen-derness ratios, three concrete strengths,and three eccentricities are shown. Forshort, concentrically loaded struts, ca-
54
36,000
35,000
PRESTRESSFORCEKIPS
34,000
33,000
32,000
31,000
30,0005 10
CONCRETE STRENGTH K51
Fig. 11. Variation of prestress force and flange thickness with concrete strength —Shubenacadie Bridge-
Table 1. Axial load capacities for 10-in, (254 mm) square strut.
Eccentricity
e= 0 a=0.1D a=0.2D
P P PLHr (ksi) (kips) PIP, (kips) Pmg (kips) P/P,
6 375 1.00 284 1.00 217 1.000 8 518 1.38 389 1.37 294 1.35
10 662 1.77 478 1.69 362 1.67
6 320 1.00 233 1.00 168 1.0050 8 437 1.37 317 1.36 220 1.31
10 545 1.70 390 1.67 267 1.59
6 205 1.00 133 1.00 80 1.00100 8 270 1,32 165 1.24 90 1.13
10 325 1.59 190 1.43 100 1.25
Note: 15 factorsP in accordance with ACI 318-77, Section 9.3.2c.' and AASHTOSpecification, Section 1.5.33.4, are included in this table.
PCI JOURNALJMay-June 1984 55
01rn PRECAST, PRESTRESSED CONCRETE
• • r r- -'0 —f^
Fig.
Fig, 13. Strut details.
0 10 20 30 40 50 60 70ULTIMATE MOMENT 1K-FT)
Fig. 14. Interaction diagram — 10 in. (254 mm) square prestressed strut.
'^^as fC =10 KSI
so
♦
\ .': ^;:^,.
o
"5+
, ♦
\\\
':',
f ^= 6 KSI ^ ' \y ^ 1
y yt ^
t i
y r
r; i
700
600
500
Ca 4000J
J
X
300WI-
I--JD
200
100
PCI JOURNAL(M1ay-June 1984 57
20
19
Q 1s
17
164
15
14
0 13
0 12J
11
J 104
?C 9
8
WH
6
5-JD 4
■■ M
u•iu•■■, ■■ • •••••••
^^•••••u•••
••^^^ ■ ■ ■■•■■■^_^^
■■■■■N.^i_ .__ ■■■■■i.0 ■.. ■■• ■ • ■■ ■a■■
■■ MM ■■^ ■SEE■ ■■ME.. ■R,.:
■ ■■^1■W ■GGi■5 10 15 20 25 30 35 40 45 50
ULTIMATE MOMENT (K-FT x 103)
Fig. 15. Interaction diagram —15 ft (4.57 m) diameter circular hollow pier.
30
pacity is determined by multiplying thecylinder strength by the cross-sectionalarea and then subtracting the prestressforce. Because the prestress force is con-stant, the benefits of increasing concretestength are more than the strength in-crease itself. This is illustrated by thestrength ratios in the table that exceed1016 = 1.67 for those struts.
For combined axial load and bending,using high strength concrete even forrelatively slender columns is benefi-cial. It can he concluded that compres-sion members are an excellent applica-tion for high strength concrete. Smallersections can be used for a given numberof members, or fewer members can heused in a given location. In either case,weight as well as material and construc-tion cost are reduced.
Thin Walled Sections
A circular and a square pier with 6-in.(152 mm) wall thicknesses were chosenfor investigation. A 15-ft (4.6 m) outsidediameter hollow circular pier had a pre-stress steel area of 20.2 sq in. (130 ema).A 10-ft (3 m) square hollow pier had aprestress steel area of 18.4 sq in. (118cm2) concentrated in the corners. Con-crete strengths studied were 6 and 10 ksi(42 and 70 M Pa).
The piers were considered to havepinned end connections which werefree to rotate but not free to translate.The 1977 AASHTO Specifications, Sec-tion 1.5-31, 11 were used to design thepiers. Interaction diagrams were con-structed using the assumption that theconcrete stress equal to 0.85f was dis-
58
IIIIIliL HISQUARE
f^ =
HOLLOW
6KSI, 1OKS1
PIERS
oo'
O S °10 O"/rs /V o0
Sç-7--^ ^ 1
1
fL
`
35 40
20
19
1s
17
Q 16
x 15N
14
13
12
0J 11
J 10
,[ 9
swI-a 7
r 6J
5
4
3
20 5 10 15 20 25 30
ULTIMATE MOMENT (K-FT x 103)
Fig. 16. Interaction diagrams --10 ft (3.05 m) square hollow pier.
tributed as a rectangular stress block. Abilinear moment-curvature relationshipwas used to approximate the actualparabolic moment-curvature of the sec-tion. The final deflected shape of thecolumn from which the information forthe interaction diagram was obtainedwas constructed using the same princi-ples as outlined for compression struts.Concrete stresses were obtained fromthe parabolic stress-strain curve for the
concrete, assuming a maximum strain of0.003. From the column-deflectioncurves, a combination of eccentricity ofaxial load and height of column is foundfor a particular axial load. A range ofaxial loads were considered from zero tothe case of pure compression.
The interaction diagrams in Fig. 15and in Fig. 16 show the increased ca-pacity of the piers when concretestrength is increased from 6 to 10 ksi (42
PCI JOURNALJMay-June 1984 59
PrecastPlate
PrecastDiaphrag,
Cast in PlaceCornerPost - Tens
Duct
Fig. 17. Square hollow pier — Thin plate precast segments.
to 70 MPa). The benefit of increasingconcrete strength is more than directlyproportional to the strength increase.
High strength concrete permits theadaptation of thin wall members to pierconstruction for major bridge spans. Thestrength of these sections permits theiruse when tall piers are required. The in-creased load carrying capability permitslonger spans or fewer piers. The lightersection permits construction with mini-mal disruption to surrounding terrain.
Two construction procedures areshown in Figs. 17 and 18. In Fig. 17,
thin plates are joined at the corners withcast-in-place concrete. By building thepiers out of sections of plates and in-stalling the precast diaphragms shown at5 to 10 ft (1.5 to 3 m) levels, these platescould be erected with the utilization of agin pole and minimum winch capacity.
In Fig. 18, the square prismatic boxsections are match cast, transported tothe site, and post-tensioned to form thepier using standard segmental construc-tion methods. The size of the sectionscould he controlled by site require-ments and crane capacities.
60
Post – Tensioninn
Preca$Box
/
Fig. 18. Square hollow pier — Match cast precast segments.
Evaluation of Benefits
The results of the study of applica-tions of high strength concrete to bridgemembers are evaluated qualitatively interms of a number of criteria listed inTable 2. Struts and hollow piers offerthe best possibility for materials savingspercentage-wise over the other mem-bers since either the sizes of the indi-vidual members or the total number ofmembers can be reduced. The potentialfor reduced shipping weight for hollowpier prismatic members or plates offers a
relatively new construction procedurefor this bridge member. These precastsections can be adapted to changing re-quirements by form modification if thechange is not major. These factors led tothe selection ofatest program for squarehollow piers.
DESIGN OF TEST SPECIMENObjective
A square hollow pier specimen wasselected to determine that design could
PCI JOURNAL/May-June 1984 61
Table 2. Evaluation of benefits of using high strength concrete.
Reduced Reduced Difficulty AdaptableMaterials shipping structure of standard
Member or system savings weight depth production section
Solid section girder,integral deck Medium Medium High Medium Medium
Solid section girder,cast-in-place deck Medium Medium High Medium Medium
Commonly usedbox girders Medium — High High Low
Segmentally post-tensioned boxgirders
Plant precast Medium High Medium Medium --Cast-in-place Medium — Medium High
Struts, solid square Medium High — Low High
Piers, hollow thinwalled Medium High — Medium High
P max
ter—(Based on Strength
M max pmax ^^♦^
{Controlled by Compression'Flange Buckling) - F
Fig. 19. Interaction diagram — Thin plate buckling.Mmax
be controlled by overall buckling of thepier rather than by buckling of the thinplate wall of the pier. Fabrication pro-cedures for thin plate segments withcast-in-place corners would be com-pared with plant fabricated match castprismatic boxes. Post-tensioned con-
struction techniques would be con-firmed.
Analytical DevelopmentAlthough studies of thin plate buck-
ling of steel plates have been made, theapplication of these studies to concrete
62
C
thin plates was questionable. Applyingformulas developed by Priest' for highstrength steels and using these formulaswith k factors suggested by Bleich12 forbuckling strength of metal structures re-sults in:
xirz Et t(1)
a"-3(1–p.2) b
in which
vcr = critical unit compressive stress,psi
Et = tangent modulus of elasticity, psiN. = Poisson's ratiob = width of plate, in.t = thickness of plate, in.
As can he seen from Eq. (1), the criti-cal stress is directly proportional to themodulus of elasticity and to the squareof the thickness-to-width ratio of theplate.
In applying this equation to hollowrectangular piers, the following com-ments can be made:
1. Under concentric axial load wherethe entire section is subject to uniformcompression, all four sides of the crosssection are equally susceptible to buck-ling,
2. When the section is subject to ec-centric load, the compression flange willbe the most likely to buckle first.
3. Considering that only eccentricloads are encountered in practical appli-cations, local buckling of the thin com-pression flange should be considered inthe capacity investigation of the section.
4. An interaction diagram that willtake compression flange stability as wellas overall stability into consideration inaddition to material failure appears torepresent the only logical solution to theproblem.
Fig. 19 shows a schematic of such aninteraction diagram. Note that the ACICode$ concrete stress block cannot beused to develop this diagram. Instead,representative stress-strain diagrams, asshown by Fig. 20, must be used in thecomputation. In determining critical
EEmax ^u
Fig. 20. Typical concrete stress-strain curve.
stress, average rather than maximumstress in the compression flange shouldbe used.
The comprehensive interaction dia-gram shown in Fig. 19 will handle anybuckling situation for square or rectan-gular piers. A computer program writtenin Fortran IV language was prepared forthis project enabling the analysis ofhollow rectangular concrete sectionscontaining prestressed and nonpre-stressed reinforcement.
Interaction diagrams developed fromthis program for 10-ft (3 m) square hol-low piers with 3, 4, 5, and 6 in. (76, 102,127, and 152 mm) wall thicknesses areshown in Fig. 21. In order to show theeffect of local buckling of the compres-sion flange on the load resistance of thecross section, curves representing loadcapacities based on both material failureand local buckling are portrayed. An ex-amination of the figure indicates that thegap between capacity based on materialfailure and that controlled by localbuckling increased with the decrease ofthe thickness-to-width ratio of the com-pression flange.
SQUARE HOLLOW PIERTEST SPECIMEN
A 5-ft (1.5 m) square cross sectionmember with a 1½-in. (40 mm) thick
PCI JOURNAUMay-June 1984 63
30.00010'-0" SQUARE HOLLOW PIES
f , = 10,000 psi
M – Pn ,5
6" Wall Thickness (Material Failure)Prestressing = 1,060 psi Max Pmax max
(Compression\ ^^ Flange Buckling)
L '
` 1,045psi Y 1
1,030psI ^\
^1 '011}psi "' \ 1
S
__
25.000
}F- 20,000U
U15,000
ILU
0LL
10,000Jax
5000
00 10,000 20,000 30,000 40,000 50,000 60,000
MOMENT CAPACITY (K- FT.)
Fig. 21. Interaction diagram for a 1 0-ft (3.05 m) square hollow pier.
wall was selected as representative of ahalf-scale model of a 10-ft (3 m) squarepier with 3-in. (76 mm) wall thicknessesas shown in Fig. 22. A 15-ft (4 m) lengthwas selected to provide short columnaction removing the parameter of overallpier stability.
The interaction diagram for the spec-imen is shown in Fig. 23. The predictedtest results based on the specimen per-forming in accordance with the theorydeveloped are tabulated in Table 3.
Fabrication of Test SpecimensMatch cast precast segments 5 ft (1.5
m) square with 5 ft (1.5 m) long and
1'/2-in. (40 mm) thick walls were se-lected for testing. As shown in Fig. 24,three segments would be match cast andconnected together by post-tensioning,making a 15-ft (4.6 m) long pier to betested for local buckling. The match castforming is shown in Fig. 25. Specimenswere cast in the architectural plant atConcrete Technology Corporation,epoxied and tensioned together, andmoved into the laboratory as shown inFig. 26.
The fabrication procedures confirmedthe difficulties that could be experi-enced in match casting square boxeswith thin walls. It was found to be flit
64
Reinforcing
i.3"Post - TensionStrand
PLAN
U.
00N
101-01. -.J
Fl FVATION
Fig. 22. Square hollow pier.
Table 3. Predicted test results.
e(in.)
P(kips)
M(kip-ft)
Curvature(rad/in.)
Deflection(in.)
Endrotation
(degrees)
6 2300 1150 1.041 x 10-6 0.04 0° 3' 14'
12 1890 1890 1.46 x 10-5 0.06 0° 432'
24 1390 2780 2.01 s 10-5 0.08 0° 6' 14"
36 990 2970 3.43 x 10-' 0.14 0° 1(Y3&
ficult to maintain the alignment of thewalls. Furthermore, the thin sectionscreated a problem in placing highstrength concrete with the necessaryslump requirements into the thin sec-tion. A preferable construction would beto cast thin plates on acontinuous bed asis customary in precasting plants, truckthem to the site on flatbed trucks, erectthem in position and cast corners inplace with longitudinal post-tensioning.
Testing ProceduresThe test arrangement is shown in Fig.
27. Loading rams were located to pro-vide the necessary eccentricities for thespecimens as required in Table 3.Loading was applied in two phases. Thefirst load phase was planned to study theperformance of the specimen tinder ser-vice load conditions by cycling the loadin small increments from zero to serviceload level. The second load phase wasplanned to verify the ultimate load ca-pacity of the specimen as well as themode of failure. The load was applied insmall increments from service load tofailure load.
Instrumentation included load cellsfor measuring forces, potentiometers formeasuring displacements and curva-tures, and strain gages for measuringstrain in concrete. Experimental infor-mation was recorded using a high speedautomatic data acquisition system. Theload cell arrangement for Specimen f isshown in Fig. 28.
PCI JOURNAUMay-June 1984 65
4000
3000
U
aaU
2000WU
0U-
1000Jaxa
n n-.. (Material Fallure)
S
S.S.
M max - P max 0-{Comp. Flange Buckling)
Oft ,1
0 1000 2000 3000 4000MOMENT CAPACITY (K-FT)
Fig. 23. Interaction diagram —5 ft (1.52 m) square hollow pier specimen.
60lr
Fig. 24.
66
Fig. 25. Forming of match-cast segmentwhich was post-tensioned to adjoiningsegments before being moved to thetesting lab.
Analysis of Test DataLoading was applied to Specimen 1 to
place the total reaction of all loads at a6-in. (152 mm) eccentricity. The speci-men failed explosively in the centersegment at a load of 1,613,500 lbs (7172kN) at an eccentricity of 4.9 in. (124mm). Fig. 29 shows the buckling of thetop plate. The failure extended totallythrough the specimen at approximatelythe same location. The concrete strengthin the failed segment at the time oftesting was 8680 psi (60 MPa). An inter-action diagram based on this strength isshown in Fig. 30. The failure point indi-cated on the diagram shows that the ac-tual failure load is reasonably close tothe predicted envelope.
Specimen 2A was loaded at an eccen-tricity of 20 in. (508 mm). It failed explo-sively like Specimen I at a load of1,346,000 lbs (5983 kN) at an eccentric-ity of 20.2 in. (513 mm). The failed spec-imen is shown in Fig. 31. The cornpres-sive strength for Segment 6A at time offailure was 9880 psi (68 MPa). The in-
Fig. 26. Specimen being moved from precasting plant to testing lab.
PCI JOURNAUMay-June 1984 67
s'- 101/2"
Loading _Rams rr ^^—f
I I I'Segment
I I I^
End BlockEND VIEW
End Platform Consists of-Concrete End Block
to
N
P.T.Bars
C.G. OfBars
SECTION__
9 14'-B"Jacking ABlock II n rAam
Load— JackingCell Blocks
A 3- 5' Long Specimen
ELEVATION VIEW
2 3,16"WWF
1 1/2"
4 1/2"
1 3/4"m Duct For1 1/4' Dywldag Bars
TYPICAL CORNER
Fig. 27, Square hollow pier test arrangement.
Fig. 28. Load cell arrangement for Specimen 1.
Rams
Segment3
Segment ,^yN2
Segment
Load Cells
PLAN
Fig. 29. Specimen 1 — Top plate — View of top plate looking west showing buckling ofwest plate over east plate. Diagram at right shows orientation of specimen and location offailure.
PCI JOURNALIMay-June 1984 69
5'-0" SQUARE HOLLOW PIER11/2" WALL THICKNESSI" = 8,680 psi PRESTRESSING = 1.211psi
4000
}3000
UQ
4U
2000wUtr0
1000Ja
Q
M n •n
_
(Material Failure)
Fallure^Point
G
M -max Pmax(CompressionFlange Buckling}
00 1000 2000 3000 4000
MOMENT CAPACITY (K-FT.)
Fig. 30. Specimen 1 — Interaction diagram.
teraction diagram shown in Fig. 32 for10,000-psi (70 MPa) concrete shows thefailure point. Once again, the failureload is very close to the predicted failureenvelope. These two specimens con-firmed two points in the interaction dia-gram at widely separated eccentricities.
The two points are on the flat portionof the diagram which are points of criti-cal concern in most pier failures,Therefore, this data can be consideredadequate as a verification of that portionof the diagram. Testing of additionalspecimens at eccentricities between thetwo tested would serve as additionalconfirmation of the straight portion ofthe diagram. When eccentricities aregreat enough to cause tension in thebottom plate, the interaction diagramcloses rapidly to the point where overallstability and local plate buckling are ap-
proximately equal. Therefore, this con-dition is not crucial; however, an addi-tional test at the apex of the curve is rec-ommended.
CODE PROVISIONSRequirements such as minimum steel
and concrete cover limit the thinness ofplates used in bridge construction to acondition where thin plate bucklinggenerally is not a problem. Thin sec-tions are not considered due to durabil-ity and impact concerns. Therefore, cri-teria for concrete thin plate buckling arenot included in the current codes. Withthe advent of high quality dense con-crete, these limitations have the poten-tial ofheing removed, which would thenrequire the incorporation of bucklingcriteria into codes of practice,
70
Seg .Seg Seg. LoadRams 6A 5 4 Cells
N
Fig. 31. Specimen 2A - Orientation of specimen and overall view of north plate showingbuckling in Segment 6A.
The current AASHTO Specificationscontrol thin plate buckling of steelmembers by restricting the width ofplates through utilization of stiffeners.These restrictions are based on thestudies of thin plate buckling character-istics of steel members. Similar designcriteria have been developed for thinplates of concrete as a result of thisstudy. The buckling failure mode forthin plates of concrete is brittle com-pared to the ductile behavior of steel.This brittle type failure indicates that aconservative code provision should beadopted pending further studies of thinplate buckling.
Therefore, it is reconunended thatplate thinness he limited to:
Width b -- IQ (2)Thickness t
without the rational analysis providedby this study. This limit applies only torectangular hollow piers with k = 4.0.Other limits must be established fordiffering bridge member configurations.This limit would insure that pier designwould be controlled by overall stabilityrather than local plate buckling.
It was previously stated that the ten-sile strength and modulus of rupture ofhigh strength concrete follow the trendsestablished for normal strength con-crete. This implies that high strengthconcrete flexural members designed tothe normal allowable tension stress cri-
PCI JOURNAUMay-June 1984 71
5'-0" SQUARE HOLLOW PIER1 1/2" WALL THICKNESSf, = 10,000psi PRESTRESSING = 1,218psi
4000
3000Q
a
02000
wQcc0U-
1000.1
x
Mn-Pn_
(Material Failurel
M max - Pmax
ICampression 'Flange Buckling) Failurel
Point i
00 1000 7000 3000 4000
MOMENT CAPACITY (K-FT.I
Fig. 32. Specimen 2A — Interaction diagram.
teria will exhibit a superior margin ofsafety against cracking under serviceload conditions, because the nominalmargin of 1.5 f' is approximately 40percent greater for 10,000-psi (70 MPa)concrete than for 5000-psi (35 MPa) con-crete. Full advantage could be taken ofthis fact by revising allowable stresscriteria in the following manner:
1.6.6—ALLOWABLE STRESSES(B) Concrete(2) Stress at service load after losses
have occurred; tension in theprecompressed tensile zone(a) For members with bonded
reinforcement:For normal weight
concrete ..7.5/] - 106 psiFor sand-lightweight
concrete ..6.3/ f^ - 106 psiFor all other lightwe ight
concrete ..5.5^ f,' - 106 psi
(b) For severe corrosive exposureconditions, such as coastalareas:
For normal weightconcrete ..7.5 ,: - 318 psi
For sand-lightweightconcrete ..6.3,f f, - 318 psi
For all other lightweightconcrete ..5.5VT - 318 psi
This proposed code revision wouldenable the designer to pmvide a con-stant margin of safety against flexuralcracking in prestressed girders. That is,the same increase of applied momentover design service load moment willcause flexural cracking, irrespective ofconcrete strength. The allowablestresses given here are calibratedagainst the AASHTO standard of 5000psi (35 MPa) (see Article 1.6.6) and thepresent criteria of 6f] and 3,J f allow-able tension allowed for normal and se-vere exposure, respectively.
72
SUMMARYProduction and utilization of high
strength concrete is rapidly becoming aviable concept in construction. Its ap-plication to precast prestressed concreteis readily apparent. The analytical de-sign studies discussed demonstrate thebenefits of using high strength concretein flexural members in addition to com-pression members. These benefits in-clude increased span lengths, reduceddead loads, and greater load capacities.
The potential for using thin plates ofconcrete fabricated on flat beds in thecustomary manner of the precast con-crete industry is shown to be advanta-geous. The analytical studies resulting inthe development of a computer pmgrammake a rational analysis of these thinplates feasible. This would insure con-trol of the section by design criteriaother than local buckling of that thinplate.
The current AASH1'0 8 Specificationsare not conducive to the use of highstrength concrete. Although use is notrestricted, Article 1.6.6 does not encour-age its use. Further, the restrictions onthickness of member and cover oversteel do not permit full utilization of thehigh stength dense concrete that couldbe provided for thin sections. Thesecriteria need to he constantly reviewedby code authorities and changes madesimilar to the provisions presented inthis report. Continuing efforts should bemade by professional bodies, gov-ernmental agencies, and code authori-ties to more rapidly implement intopractice application of the knowledgebeing developed by numerous re-searchers on the application of highstrength concrete to structural members.
REFERENCESFreedman, S., "High Strength Con-crete," Modern Concrete, V. 34, Nos.
6-10 , October 1970, pp. 29-36; Novem-ber 1970, pp. 28-32; December 1970, pp.21-24; January 1971, pp. 15-22; and Feb-ruary 1971, pp. 16-23.
2. Blick, R. L., "Some Factors InfluencingHigh Strength Concrete," Modern Con-crete, V. 36, No. 12, April 1973.
3. Chicago Committee on High-RiseBuildings, "High Strength Concrete inChicago High-Rise Buildings," ReportNo. 5, February 1977.
4. Nilson, Arthur H., and Slate, Floyd O.,"Structural Properties of Very HighStrength Concrete," Second ProgressReport, NSF Grant ENG 78 05124,School of Civil and Environmental En-gineering, Cornell University, Ithaca,New York, 1979.
5. Wang, P. T., Shah, S. P., and Naaman,A. E., "Stress-Strain Curves for Normaland Lightweight Concrete in Compres-sion," ACI Journal, V. 75, No. 11, No-vember 1978, pp. 603.611.
6. Kaar, P. H., Hanson, N. W., and Capell,H. T., "Stress-Strain Characteristics ofHigh Strength Concrete," Research andDevelopment Bulletin RD051.O1D,Portland Cement Association, Skokie, Il-linois, 1977.
7. Perenchio, W. F., and Klieger, P., "SomePhysical Properties of High StrengthConcrete," Research and DevelopmentBulletin RD056.01T, Portland CementAssociation, Skokie, Illinois, 1978.
8. AASHTO, Standard Specifications forHighway Bridges, Twelfth Edition,American Association of State Highwayand Transportation Officials, Washing-ton, D.C., 1977.
9. ACI Committee 318, "Building CodeRequirements for Reinforced Concrete(ACI 318-77)," American Concrete In-stitute, Detroit, Michigan, 1977.
10. Franklin, D. H., "Shuhenacadie BridgeTames Angry River," Concrete Interna-tional, V. 1, No. 2, February 1979, pp.73-82.
11. Priest, H. M., Design Manual for HighStrength Steels, U.S. Steel Company,Davis & Ward, Inc., Pittsburgh, Pennsyl-vania, 5th Printing, 1957.
12. Bleich, F., Buckling Strength of MetalStructures, McGraw-Hill Book Com-pany, 1952.
PC i JOURNAL/May-June 1984 73