Welding Innovation Vol. XVIII, No. 3, 2001

Engineering for Rehabilitation

of Historic Metal Truss BridgesBy Frank J. Hatfield, P.E.

Professor Emeritus, Civil and Environmental Engineering

Michigan State UniversityEast Lansing, Michigan

Editor’s Note: An earlier version of thispaper was presented at the 7thHistoric Bridges Conference inCleveland, Ohio, in September 2001,and was published in the proceedingsof that conference.

Introduction

The Calhoun County Historic BridgePark southeast of Battle Creek,Michigan, displays a collection of reha-bilitated metal truss bridges for useand enjoyment by pedestrians. Fromthe perspective of a structural engi-neer, it was instructive to investigatethe general feasibility of rehabilitatingcentury-old metal truss highwaybridges for pedestrian service consis-tent with modern standards for safe-ty1,2,3 and historic integrity14.Engineering aspects of rehabilitationare discussed for bridges that are nowin the Park, specifically:

• 133rd Avenue bridge (Figure 1), apin-connected half-hip Pratt ponytruss spanning 64 ft. (19.5 m), erect-ed in 1897 by the Michigan BridgeCompany to cross the Rabbit Riverin Allegan County, Michigan.

• Twenty Mile Road bridge (Figure 2),a 70 ft. (21 m) long riveted Pratt ponytruss that spanned the St. JosephRiver in Calhoun County. Physicalfeatures hint that this bridge wasdesigned for railway service. Thebuilder has not been identified andseveral sources date construction tothe early twentieth century.

• Gale Road bridge, a pin-connectedskewed Pratt through truss built in1897 by the Lafayette BridgeCompany. Originally spanning 122ft. (37 m) over the Grand River inIngham County, Michigan, thisbridge currently is being re-erectedin the Park.

Six other bridges have been procuredand are awaiting rehabilitation beforebeing put in the Park, including thesethat also will be discussed

• Tallman Road and Bauer Roadbridges, nearly identical pin-connect-ed Pratt through trusses thatspanned about 90 ft. (27 m) over theLooking Glass River in ClintonCounty. Manufactured by the PennBridge Company and erected in1880, they are two of Michigan’soldest through trusses9.

• Charlotte Highway bridge, manufac-tured by the Buckeye BridgeCompany and erected in 1886. Priorto its recent removal (Figure 3), itcrossed the Grand River in IoniaCounty with a span of 177 ft. (54 m)and was one of very few double-intersection Pratt truss bridgesremaining in Michigan9.

FeasibilityInvestigation of feasibility involvescomparing historic and modern specifi-cations for bridge design, particularlythose governing materials and loads.During the period when the projectbridges were built, standards werepromulgated by individual iron andsteel producers, bridge designers andmanufacturers, owners (typicallymunicipal governments) and textbookauthors. These standards were numer-ous and varied; those cited are repre-sentative rather than comprehensive.

Figure 1. The rehabilitated 133rd Avenue bridge, installed at the CalhounCounty Historic Bridge Park.

Welding Innovation Vol. XVIII, No. 3, 2001

Strength of MetalsAlthough the quality of structural steelhas been perfected over the past cen-tury, the strength of low carbon steelsusually used in bridges has notchanged significantly (Table 1).However, the allowable stresses usedby bridge designers increased as con-fidence and understanding developed.This is reflected in the trend towardlower factors of safety illustrated byTables 1 and 2. Early bridge designersused factors of safety as high as six tocompensate for lack of quantitativeinformation. Today, based on results ofa century of research and experience,factors of safety of two or less are typ-ical. Modern specifications may allowlarger stresses in the old steel andwrought iron members of a historicbridge than did its designer.

Live Load

An old highway bridge may havebecome deficient in strength due tothe increased weight of trucks. In 1916Waddell17 advocated designing ClassC bridges for a single 6 ton (53 kN)truck weight, and Class A bridges foran 18 ton (160 kN) truck, noting that“Almost all of the old highway bridgesare incapable of carrying these newlive loads with safety.” The smallest

design vehicle load currently recog-nized is a two-axle truck weighing 15tons1 (133 kN). However, historicmetal highway bridges were designedto carry uniformly distributed loads inaddition to, or in lieu of, concentratedaxle loads to assure safety for lines ofwagons or automobiles, livestock, andcrowds of people, the latter being thelarger, or governing, distributed load.

Table 3 traces the trend and variationsin design values for distributed liveloads on highway bridges as well as

listing current design values for pedes-trian bridges2. Ranges reflect levels ofservice. This table demonstrates that,in general, the published design loadsfor old highway bridges exceed thecurrent requirement for pedestrianbridges. Bridges with long spans anddesigned for rural service may beexceptions.

Wind LoadIn contrast to distributed live loads,design wind loads have increased sig-nificantly. In 1901 Waddell advocateddesign loads of 250 and 150 lb/ft.(3.65 and 2.19 kN/m) on the loadedand unloaded chords, respectively, forclass A bridges with spans of 150 ft.(46 m) or less16, but by 1916 he had

Modern specs may allow larger stresses in the members

of a historic bridge than did its designer

Figure 2. The rehabilitated Twenty Mile Road bridge, shown in its new positionat the Historic Bridge Park.

Figure 3. Lifting the Charlotte Highway bridge from its original abutment. Thisend was lowered onto a barge prior to hauling the bridge across the river andup the other bank.

Welding Innovation Vol. XVIII, No. 3, 2001

increased those values to 320 and180 lb/ft.17 (4.67 and 2.63 kN/m). TheIllinois Highway Department designedfor the larger of 25 lb/ft.2 (1.2 kN/m2)on the vertical projection of each trussand of the deck, or 300 and 150 lb/ft.(4.38 and 2.19 kN/m) on the loadedand unloaded chords, respectively12.Modern specifications1,2 are muchmore demanding, requiring design forwind loads of 75 lb/ft.2 (3.6 kN/m2) on

the vertical projection of each trussand of the deck, plus 300 and 150lb/ft. (4.38 and 2.19 kN/m) on theloaded and unloaded chords, respec-tively (this lineal load is not requiredfor pedestrian bridges), plus 20 lb/ft.2

(0.96 kN/m2) upward on the deck.Clearly, historic bridges are unlikely tohave been designed for the wind loadscurrently mandated.

Structural Analysis and Design

The components of each of the reha-bilitated project bridges were analyzedto estimate design stresses associatedwith internal forces caused by speci-fied combinations of loads1 and actingon the original uncorroded membercross-sections. Allowable stresseswere computed from assumed materi-al properties3 and specified factors of

Table 1. Tensile strengths of steel and factors of safety for tension fracture at net section.

Table 2. Tensile strengths of wrought iron and factors of safety for tension fracture.

Source

Pottsville Iron & Steel Co.7

Carnegie Phipps & Co.7

IATM10

Waddell16

Burr and Falk4

Copper12

Michigan13

Bethlehem Steel Co.7

Waddell17

Ketchum12

AASHTO3

AASHTO1

1887

1889-1893

1900

1901

190119091910

1907-1119161920

pre 19051905-36current

Year Grade ofSteel

Yield stress,minimum,

ksi(MPA)

Ultimate stress,minimum,

ksi(MPA)

Allowable stresson net section,

ksi(MPA)

Factor ofsafety for fracture

for bridges

medium

medium

mediummedium

moving loadsmediummedium

ASTM A36

35 (241)

35 (241)

30 (207)

35 (241)

26 (179)30 (207)36 (248)

60 (414)

60 (414)

60 (414)

60 (414)

52 (358)60 (414)58 (400)

15.6 (108)

12.5 (86)

16 (110)18 (124)

10 to 25 (69 to 720)#15 (103)12.5 (86)16 (110)16 (110)26 (179)*30 (207)*29 (200)*

3.83.3

3.5 to 6.0@2.4 to 6.0#

4.0

3.8

2.0*2.0*2.0*

* for inventory rating # depending on service class and influence area

Source

Carnegie Kloman & Co.7

Waddell15

Phoenix Iron Co.7

IATM11

Waddell16

AASHTO3

1873188318851900

1901

Year Grade of Steel

Yield stress,minimum,

ksi(MPA)

Ultimate stress,minimum,

ksi(MPA)

Allowable stressksi

(MPA)

Factor ofsafety for fracture

wrought ironiron

refined irontest iron class Atest iron class B

stay-bolt ironwrought ironwrought iron

26 (179)

25 (172)25 (172)25 (172)25 (172)26 (179)

50 (345)

48 (331)48 (331)50 (345)46 (317)50 (345)

14 (97)8 to 12.5 (55 to 86)#

12 (83)

13 (90)14.6 (101)*

34.0 to 6.2#

3.8

* for inventory rating, less than 100,000 load cycles@ depending on span # depending on type of load, including impact factor

Welding Innovation Vol. XVIII, No. 3, 2001

safety1. For each component and loadcombination, the allowable stress wasdivided by the design stress. A ratioless than unity indicates need for mod-ification, while a ratio greater thanunity suggests that an acceptable levelof safety may be achieved withoutcompletely restoring corroded sections(in general, significant damage wasrepaired in the interest of historicintegrity and esthetics). The threerehabilitated project bridges werefound to have adequate capacity forpedestrian loading.

Unusual Features

The structural analysis of a truss usu-ally is a routine procedure. To simplifycomputations, the structural engineerassumes that each member transmitsforce only in the direction of its longitu-dinal axis. That is, the member is not

subject to transverse force (shear) orbending. This assumed behavior isachieved if the members are straightand connected at their ends by friction-less pins, longitudinal axes of membersare concentric at connections, andloads are applied to the truss only atconnections. Real trusses conform tothis idealization only approximately butmember forces may be computed withsufficient accuracy if the designapproaches the ideal conditions.

The Tallman Road bridge displays twopeculiar details that are contrary to theideal conditions and to subsequentpractice. The most obvious is the hipjoint, which has two pins rather thanone. One pin carries the vertical eyebarand the other carries the diagonal eye-bar pair. Because the longitudinal axesof the inclined end post, top chord, vertical and diagonal members do not

meet at a common point, bending isinduced in the end post and top chord.

The second peculiarity of the TallmanRoad bridge is that each lower chordeyebar spans two deck panels andhas three eyes: one at each end andone in the middle. When gravity loadis applied to a truss, the panel pointsnear midspan typically deflect down-ward more than those near the ends.If the truss conforms to the ideal con-ditions, the members rotate but remainstraight as the panel points deflect.Obviously this behavior cannot beachieved by a three-hole eyebar.Thus, these unusual lower chord eye-bars are subject to bending as well asaxial tension.

Strength Not Predicted byConventional Truss AnalysisConventional analysis predicts that thelower chord of a single-span throughtruss is always in tension when thebridge is carrying gravity load.However, the lower chords in the endpanels of the Charlotte Highwaybridge were observed to be slack (i.e.,subjected to compression rather thantension) when the bridge was in ser-vice in its original location. Thosemembers remained slack after thevehicular railings and deck wereremoved in preparation for moving the

bridge from its masonry abutments.However, when the bridge was freedfrom its inoperative expansion bear-ings, that end appeared to moveinland several inches and cracksopened where the wingwalls join theabutments. Apparently the upperchord and end posts had been func-tioning as an arch as well as restrain-ing displacement of the heavyabutments and fill.

Design wind loadshave increasedsignificantly

Table 3. Uniformly distributed design live loads for highway bridge trusses inpounds per square foot (kN/m2).

* Prescribes an impact factor, which is included in the tabulated values # For 16 foot (4.88 m) deck width

Whipple5

ASCE5

Waddell15

Waddell*16

American Bridge Co.*4

Michigan Highway Comm.13

Waddell*#17

Ketchum*12

Illinois Highway Comm.12

Wisconsin Highway Comm.12

AASHTO (pedestrian)#2

1846

1875

1883

1901

1901

1910

1916

1920

1920

1920

1997

100(4.79)100-70

(4.79-3.35)100-80

(4.79-3.83)170-113

(8.14-5.41)125-100

(5.99-4.79)100

(4.79)161-107

(7.71-5.12)151-116

(7.23-5.55)125

(5.99)120

(5.74)67

(3.21)

100(4.79)75-50

(3.59-2.39)90-80

(4.31-3.83)149-98

(7.13-4.69)125-94

(5.99-4.50)100

(4.79)144-95

(6.89-4.55)126-89

(6.03-4.26)100

(4.79)93

(4.45)65

(3.11)

100(4.79)60-40

(2.87-1.92)70-60

(3.35-2.87)120-80

(5.75-3.83)100-69

(4.79-3.30)100

(4.79)119-80

(5.70-3.83)103-60

(4.93-2.87)85

(4.07)50

(2.39)65

(3.11)

Source YearSpan

50 feet 100 feet 200 feet(15.2 m) (30.5 m) (61.0 m)

Welding Innovation Vol. XVIII, No. 3, 2001

Prior to lifting the six-panel BauerRoad bridge from its original abut-ments, the contractor removed railings,decking and stringers. Then a liftingsling was attached to the upper lateralstruts at the third points of the span.Conventional truss analysis predictsthat the bottom chord will be com-pressed when the bridge is lifted in thismanner. Since the bottom chord con-sists of eyebars, which have negligible

resistance to compression, it seemedlikely that the trusses would collapse.The fact that the lift was accomplishedwithout damage attests that the upperchord, hip joints and end posts pos-sess significant bending strength.

Conventional truss analysis mayunderestimate the strength of a metaltruss bridge. More comprehensiveanalysis techniques coupled with

detailed modeling of connections maymake it possible to quantify additionalstrength.

Inadequate Resistance to Wind LoadBy modern design standards, therehabilitated project bridges had inad-equate resistance to wind load. It wasnecessary to employ a provision1 thatpermits design wind speed to beadjusted from a nominal 100 MPH(45 m/s) to reflect favorable localconditions. The inland location of thePark and the low and sheltered sitesof the project bridges justify a designwind velocity of 70 MPH (31 m/s).Despite the resulting 50% reductionof wind force, the original anchorbolts typically were inadequate, andeach of the three bridges manifestedother deficiencies.

Analysis of the 133rd Avenue bridgepredicted that modern design windloads would cause net axial compres-sion of the windward lower chord eye-bars. Since eyebars have negligibleresistance to compression, they wouldbuckle and the truss would becomeunstable. This was corrected byinstalling an unusually heavy deck tocreate enough tension in the lowerchord to counteract the compressioninduced by wind. Alternatively, it mayhave been possible to rely on the deckor upper chord to stabilize the trussesas suggested in the preceding section.

The deck lateral ties of the TwentyMile Road bridge were evaluatedusing the assumed strength of steelproduced before 19053 and found tobe inadequate. The ties, like otherparts of this bridge (Figure 4) were toobadly corroded to be salvaged.Replacing them with new steel, in theoriginal sizes, was sufficient to providethe required wind resistance.

Structural analysis showed that theoriginal portal braces of the Gale Roadbridge were inadequate. Vertical strutshad been arc welded to the latticepanels sometime after construction,apparently to correct perceived weak-

Figure 4. Severely corroded sections of the Twenty Mile Road bridge werereplaced by welding new steel to sound original material.

Figure 5. Forge-welded loop eyebars like these are obsolete.

Welding Innovation Vol. XVIII, No. 3, 2001

ness, and localized bending of hori-zontal members occurred after thesereinforcements were installed. Theoriginal portal braces will be retained

for display but not installed on therehabilitated bridge. The replacementportal braces have larger connectiongussets than the originals, and the lat-tice is steel angles of the same widthas the original flat bars. The configura-tion and overall dimensions of theoriginal portal braces are duplicated.

Features Not Covered in CurrentSpecificationsPony trusses and loop eyebars (Figure5) are obsolete, and there are no cur-rent standards to guide assessment ofthese features. Pony trusses are proneto lateral instability of the top chords.That is, the bridge tends to fold inwardunder heavy load. The two rehabilitat-ed pony trusses were checked for sta-bility by Holt’s method8 and both werefound to have adequate factors ofsafety for pedestrian loading.

Single-load tests of seventeenwrought iron loop eyebars reported by

Ellerby et al6 demonstrated that frac-ture may occur at a forge weld ratherthan in the body of a bar, sometimesat a load significantly less than thedesign strength of the bar. As part ofthe same investigation, twenty-sixwrought iron loop eyebars wererepeatedly loaded to working stresslevel. The number of load cycles tofailure suggests that the bars couldhave remained in highway service formany more decades. When fatiguefractures finally did occur, they were inthe loops (except for two bars, whichinitially had large cracks at forgewelds). The investigators speculatedthat repeated flexing of the loops wasa critical factor and noted the deleteri-ous effect of poor fit on the pin.

The usual practice for the projectbridges is to inspect eyebar eyes andforge welds visually and by ultrasonicand dye penetrant methods (Figure 6).Cracks are ground out and bars arebuilt back to original profile by arcwelding. Testing has shown that care-ful arc welding restores full strength6.

Conclusion

Selected historic metal truss bridgesthat are rehabilitated to near-originalcondition can satisfy modern safetystandards for pedestrian service. Thisis demonstrated by the bridges on display in the Calhoun County HistoricBridge Park.

Acknowledgements

Dennis A. Randolph, ManagingDirector, Calhoun County RoadCommission and Board of PublicWorks, developed the concept for the Historic Bridge Park and providesdirection and support. The projectdirector is Vern Mesler and the historian is Elaine Davis.

References

1. AASHTO (1996), Standard Specifications forHighway Bridges, sixteenth edition, Am.Assoc. State Highway and TransportationOfficials.

2. AASHTO (1997), Guide Specifications forDesign of Pedestrian Bridges, Am. Assoc.State Highway and Transportation Officials.

3. AASHTO (2000), Manual for ConditionEvaluation of Bridges 1994, second edition asrevised by the 1995, 1996, 1998 and 2000Interim Revisions, Am. Assoc. State Highwayand Transportation Officials.

4. Burr, C. E., and Falk, M. S. (1905), TheDesign and Construction of Metallic Bridges,John Wiley and Sons, New York.

5. Edwards, L. N. (1959), A Record of Historyand Evolution of Early American Bridges,University Press, Orono, Maine.

6. Ellerby, H. A., Sanders, W. W., Jr., Klaiber, F.W., and Reeves, M. D. (1976), Service Loadand Fatigue Tests on Truss Bridges, J.Structural Division, Am. Soc. Civil Engineers,n. ST 12, Dec., p. 2285-2300.

7. Ferris, H. W., ed. (1990), Iron and SteelBeams 1873 to 1952, tenth printing, Am. Inst.of Steel Construction.

8. Galambos, T. V., ed. (1988), Guide to StabilityDesign Criteria for Metal Structures, fourth edi-tion, John Wiley and Sons, New York.

9. Hyde, C. K. (1993), Historic Highway Bridgesof Michigan, Wayne State University Press,Detroit.

10. IATM (1900), Proposed StandardSpecifications for Structural Steel for Bridgesand Ships, Bulletin 8, May, Proceedings, v. I,Int’l. Assn. for Testing Materials - Am. Section,p. 81-86.

11. IATM (1900), Proposed StandardSpecifications for Wrought Iron, Bulletin 17,May, Proceedings, v. I, Int’l. Assn. for TestingMaterials - Am. Section, p. 129-134.

12. Ketchum, M. S. (1920), The Design ofHighway Bridges of Steel, Timber andConcrete, second edition rewritten, McGraw-Hill Book Co., New York.

13. Michigan, State of (1910), GeneralSpecifications for Highway Bridges, ThirdBiennial Report of the State HighwayCommission, Wynkoop Hollenbeck CrawfordCo., Lansing, Mich.

14. NPS (1992), The Secretary of the Interior’sStandards for the Treatment of HistoricProperties, U. S. Department of the Interior,National Park Service, PreservationAssistance Division, Washington.

15. Waddell, J. A. L. (1883), General Suggestionsas to the Conditions Proper to be Required inOrdinary Iron Highway Bridge Construction,Transactions, Am. Soc. Civil Engineers, v. 12,p. 459-478.

16. Waddell, J. A. L. (1901), De Pontibus: aPocket-Book for Bridge Engineers, first editionsecond thousand, John Wiley and Sons, NewYork.

17. Waddell, J. A. L. (1916), Bridge Engineering,first edition second thousand, John Wiley andSons, New York.

Conventional truss analysis may underestimate

the strength of a bridge

Figure 6. Dye penetrant inspection ofa forge weld.