United StatesDepartment ofAgriculture
Forest Service
ForestProductsLaboratory
ResearchPaperFPL-RP-511
Behavior ofStress-LaminatedParallel-ChordTimber Bridge DecksExperimental andAnalytical StudiesAl G. DimakisMichael G. OlivaMichael A. Ritter
Abstract AcknowledgmentsThe authors wish to express their sincere apprecia-tion to Mr. Russ Moody, Dr. Bill McCutcheon, andMr. Earl Geske of the Forest Products Laboratory;Mr. Dave Summy of the U.S. Department of Agricul-ture, Forest Service Eastern Region, Regional Office;and Mr. Howard Haselschwardt and Mr. Raino Makiof the Hiawatha National Forest, Michigan. Their as-sistance and encouragement in the construction of theexperimental prototype parallel-chord bridge was a pri-mary factor in the completion of this project.
The use of stress lamination for constructing timberbridges may provide a solution to the urgent need forrehabilitating and replacing U.S. highway bridges. Thisreport describes the development, construction, test-ing, and analysis of a new type of stress-laminatedtimber bridge: the parallel-chord bridge. A full-scalelaboratory test was conducted on a stress-laminatedparallel-chord bridge made with Vierendeel trusses. Asimilar set of shorter trusses was built for a prototypebridge on the Hiawatha National Forest in Michigan.Test results showed that both of these bridges havegreater stiffness and can span longer distances thanstress-laminated solid-sawn timber bridges. The stress-laminated parallel-chord bridge system effectively trans-fers applied loads to a wide portion of the deck trusses.Anchorage configurations have little effect on load re-sisting behavior as long as the anchorages transfer pre-stressing force into both chords and webs. Good corre-lation was found between analytical and experimentalresults. The stress-laminated parallel-chord bridge iseasy to build, but the cost of the superstructure may belimiting.
Keywords: Timber, wood bridges, stress laminating,parallel-chord truss
October 1992
Dimakis, Al G.; Oliva, Michael G.; Ritter, Michael A. 1992.Behavior of stress-laminated parallel-chord timber bridgedecks: Experimental and analytical studies. Res. Pap.FPL-RP-511. Madison, WI: U.S. Department of Agricul-ture, Forest Service, Forest Products Laboratory. 19 p.
A limited number of free copies of this publicationare available to the public from the Forest ProductsLaboratory, One Gifford Pinchot Drive, Madison, WI53705-2398. Laboratory publications are sent to morethan 1,000 libraries in the United States and elsewhere.
The Forest Products Laboratory is maintained incooperation with the University of Wisconsin.
ContentsPage
Introduction . . . . . . . . . . . . . . . . . 1
Background . . . . . . . . . . . . . . . . 1
Problem . . . . . . . . . . . . . . . . . . 1
Objective and Scope . . . . . . . . . . . . . . 2
Methodology . . . . . . . . . . . . . . . . . 2
Member Feasibility . . . . . . . . . . . . . 3
Individual Truss Analysis . . . . . . . . . . 3
Evaluation . . . . . . . . . . . . . . . . . 4
Results and Discussion . . . . . . . . . . . . 12
Behavior of Individual Trusses . . . . . . . . 13
Simulated Truck Loading . . . . . . . . . . 13
Panel Joints . . . . . . . . . . . . . . . . 14
Prestress Distribution . . . . . . . . . . . . 15
Cyclic Loading . . . . . . . . . . . . . . . 16
Analytical and Experimental Results . . . . . 16
Conclusions and Recommendations . . . . . . . 16
Resistance to Truck Loading . . . . . . . . . 16
Ease of Construction . . . . . . . . . . . . 17
Prestress Rod Anchorage System . . . . . . . 17
Repeated Load Resistance . . . . . . . . . . 17
Composite Action of Parallel-Chord Truss . . . 17
Analytical Modeling . . . . . . . . . . . . . 17
Future Use of Parallel-Chord Stress-Laminated Bridges . . . . . . . . . . . . . 18
References . . . . . . . . . . . . . . . . . . 18
Appendix . . . . . . . . . . . . . . . . . . 19
Behavior of Stress-LaminatedParallel-Chord Timber Bridge DecksExperimental and Analytical Studies
Al G. Dimakis, Research Engineer1
Weyerhaeuser Corporation, Tacoma, Washington
Michael G. Oliva, Associate ProfessorDepartment of Civil and Environmental Engineering,University of Wisconsin, Madison, Wisconsin
Michael A. Ritter, Research Engineer2
Forest Products Laboratory, Madison, Wisconsin
Introduction
The size of the highway system in the United Statesand the continued need for maintenance and repairhave created tremendous pressure on local, state, andfederal governments to seek new, efficient, and eco-nomic means of insuring safety. Bridges, critical linksin the transportation system, have begun to receivelong-overdue attention. The USDA Forest Service,which is directly responsible for thousands of miles inthe highway system, has recently taken a lead role inassessing new technological means of addressing high-way bridge deficiencies through innovative uses of tim-ber in bridge construction. As part of this role, directfunding has been allocated toward basic and appliedtimber bridge research through the Forest ProductsLaboratory. Efforts are also underway to bring the nec-essary expertise to design engineers. This report de-scribes the development of a new type of timber bridge,which may prove to be an efficient and economical sys-tem for replacing many bridges that are beyond repair.
Background
Stress laminating, a new technique for efficiently andrapidly building new bridges or replacement decks, was
1 Formerly Graduate Student, University of Wisconsin,Madison, Wisconsin.2 Formerly Structural Engineer, National Forest System,Division of Engineering, Washington, D.C.
introduced in Ontario, Canada, in 1976. In this bridgesystem, a solid timber plate, which acts as a deckbridge, is constructed by vertically laminating individ-ual pieces of lumber using compressive prestress.
The prestress causes compressive stress between thefaces of the individual laminae. This compression cre-ates friction, which allows shear forces to be transmit-ted between the laminae. In addition, when transversebending occurs, the laminae are held together by thecompressive force, which is designed to be high enoughto counteract the flexural tension stress likely to be in-duced by applied bending loads. A bridge deck con-structed of stress-laminated solid timber is shown inFigure 1. The entire set of laminae is squeezed to-gether tightly so that the laminae act as a single, solidunit.
Problem
Although stress-laminated solid-lumber decks appear tobe a very attractive solution for certain bridge applica-tions, the stiffness of a solid deck is limited. The spanof a stress-laminated solid-timber bridge deck should belimited to <40 ft for HS 20-44 truck loading if 16 in.laminae are used and deflections are to be controlledat a reasonable level (Oliva and Dimakis 1988). (SeeTable 1 for conversion factors for SI units of measure-ment.) Elimination of cracking or deterioration in anasphalt wearing surface may require a shorter span.
Figure 1—Bridge deck constructed of stress-laminated solid timber. Dimensions are in feet.
Table 1-Factors for converting English units ofmeasurement to SI units
English unitConversion
factor SI unit
inch (in.) 25.4foot (ft) 0.3048board foot (fbm) 0.0024lb/in2 (stress) 6.89
millimeter (mm)meter (m)cubic meters (m3)kilopascal (kPa)
The stress-laminating technique could be used forlonger spans if a method could be developed for in-creasing the deck depth, thereby providing the requiredstiffness. Attaining a greater depth while maintainingthe stress at efficient levels can only be accomplishedby changing the shape of the laminae. Thus, the prob-lem in developing longer span stress-laminated bridgesis in identifying or developing fabricated laminae witha deeper cross-sectional configuration, which providesimproved efficiency in both strength and stiffness.
Objective and Scope
The objective of this study was to develop and evaluatebridge systems consisting of stress-laminated parallel-chord trusses. The scope of the work included anevaluation of previous work on individual members,development of individual trusses, and construction andevaluation of a full-scale laboratory bridge and a proto-type field structure.
Methodology
To meet the study objectives, we conducted an ana-lytical investigation of the behavior of individual fab-ricated members with various cross-sectional config-urationss, followed by experimental evaluation to de-termine whether bridges could be assembled by stresslaminating the members together. The feasibility ofvarious potential member sections and preliminary eval-uation of individual components were based on previ-ous work. Within this study, methodology involved apreliminary analytical analysis of bridge systems com-posed of stress-laminated members followed by theexperimental and analytical evaluation of two full-scalebridge structures.
2
Figure 2—Possible configurations for parallel-chord members initially identified as likely com-ponents for a stress-laminated parallel-chordbridge deck.
Member Feasibility
Four types of member configurations were initiallyidentified by Oliva, Tuomi, and Dimakis (1986) aspotentially feasible for stress-laminated bridge appli-cations (Fig. 2). Each member was subjected to com-putational analysis to determine necessary elementsizes and connection strengths between elements. Onthe basis of those initial studies, the composite I-beammembers and the multi-leaf truss were discarded. Thecomposite I-beam exhibited a weakness within its ply-wood webs. Considerable horizontal shear occurs inthe webs of laminae from bridge truck loadings, andseveral layers of plywood are apparently necessary tomeet strength requirements. The multi-leaf truss sec-tion could be provided with sufficient individual ele-ment strength. However, economic fabrication of multi-leaf trusses is limited by the connections between weband chord members and their strength requirements.The Vierendeel-type truss and the metal-plate trussappeared to be feasible and were examined in furtherdetail.
Individual Truss Analysis
Oliva and Lyang (1987) examined Vierendeel parallel-chord trusses in detail. This study entailed testing ofthe connections and trusses, and analytic simulation
Figure 3—Chord-to-web joint test specimen.Shear stiffness of connection was determined byapplying downward load on center wood block.
of their behavior. Shear tests on individual specimenssimulating chord-to-web joints were used to determinetruss shear stiffness. Variables considered in the studyincluded the size of individual members in the joint,and the size, number, and length of connecting fasten-ers (steel dowels). A typical test specimen is shown inFigure 3. Each specimen was tested to failure (that is,large slip with no increase in load) by applying a down-ward load on the center wood block while the outerpieces were supported.
In addition to shear tests, three types of 25-ft-longVierendeel trusses (Fig. 4) were tested to determineindividual stiffness and to allow correlation between an-alytically predicted behavior and measured response.Two trusses were 24 in. deep with 4- by 8-in. chordsand webs. The third truss was 30 in. deep with 4- by10-in. chords and webs. Each web block was attachedto the chords by two steel dowels. The trusses were ini-tially built without butt joints in the chords. Midwayin the testing program, the chords were sawn throughto simulate butt joints.
In the study by Oliva and Lyang (1987), Vierendeeltrusses were tested individually and in groups to iden-tify overall section stiffness. The known web shear stiff-ness for the Vierendeel truss was then used in an an-alytical model to predict the behavior of a complete
3
Figure 4—Configurations of three types of Vierendeel trusses subjected to individual loading.
truss. This information was compared with test re-sults to verify the analytical modeling techniques andto determine whether the behavior of the truss couldbe predicted. Verified techniques were used to ana-lyze a series of trusses with different configurations toidentify a truss design with sufficient stiffness andstrength for bridge applications. Because of the su-perior performance and simplicity of fabrication, theVierendeel truss was selected for the full bridge andbecame the focus of this study.
The analytical model developed by Oliva and Lyangwas used to develop Vierendeel truss configurationsfor evaluating stress-laminated bridges. Initially, themodel was used to simulate a truss with continuous topand bottom chords and semirigid web-chord connec-tions. The shear stiffness obtained from previous testswas used at the chord-to-web joints. The load defor-mation response of a 52-ft Vierendeel truss was thenpredicted using the verified analytical model. Becausediscontinuous top and bottom chords would be requiredfor bridge evaluations, the model was modified to ac-count for three types of butt joints. Analyses were con-ducted for 51-, 41.5-, and 32-ft spans, measured center-to-center of supports.
4
Evaluation
Based on the analysis of individual trusses and the an-alytic evaluation of a stress-laminated truss bridge, twoparallel-chord stress-laminated bridges were designedand built for this study. The first bridge was assembledin the laboratory and used to obtain performance data,which were correlated with the analytically predictedbridge response. Based on the knowledge gained fromthe analytical studies and laboratory testing, a secondexperimental bridge was designed and constructed overMormon Creek on the Hiawatha National Forest. Thelaboratory bridge and the Mormon Creek Bridge areshown schematically in Figure 5.
Laboratory Bridge TestsThe laboratory tests were designed to meet the fol-lowing objectives: (1) to test the capability of theparallel-chord system to resist truck loading, (2) to de-termine whether the type of prestress rod anchorageaffects resistance to truck loads, (3) to test the stiffnessacross the proposed joint between prefabricated panels,(4) to appraise the stiffness changes and accumulateddeflection after one-half million cycles of simulatedtruck loading, (5) to determine how the force from a
Figure 5—Schematic view of the laboratory andMormon Creek parallel-chord bridge decks.
prestress rod anchorage flows into the wood and howthe stress varies within the wood, and (6) to providesufficient data on the response of the system to validatean analytical model for predicting response.
Bridge Configuration —A laboratory test bridge was con-structed of stress-laminated trusses. The 52-ft-long by9-ft-wide bridge was made with the following configura-tion:
l 27 Vierendeel-type trusses, 3.88 in. wide by 24 in.deep with rough-sawn 4- by 6-in. chords and 4- by12-in. webs planed to 3.88 in. Material was visuallygraded No. 1 or better Douglas Fir, pressure treatedwith creosote.
l Butt joints on top and bottom chords, only at web-to-chord connections. Joints were placed in a precisesequence:
(1) Only one in four trusses was allowed to have abutt joint, at top or bottom chord, within any 8-ftlength parallel to the bridge span.
(2) Butt joints in adjacent trusses were separated byat least 8 ft in the direction of the span.
l Chord member connected to web block by three0.63-in.-diameter steel dowels.
l Camber of approximately 2 in. provided in eachtruss to offset dead load deflection.
The individual parallel-chord trusses were very flexibleas a result of the butt joints in the chords. Temporarysupports had to be provided at midspan locations topreserve the camber in the trusses until it was lockedin by the stress-laminating process. After stressing,the intermediate support was removed. The laboratorybridge system is illustrated in Figures 6 to 8.
Figure 6—End view of laboratory parallel-chordbridge deck. Deck rests on 12- by 12-in. timberon concrete abutment. Instrumentation referenceframe is mounted above deck.
Figure 7—Side view of laboratory bridge deck.Posts and steel plates provide anchorage fortransverse prestressing rods.
Figure 8—Hydraulic jack used to apply tensionforce on prestress rod of laboratory parallel-chordbridge.
5
Rod Anchorage Configurations — Adequate transfer oftransverse prestressing from the tensioning rods intothe wood of the trusses, particularly those trusses ad-jacent to the edge, is essential to the performance ofa bridge system. Because stress laminating had notbeen used previously on parallel-chord members, wedid not know if prestressing rods were needed in bothtop and bottom chords, as would be required if the sys-tem were considered two solid decks spaced a short dis-tance apart. Another uncertainty was whether the rodswould perform best if placed within the web blocks orif threaded through the web openings, which would fa-cilitate and simplify construction. A series of these dif-ferent options was investigated to determine the actualprestress distribution from the tensioning rods as wellas the capability of the anchorage system to effectivelylaminate the individual trusses. When the rods wereplaced within the web openings, a separate bulkheadmember was required at the rod anchorage to trans-fer the compression equally into both top and bottomchords.
Coincidentally, the bulkhead member appeared to beable to serve a dual purpose. If a vertical post wereused as the bulkhead, it could be extended upwardabove the deck and possibly serve as a rail post (pend-ing further studies to determine vehicle impact effectson system behavior). With this dual use in mind, theanchorages were designed to be composed of a set ofsteel plates bearing on the post, which transferred theforce to the top and bottom chords, its illustrated inFigure 9. The specification for the posts required a24FV4, 8.75- by 12-in. Douglas Fir glulam member;the actual posts delivered to the laboratory and used inthe tests were solid-sawn 8- by 10-in. members, whichperformed poorly, as described later in this report.
The rods and anchorages used in the laboratory bridgewere constructed in four configurations:
Configuration 1: One 1-in.-diameterrod in each web opening and two 1-in.-diameter rods in each web block.Because the goal was to obtain a uni-form compressive stress between everyportion of the trusses, more force (andhence more rods) was required in theportions of the trusses where the webblocks were placed. The rod in the webopening was assumed to cause prestressin the chords in its vicinity. The rodsin the web block were assumed to pre-stress the blocks together as well as theportions of the chords in the vicinityof the rods. The specific placement ofthe rods and the size of steel anchorage
6
Figure 9—Posts used to distribute prestressingforce into top and bottom chords.
plates for this configuration are shownin Figure 10.
Configuration 2 : Similar to configu-ration 1, but effect of joint betweenpanels simulated by adding 6- by 4-in.members on one side of deck (Fig. 11).
Configuration 3 : Similar to configu-ration 2 in respect to rods within webopenings. Two 1-in.-diameter rods ineach web block replaced by single 1.25in.-diameter rod (Fig. 12).
Configuration 4 : Four 1-in.-diameterrods inserted only through web open-ings (Fig. 13).
In configurations 1, 2, and 3, insertion of the rodsthrough holes in the web blocks required tight toler-ances and drilling of adequately sized holes to makeassembly possible. This could be a difficult or impos-sible task, particularly with bridges consisting of morethan one traffic lane. Configuration 4 is practical forbridge construction under such conditions. However,trusses without butt joints would probably be requiredunless temporary midspan support could be provided tomaintain camber until prestress was applied.
Simulated Truck Loading—The stressed bridge was loadedwith simulated double-wheel truck loading applied atmidspan for each anchorage configuration at internalprestress levels of 10, 20, 40, 50, and 70 lb/in2. This“point” load was actually applied over an area of 20in. transversely and 15 in. longitudinally to simulatethe tire contact area of a truck. The standard test con-sisted of a point load at the center of the span and the
Figure 10—Anchorage configuration 1. Tworods were placed through each web block andone rod through each web opening. Three typesof rod anchorages were used, as shown.
Figure 11—Anchorage configuration 2—detail ofone side. Spacer blocks were used to simulatetransfer of stress into deck at joint between twopanels.
Figure 12—Anchorage configuration 3. One rodwas placed through each web block and one rodthrough each web opening.
Figure 13—Anchorage configuration 4. Fourrods were placed in each web opening. Twoposts were used with two rods through eachpost with anchorages as shown. The anchor-age on one side used spacer blocks, as shown inFigure 11.
center of the lane. Other tests included (1) two point-loads spaced in the transverse direction and applied 1.5ft from each edge of the deck and (2) two point-loadsspaced in the longitudinal direction and applied 4 ft toeither side of the midspan (directly above web open-ings). In each case, the loads were applied until themaximum deflection reached approximately 1 in. Theseload tests were conducted with clear spans of 50, 39.5,and 31 ft by changing the location of the bridge sup-ports. The 39.5-ft clear span was tested under longitu-dinally asymmetric loading. Table 2 lists the 50 typesof simulated truck load tests used to determine bridgecapacity and effects of various prestress levels and an-chorage types.
Panel Joints-Continuity across the joint in a multi-panel prefabricated bridge is essential to transverse loaddistribution and wearing surface longevity. In mostconstruction situations, field labor time and cost canbe reduced by assembling trusses into panels and pre-stressing together before they are shipped to the erec-tion site. This approach requires a joint between pan-els, which may create a problem that stress laminatingspecifically tries to circumvent: if the panels are notprestressed together, a weak joint remains in the bridgesystem where differential movement may occur undertruck loading.
Previous work on solid stressed decks showed that ade-quate friction is provided between the laminae to avoidslip even under very low levels of prestress (Oliva andothers 1990). A 16-in.-deep solid stressed deck, with acoefficient of friction of 0.35, requires a theoretical levelof prestress of 10.4 lb/in2 to prevent interlaminar slip.Because the effective area for shear transfer across ajoint in a parallel-chord truss is potentially less thanthat provided by a 16-in.-deep stressed deck, a level ofprestress >10.4 lb/in2 would be necessary to preventinterlaminar slip between the panels, assuming a similarcoefficient of friction.
To investigate the integrity of the joint subjected tovehicular truck loading, two panels were constructedof 12 and 13 trusses, respectively. A continuous 6-in.by 4-in. spacer block was placed between the panels,attached to the top and bottom chords of an interiortruss at the joint, to secure interior anchorage of theindividual stressed panels (Fig. 14). The two panelswere connected by pressure provided by only the 1-in.-diameter rods. The rods were inserted through the webopenings and stressed to the required levels of prestress,resulting in the “design level” prestress between thetrusses in each separate panel, but only about one-thirdthe design prestress between the trusses at the joint.
The behavior of the joint was investigated under threelevels of prestress in the deck—20 lb/in2 (less than
8
Table 2—Truck load tests on laboratory bridgea
Anchorageconfiguration b Load location
Clear Prestress
span level
(ft) (lb/in2) b
1
1
2
2
2
3
3
4
4
5
Centerpoint load 50
Two transverseloads
50
Centerpoint load
Two transverseloads
50
50
Two longitudinalloads
50
Centerpoint load 50, 39.5, 31
Two transverse 50, 39.5, 31loads
Centerpoint load 50, 39.5, 31
Two transverse 50, 39.5, 31loads
Point load near 50center
1020405070
1020405070
5070
10205070
10205070
102050
102050
50
2050
205070
a Loads were applied to a magnitude required toproduce approximately 1 in. of deflection.
b Configuration1—one rod in web opening, two rods in web
block2—same as configuration 1 with special edge
bearing3—one rod in web block4—all rods in web opening5—panels jointed
Figure 14—Detail of joint between panels. Thetwo sets of stressing rods, which would appearto strike one another when the panels are pulledtogether, are actually offset from the drawing.Spacer blocks provided clearance between panelsfor anchorage plates.
normally acceptable in design), 50 lb/in2, and70 lb/in2, resulting in net pressure across the joint of6, 14, and 20 lb/in2, respectively. Centerpoint loadswere applied to the side of the joint until a center dis-placement of approximately 1 in. was achieved, causingmaximum shear across the joint.
Prestress Distribution—Transverse compressive deforma-tion, induced from a tensioning rod, is an indicator ofthe magnitude of compressive stress at any location inthe deck. The anchorage system used to transfer theprestressing force into the bridge dictates how concen-trated the compressive stresses are near the edge of thedeck. With small concentrated anchorage areas, it isimpossible to have uniform compression near the an-chorage. Uniform compression, or at least a minimumcompression above the amount associated with a lowerdeck stiffness, is desirable within the area of the deckwhere vehicular wheel loading is possible. The actualcompressive stress in a parallel-chord bridge is expectedto vary not only with distance from the anchorage inplan but also in the vertical direction.
Compressive deformations in the transverse directionwere measured within a region of the deck adjacentto a single anchorage point as the prestress in the rodor rods at the anchorage was varied. The amount oftransverse compressive displacement in the wood atany measured location was taken to be proportional tothe compressive prestress level at that location. Threetransverse stress distribution tests were conducted. An-chorage configuration 1 was used in two of these testsand configuration 3 in the third test (Table 3).
Table 3—Transverse stress distribution tests
Anchorage PostTest configuration location
1 1 Opening2 1 Web3 3 Web
Figure
101012
Three rows of linear variable differential transformers(LVDTs) were installed on the deck surface to measurethe variation in transverse compressive deformation asthe distance from the anchorage point increased. Therecorded displacements were normalized to units ofaverage strain and then plotted. Generally, three in-crements of force were applied to a single prestressingrod, and the associated compressive deformations in thetimber were recorded. These tests allowed us to deter-mine the effective level of compressive prestress withinthe deck and to estimate the average distance fromthe anchorage point at which the compressive stressbecame approximately uniform.
Cyclic Loading—Cyclic loading was imposed on thejointed deck to determine whether cumulative slipwould occur at the joint and whether repeated loadwould reduce elastic stiffness. The deck was tested atthe lowest level of average prestress expected under ser-vice conditions, 50 lb/in2 within the deck panels and14 lb/in2 between the two panels. The load was appliedat midspan, adjacent to the construction joint. Theloads were applied in a cyclic sinusoidal variation.
The first 100,000 cycles of load were applied at a rateof two cycles per second with a cyclic displacement of0.1 in. about a mean of 0.5 in. An additional 400,000cycles of load were then applied with a mean level ofdisplacement at 1.1 in. and peak amplitudes of 0.6 and1.6 in., and the rate of loading was decreased to 1 cycleper second. The peak displacement amplitude of 1.6 in.was calculated to be consistent with the expected de-flection caused by an HS 20-44 design truck on the 50-ft span.
Before the cyclic loading test began, a static stiffnesstest was conducted to determine the initial stiffnessof the system. Static tests were also conducted af-ter 100,000, 400,000, and 500,000 cycles to determinereductions in structural stiffness.
Data Measurement-Data measuring devices used inthis study included load cells (with strain gauges in afull-scale bridge arrangement), and LVDTs and dialgauges to measure displacements. Load cells were usedto measure the variation in load in a prestressing rod.Twenty full-scale bridge load cells were fabricated and
9
calibrated to a maximum load of 60,000 lb. In thisload range, the relationship of load to strain was lin-ear. Transverse compression deformations in the wood,caused by the force from a prestressing rod, were mea-sured by LVDTs. The LVDTs were installed in a spe-cial arrangement to study the variation of compressiondeformations in the transverse direction.
Taking advantage of symmetry, deflections were mea-sured over only one-quarter of the deck. For the truckload tests using anchorage configuration 1, deflectionswere measured at six “stations” between the supportand the center of the bridge in the longitudinal direc-tion. For all other anchorage configurations, displace-ments were only measured at station 1, located nearthe center of the span. At each station, seven deflec-tions were measured in the transverse direction betweenthe centerline of the bridge and the deck edge. This in-formation was used to verify the analytical model andto compare the elastic stiffness of the deck with differ-ent anchorage configurations, at comparable levels ofprestress. Deflections were measured with LVDTs, witha minimum linear range of ± 1 in. All LVDTs were cali-brated before testing. The LVDT linear range was wellbeyond the 1-in. limit.
A computer-aided data-acquisition system was used torecord the output from the differential transformers.The output from the load cells (load in the tensioningrods) was read by a strain indicator and two lo-channelswitch- and balance-devices. Certain residual deforma-tions occurring as a result of the dynamic truck load-ing tests were monitored by displacement dial gauges,which were read manually.
Analytical Modeling—The development of a closed formmathematical model to predict the behavior of openweb systems is a difficult task. Orthotropic plate the-ory was used in the past to analyze stiffened platesof constant thickness (Huffington 1956). Bridges con-structed with longitudinal open sections were inves-tigated successfully using orthotropic and articulatedplate theories (Bakht and Jaeger, 1985), but the lit-erature does not provide any information on methodsof analysis of open web systems with web openingslocated transversely.
The analysis of a system idealized as an orthotropicplate requires the determination of the effective plateelastic rigidity constants Dx, Dx, and Dxy, and a suit-able solution to the orthotropic plate equation satisfy-ing the existing boundary conditions. This latter prob-lem has been addressed in the literature for various setsof boundary conditions, and solutions exist for classicalcases.
10
The flexural and twisting rigidity constants Dx, Dy,and Dxy may be conceived of as applying to a homoge-neous orthotropic plate of constant thickness, whichis equivalent to the actual parallel-chord open-websystem. The term equivalent requires careful defini-tion because the orthotropic plate obviously cannotbe equivalent to the actual system in every respect.The deflection of the actual bridge and equivalent or-thotropic plate might need to be approximately thesame when both have the same loading and boundaryconditions.
The flexural and twisting rigidity constants of an or-thotropic plate do not depend on the boundary condi-tions of the plate nor on the distribution of the trans-verse load. This principle, coupled with an equivalencecriterion, provides a basis for determining the effectiverigidity constants. If the deflection of both the actualbridge and equivalent orthotropic plate can be foundfor some applied loading and a particular set of bound-ary conditions, the application of the equivalence cri-terion will provide an equation in which the only un-known quantities are the desired rigidity constants.
Care must be exercised in using the results from theanalysis of such an “equivalent” plate. For a givenbridge, the transverse variation of all response quan-tities, that is, deflection, moments, and shears, is notthe same. The variation pattern becomes more concen-trated (higher intensity at the position of the load) asthe order of the deflection derivative increases. Thus, adistribution width derived from a transverse deflectiondistribution pattern is larger than that obtained froma transverse moment distribution pattern. Distributionwidth design charts are usually based upon transverseprofiles of longitudinal moments.
A common commercial finite element analysis pro-gram was used to analyze an equivalent plate, whichwas intended to have the same response as the actualstress-laminated parallel-chord bridge deck. The re-quired rigidities for the orthotropic plate analysis wereobtained by first studying the stiffness of an individualtruss. An equivalent orthotropic plate thickness wascalculated from the average truss stiffness by settingthe modulus of elasticity equal to the tabulated valuefor visually graded No. 1 Douglas Fir (1.8 x 106 lb/in2)and equating the longitudinal rigidity of the plate withthe stiffness of the truss. The transverse plate rigidityand shear rigidity were estimated from relations deter-mined previously for solid-sawn stress-laminated decks(Oliva and others 1990). The rigidity relations weretaken as
Et = 149 f + 15,360
G = 134 f + 16,600
Table 4—Properties of equivalent orthotropic plateused in correlation studies
Deckclear Equivalentspan Trussa
(ft) (×103 lb/in) (×103 1b–in2/in)D x thickness
(in.)
31 1.9739.5 1.2450 0.72
a Mean values.
580,977 15.7715,838 16.8861,966 17.9
where
f is average prestress in the plate (lb/in2),
Et transverse modulus (lb/in2), and
G shear modulus (lb/in2).
These properties were used as initial assumptions be-cause there was no basis to expect that the rigiditiesin the laminated truss system would have similar rela-tions to those in the solid wood deck. The properties ofthe equivalent orthotropic plates used in the correlationstudies are listed in Table 4.
The general analytic ability of the equivalent platemodel was verified by comparison with measured re-sults of simulated truck loading in the laboratory testsover various spans and with various load locations. Thefinite element model of the equivalent solid orthotropicplate, which was used to predict the deformation re-sponse of the actual parallel-chord system, is displayedin Figure 15.
Field Bridge TestsThe Mormon Creek Bridge was installed with the coop-eration of Hiawatha National Forest personnel. Sinceconstruction, the bridge has been field tested undertruck loading, and it is being continuously monitoredfor stress loss and creep deflection. Results of thefield performance of this bridge have been reported(McCutcheon 1992).
The Mormon Creek Bridge is generally similar to thelaboratory test bridge but has a shorter span. The con-struction details are as follows:
• 40-ft-long, 38-ft clear span, 16.5-ft-wide, single-lanebridge.
• 48 Vierendeel-type trusses, 3-7/8 in. wide by 24 in.deep. Material visually graded No. 1 Douglas Fir,pressure treated with creosote.
• Chords made from rough-sawn 4- by 6-in. members;webs made from rough-sawn 4- by 12-in.-deep blocks.
Figure 15—Finite element grid used to modelone-quarter of bridge deck for 51-ft span. Sym-metry of deformation and support conditionswere used to allow modeling of portion of deck.Loads were applied on element near node 1.
• Butt joints in top and bottom chords the same asjoints in laboratory test bridge.
• Chord members and web blocks connected by three5/8-in.-diameter steel dowels.
• Prestressing rods spaced at 4-ft intervals.
• 150 x 103 lb/in2 steel rods in two diameters: 1 in.and 1.25 in.
• Trusses assembled and stressed together in two sep-arate panels by stress rods in solid web blocks forshipping to bridge site. No rods were placed in webopenings of panels until the bridge was finally assem-bled.
The Mormon Creek prototype bridge was designed as apart of this research project and fabrication bids weresought from private contractors. The bridge superstruc-ture was prefabricated in two panels composed of 24truss laminations each and prestressed by the fabrica-tor before shipping to the job site. The prefabricatedpanels were shipped to the site and lifted into place onthe timber pile caps of the substructure. The bridgewas then assembled by laminating the individual pan-els together through prestressing rods placed in thelarge web openings. Timber curbs and a plank wearingsurface were later added to the deck. Load cells wereplaced on specific prestressing rods to monitor long-term prestress loss resulting from creep, relaxation, andmoisture content variation within the timber.
11
Figure 16—Placement of prefabricated bridge panel for Mormon Creek parallel-chord bridge.Panels were set in place on prefabricated abutments and then stressed together. (M87 0281-14)
Deflection measuring devices ‘were permanently at-tached to the bottom of the superstructure to moni-tor creep deflection over time. Figures 16 and 17 showthe placement of the prefabricated panels and the finalstate of the bridge. Itemized costs of the superstructureare listed in the Appendix.
Results and Discussion
The results of the laboratory tests on both the indi-vidual trusses and the stress-laminated bridge deck aresummarized in this section. Included are laboratorytest results involving simulated truck loading, responseof a bridge with individual panels connected by a weak
Figure 17—Side view of assembled MormonCreek parallel-chord stress-laminated bridge.(M87 0324-11)
12
joint, determination of compressive prestress within thewood, response of the bridge to cyclic load tests, andcomparison of analytical and experimental results.
Behavior of Individual Trusses
From research reported by Oliva and Lyang (1987),a plot of load and associated shear slip for a typicalchord-to-web-block joint is shown in Figure 18. Fora complete individual truss, the average stiffness fora point load at center span, taken as the load dividedby the center displacement, was 3.98 x 103 lb/in for atruss without butt joints in either chord. Three of the25-ft trusses were stress-laminated together. Figure 19shows the load resisting behavior of the trusses; averagestiffness per truss was 6.73 x 103 lb/in. The increase instiffness after the trusses were laminated together wasdue to the stiffer web-to-chord connection created bythe prestress rod anchorage system. When a butt jointwas placed in the top and bottom chords of the middletruss, the stiffness was initially reduced, but the initialstiffness was regained with a second load repetition.
Three sets of 50-ft trusses from the laboratory bridgewere individually loaded after all the truck load testswere completed. Stiffness values obtained from threeindividual trusses with different butt-joint layouts were0.69, 0.33, and 0.55 x 103 lb/in with a centerpoint load.An additional test on a small group of four trusses,which were not stress-laminated together but wereforced to undergo equal simultaneous displacements,resulted in an average stiffness of 0.62 x 103 lb/in. Be-cause the trusses had been severely loaded during thesimulated truck tests, the results of the individual trusstests probably reflect decreased stiffness.
Simulated Truck Loading
A typical transverse displacement profile for simulatedtruck loading at center span, obtained while testingthe deck under anchorage configuration 1, is shown inFigure 20. The system behaved linear-elastically, withan average stiffness of 20.4 x 103 lb/in at a level of pre-stress representing service conditions; that is, 50 lb/in2.Stiffness is defined as the load at center divided by themidspan deflection. The displaced shape of the deckwhen subjected to two loads applied at a transversesection, each 1.5 ft from the deck edge, is shown inFigure 21.
The simulated truck wheel load was applied in five 0.2-in. increasing displacement increments and three de-creasing increments. Thus, the maximum displacementimposed on the deck in this stage of the testing wasapproximately 1 in. This displacement is below thatexpected from an equivalent HS 20-44 truck load. Themaximum load applied on the deck was 20.5 x 103 lb.That load gives approximately 80 percent of the longi-tudinal moment at the center of the deck expected fromone line of wheels of an HS 20-44 truck positioned for
Figure 18—Shear slip as a function of loadfor individual joint between chord and web inVierendeel truss.
Figure 19—Load resistance of three trusses, withand without butt joints.
maximum longitudinal moment. Thus, the actual de-flection under an HS 20-44 truck wheel line would beapproximately 1.15 in.
Figure 22 shows the variation in stiffness of the bridgewith anchorage configuration 1 as a result of changesin the “uniform” level of prestress in the deck. At aprestress <20 lb/in2, serious reduction in stiffness oc-curred, similar to that experienced with solid stress-laminated decks (Oliva and others 1990). A visual
13
Figure 20—Vertical displacements along trans-verse cross-section near center of deck result-ing from centerpoint load (anchorage configura-tion 1).
Figure 21—Vertical displacements along trans-verse cross-section near center of deck resultingfrom two centerpoint loads (anchorage configu-ration 1).
Figure 22—Deck vertical load-resisting stiffnesswith change in prestress level for anchorage con-figuration 1 with centerpoint load. (Stiffnessis centerpoint load (× 103 lb/in) per inch ofdeflection at center.)
inspection of the deck after loading at a prestress levelof 10 lb/in2 did not indicate any significant slip be-tween the laminae. Thus, we surmise that the loss instiffness at stress levels <20 lb/in2, obvious in Fig-ure 22, was a result of interlaminar opening betweentrusses at the bottom of the deck caused by the trans-verse bending moments, as noted in the solid deck re-search. An attempt to investigate the system underprestress >70 lb/in2 was unsuccessful. The posts actingas the bulkhead system deformed extensively and, inmany cases, failed under prestressing forces necessaryto achieve average compression >70 lb/in2 in the tim-ber bridge. Post failure occurred in one of two fashions:(1) the post began bulging in the direction perpendicu-lar to the compressive load and finally split along a ver-tical plane containing the stressing rod or (2) the postcompressed more on one side of the rod than the otherand post crushing was initiated. These failures were in-duced by the varying stiffness of the wood dependingon the ring structure of the solid-sawn posts. Becausehigh initial levels of prestress are necessary in actualfield installations, glulam posts are recommended tomore effectively transfer the compressive stress into thedeck.
Simulated truck loading tests were conducted under allanchorage configurations, as described previously. Re-sults for the different systems are plotted together inFigure 23. Note that at a prestress level in the range of10 to 70 lb/in2, the structural performance of the sys-tem, as measured by stiffness, was approximately thesame regardless of the anchorage configuration. Thus,all the anchorage systems are equally adequate froma performance viewpoint, and selection of a particu-lar type of detail should be based on economy and easeof construction. Further study of Figure 23 indicatesthat at a low prestress level of 20 lb/in2, configuration4, which had four rods in each web opening, exhibitedslightly higher stiffness than the other configurations.This was expected because all rods are positioned closeto the top and bottom chords of the deck in configura-tion 4, thus enhancing the transverse stiffness capacityof the deck.
Panel Joints
Figure 24 shows plots of transverse displacements ata section near the center of the deck, where the loadwas applied, for various prestress levels across the joint(6, 14, and 20 lb/in2) and in the deck (20, 50, and70 lb/in2). The measured stiffness of the deck (loadat midspan/displacement at midspan) was 14.8, 17.7,and 18.5 × 103 lb/in., respectively. At a low level ofprestress (20 lb/in2 uniform prestress in deck, 6 lb/in2
across joint), slip occurred at the joint, as indicatedby the discontinuous displacement profile. At higherprestress levels (70 lb/in2 uniform prestress in deck,20 lb/in2 across joint), the displacement profile wasnearly continuous, indicating the existence of ade-quate moment transfer across the joint. The stiffnessat 20 lb/in2 across joint and 70 lb/in2 in the panels
14
Figure 23—Deck vertical load-resisting stiffnesswith change in prestress level for all anchorageconfigurations.
Figure 24—Vertical displacements along trans-verse cross-section near center of deck, showingthe effect of joint between panels at differentprestress levels.
(18.5 × 103 lb/ in) was near the level shown in Figure23 for 20 lb/in2. It is obvious from Figure 24 that thelow prestress across the joint did not seriously affectthe shear capacity but did offset the transverse flexuralstiffness (slope discontinuity).
Prestress Distribution
Figure 25 shows how the transverse compressive de-formations, normalized to units of average strain onthe deck surface, varied in the vicinity of an anchor-age. Each curve in the figure shows deformation at alocation successfully further from the edge of the deck;location 1 was within 1.5 ft of the edge and location 3was >3 ft from the edge. A close examination of thetransverse deformation plots suggests the following:
1. Deformation is higher in the immediate region of theapplied load. This occurs regardless of rod arrange-ment (one or two rods) or bulkhead system and indi-cates that the prestress is transferred over a narrowwidth near the outside of the deck.
Figure 25—Transverse compression of deck withincrease in stress for three anchorage configu-rations. Adjacent rods located at 0 and 96 in.from deck edge. Location 1—compression innext 18 in. nearest deck edge. Location 2-compression in next 18-in. strip from edge.Location 3—compression in strip 3 ft from edge.Configuration 1—at web with two rods. Config-uration 1—at opening with one rod. Configura-tion 3—at web with one rod.
2. At some distance from the edge (that is, near lo-cation 3), the deformation approaches a “uniform”state.
3. The two deformation plots representing displace-ments at webs are qualitatively identical but differfrom the plot that represents deformation data fromthe web opening location. The reason for the higherconcentrated displacement at location 1 for the webopening configuation is reasonable, because all pre-stress is being transferred into the two chords, whilein the web configuration, there is variation over thedepth of the truss.
15
Table 5—Stiffness of deck during cyclic loading
Loss in stiffnesscompared to
Stiffness a initial stiffnessTime of test (× 103 lb/in) (percent)
Initial 18.7 —
After 100,000 cycles 18.1 -3.2After 400,000 cycles 17.7 -5.3After 500,000 cycles 17.2 - 8
a Stiffness measured as centerpoint loaddivided by center deflection (×103 lb/in).
Cyclic Loading
Table 5 lists the initial stiffness value of the two-panelbridge deck along with stiffness values measured laterin the cyclic loading test sequence. After one-half mil-lion cycles, stiffness was reduced a total of approxi-mately 8 percent. This loss of stiffness was not totallydue to deterioration caused by the repetitive action ofthe applied load. The duration of the cyclic test wasapproximately 1 week, including some idle time whilethe test sequence was changed and time spent conduct-ing the various stiffness tests. Relaxation tests con-ducted on solid stressed decks indicated that the rodslose a portion of their tension load primarily as a re-sult of creep in the wood (Oliva and others 1990). Atthe end of the dynamic tests, the average stress lossin all rods recorded was 35 percent (deck was initiallystressed to 50 lb/in2). The compressive stress betweenpanels was reduced from 14 to 10 lb/in2. We can log-ically assume that this accounted for part, if not all,of the stiffness reduction. A permanent slip did de-velop at the joint between panels. The slip was con-fined to within a distance of 2 ft, in the direction of thespan, on either side of the applied load. Maximum slipmeasured was 0.188 in. at the joint and was undoubt-edly a result of the unacceptably low level of prestressacross the joint. The measured permanent set of thedeck across a transverse section directly under the load,after one-half million cycles, is displayed in Figure 26.
Analytical and Experimental Results
The general analytic ability of the equivalent platemodel was verified by comparison with measured re-sults of simulated truck loading in the laboratory testsover various spans and with various load locations.Analyses were conducted for three different spans. Cor-relation plots, between measured and predicted dis-placements, are shown in Figure 27 for the deck with50 lb/in2 of prestress. Each plot shows the deflections
16
Figure 26—Vertical residual displacements acrosstransverse profile near center of deck after dy-namic load testing with one-half million loadcycles. Displacement offset occurred at paneljoint, where very little prestress existed.
that occurred across a transverse cross-section at thebridge midspan. Considering the complexity of the lab-oratory test model and the assumed relation betweenthe plate rigidities, the correlation between theory andexperiment is adequate; the maximum difference wasless than 10 percent.
Conclusions and Recommendations
The stiffness of the parallel-chord bridges described inthis report was improved compared to that of previ-ously studied stress-laminated solid-sawn timber bridgedecks with spans >35 ft. The parallel-chord system ap-pears to be a feasible means of improving the overallperformance and efficiency of timber bridges. However,the cost for the superstructure of the prototype bridgemust be considered ($33.45/ft2 for materials only; seeAppendix).
Resistance to Truck Loading
In this study, the parallel-chord stress-laminatedbridges could provide the desired stiffness; the stiff-ness can easily be modified by increasing the overalldepth of the trusses. Stress laminating allowed thebridge system to effectively transfer applied loads toa wide portion of the deck trusses. This lateral transferof load depends on the depth of the trusses and resultsin a wide effective distribution width or effective re-sisting width of the bridge deck if the deck is assumedto act as a simple beam. In addition to developing awide transverse load distribution, the system exhibitedno measurable slip or rotation between laminae whensufficient prestress was maintained.
Prestress Rod Anchorage System
Figure 27—Vertical deflections across transverseprofile near center span: experimental and ana-lytical results for various spans and load points.
Ease of Construction
The parallel-chord system lends itself to easier prefabri-cation of panels and connection at the bridge site com-pared to solid-sawn stress-laminated panelized decks.Prefabricated laminated panels may be erected in placeand subsequently stress-laminated together to form acomplete bridge by placing rods through the openingsin the truss webs and stressing.
Variations in configuration of anchorage systems forprestressing rods in the deck have little effect on loadresisting behavior as long as the anchorages are reason-ably designed to transfer prestressing force into boththe chords and webs. Anchorage plates must be de-signed to avoid crushing of wood under the plates andposts. If posts are used as part of the anchorage sys-tem, we recommend that they be made of glued lami-nated timber or laminated veneer lumber for improvedresistance to splitting under load.
Repeated Load Resistance
Over one-half million cycles of loading, at a level induc-ing deflections caused by an HS 20-44 truck, failed tocreate any significant deterioration of the bridge stiff-ness. The joint between panels is likely to be the oneregion of the bridge system that may be susceptible todamage by repeated loading if the prestress across thejoint is not maintained at a sufficiently high level.
Composite Action of Parallel-Chord Truss
The efficiency of any composite section depends on thetransfer of forces between materials in the cross sec-tion. The parallel-chord Vierendeel trusses used in thisstudy relied upon the action of steel dowels to transfershear forces between the chords and webs. Some defor-mation developed in the dowels, which allowed a smallslip between the chords and web elements. The trussesalso had higher vertical shear deformation than nor-mally occurs in a truss with continuous diagonal webelements because of vertical deformation in the chordsbetween the discrete web blocks. Tests results indicatedthat the effective cross-section moment of inertia was40 percent of the moment of inertia that would occur ifthe chords were perfectly connected and shear deforma-tion was neglected. This reduction in stiffness was dueto horizontal shear-slip in the doweled web connectionand vertical shear deformation at openings betweenthe web blocks. The relative size of these two defor-mation mechanisms depends on dowel size, number ofhorizontal shear interfaces in chord-to-web connection,and independent moment of inertia of chord members.The size of deformation mechanisms may also be sub-sequently affected by moisture content variations in thewood.
Analytical Modeling
Analytical modeling of a parallel-chord bridge systemwould be unacceptably complex if approached in anexact manner. The use of an “equivalent” orthotropicplate analytic model provided predictions of the bridge
17
deflection response that were within 10 percent ofactual measured values.
Future Use of Parallel-ChordStress-Laminated Bridges
The parallel-chord bridge has been shown to be anattractive alternative to other methods of bridge con-struction. Nevertheless, this type of bridge is in theexperimental stage. Before this system can be usedfor production bridges, the following issues must bedeveloped or clarified:
1. The stress-laminating technique is completely depen-dent on maintaining minimum levels of compressiveprestress between the laminae. The loss of prestressresulting from creep of the wood, relaxation of thesteel, and volumetric changes in the wood caused bymoisture variation must be thoroughly understoodbefore this method can be used on any applicationother than experimental bridges.
2. Methods must be developed for designing individ-ual trusses for the parallel-chord bridge. The de-sign process depends on the shear stiffness of theweb-chord connection. Values for shear stiffness ofdoweled or metal plate connections are not currentlywell-documented or readily available. Improvedparallel-chord systems may be developed that donot rely upon any mechanical fastenings betweenthe members in any individual truss, resulting in asimplification of the design process.
3. The ultimate capacities of individual trusses, at afailure limit state, as well as the capacities of stress-laminated subassemblies need to be documentedto provide a design engineer with a measure of theactual safety factor against failure in the system.Load sharing will probably reduce system strengthvariability and may allow design with an increasedcapacity over that expected using normal woodstress design tables.
4. Proven techniques for protecting prestressing rods orstrands are necessary. Steel is highly susceptible tocorrosion when it is in a state of elevated stress. Theprestressing elements are more exposed in timberbridges than in prestressed concrete bridges, wheregrouting or greased sheathing is commonly used inaddition to embedment in concrete.
References
Bakht, B.; Jaeger, L.G. 1985. Bridge analysis simpli-fied. New York: McGraw-Hill.
Huffington, N. 1956. Theoretical determination ofrigidity properties of orthogonal stiffened plates. Trans.American Society of Mechanical Engineers. Journal ofApplied Mechanics. 78(3): 15-20.
McCutcheon, W.J. 1992. The Mormon Creek Bridge:Performance after three years. Res. Pap. FPL-RP-509. Madison, WI: U.S. Department of Agriculture,Forest Service, Forest Products Laboratory.
Oliva, M.G.; Dimakis, A.G. 1988. Behavior of a stress-laminated timber highway bridge. American Societyof Civil Engineers. Journal of the Structural Division.114(7): 1850-1869.
Oliva, M.G.; Lyang, J. 1987. Behavior of parallelchord trusses in prestressed wood bridges. Structuresand Materials Test Lab. Report 87-3. University ofWisconsin, Madison.
Oliva, M.G.; Tuomi, R.L.; Dimakis, A.G. 1986. Newideas for timber bridges. Timber Bridges, Transporta-tion Research Record 1053, National Research Council,Washington, DC, pp. 59-64.
Oliva, Michael G.; Dimakis, Al G.; Ritter, MichaelA.; Tuomi, Roger L. 1990. Stress-laminated woodbridge decks: Experimental and analytical evaluations.Res. Pap. FPL-RP-495. Madison, WI: U.S. Depart-ment of Agriculture, Forest Service, Forest ProductsLaboratory. 24 p.
18
Appendix
Materials and costs for superstructure of Mormon Creek Bridge.a
Bridgecomponent Itemb
Number of costpieces Quantity ($)
Trusses
Curbs andscupperblocks
Miscellaneoussupplies
Anchorages
Superstructure c
4-in. x 6-in. x 8-ft Douglas Fir 50 800 fbm4-in. x 6-in. x 16-ft Douglas Fir 50 1,600 fbm4-in. x 6-in. x 24-ft Douglas Fir 50 2,400 fbm4-in. x 6-in. x 32-ft Douglas Fir 50 3,200 fbm4-in. x 12-in. x 3-ft Douglas Fir 100 1,600 fbm4-in. x 12-in. x 4-ft Douglas Fir 200 3,200 fbm
Total
4-in. x 6-in. x 10-ft Douglas Fir 16 320 fbm6-in. x 10-in. x 22-ft Douglas Fir 4 440 fbm6-in. x 10-in. x 5-ft Douglas Fir 4 100 fbm6-in. x 10-in. x 4-ft Douglas Fir 6 120 fbm
Total
10.75-in. x 12-in. x 2-ft posts0.75- x 34-in. dowels0.75- x 16-in. dome-head bolts0.75- x 16-in. lag screws0.75-in. plate washers0.75in. cut washers0.38 x 11 in.-ring-shank spikes60d (6-in.) ring-shank spikes0.6- x 16-in. drift pins
22
188
188
—
1,602
30.8 lb52.0 lb16.3 lb15.3 lb1.2 lb
20.0 lb9.0 lb
2,450.0 lbTotal
1-in.-diameter prestress rod1.25-in.-diameter prestress rod1-in.-diameter nuts1.25-in.-diameter nutsPlates for l-in. rod anchorPlates for 1.25-in. rod anchorBearing platesBearing platesBearing platesBearing platesBearing plates
TotalTotal
—34123412
6666
10
256 ft78 ft—
———————
—
8801,7602,6403,5201,3203,520
13,640
352484110132
1,078
57846782423
220
93,6774,457
397179177
85269139334174167131182
2,23421,409
a See Table 1 in text for metric conversion factors.b Width and thickness of wood are nominal dimensions (inches).Wood was treated with creosote.
c Total cost = $33.45/ft2.
19
