CONTENIDO DEL PLAN DE ZONA

Often initially in bridge design the entire structure is assumed to be constructed in a minimum of steps, with no regard to sequence. Girders are modeled complete end-to-end of the bridge. Wet deck concrete weight is applied abutment to abutment in a single step as well. Then superimposed loads are placed on the full composite structure. This is not how bridges are actually constructed.

There are many different ways that a bridge can be constructed, each with a unique loading sequence. It may be the case that the most critical load case occurs at some point during construction, for example stability of a composite girder bridge prior to the concrete deck hardening. For this reason a construction sequence analysis should be performed to make sure the bridge will not exceed any limit state at any point during construction, as well as reaching the desired stress state in the final condition. The construction analysis required depends on

the contract language and delivery type. Often only the proposed construction sequence on the contract plans needs be checked sufficiently to demonstrate that the bridge is constructible. Contractors wishing to deviate from the proposed sequence would provide their own analysis demonstrating adequacy.

Although permanent bridge dead loads are well defined for a constructability analysis, temporary construction loadings are not well defined by AASHTO LRFD, making constructability checks more difficult. One of the reasons is that construction loadings are so variable and bridge specific, depending not only on the type of bridge, but also the method of erection. The designer should make every effort to ensure that the loadings of the constructability analysis are consistent with the assumed method of construction.

Historically, in order to perform a constructability check of a relatively complex bridge, a “deconstruction” analysis was performed. The bridge was modeled in its final condition and then a series of analyses were run removing single or groups of elements. Temporary supports or bracing could also be added and/or removed in the models as necessary. This process not only permitted a check of the stresses at each step of construction, but also allowed for calculation of elevations and camber such that the bridge ended up in the correct final position, since that is where it “started”.

Another common series of analyses are the deck pour sequence. Typically the deck is not poured all at once. Due to the large volume of concrete required, it is cast in stages. Therefore the first pour is partially hardened and composite for loads from the second pour. The first and second pours are partially hardened and composite for loads from the third pour. And so on until the entire deck is cast. The modulus of the deck concrete can be adjusted for each pour in each analysis if desired to reflect the time of curing. Again, not only can stresses be checked at each stage, but by summing the deflections of the entire sequence, required camber due to the deck weight can be calculated.

A similar sequential redecking analysis may also be prudent during the initial design of a bridge. During a redecking, keeping a portion of the bridge open to traffic may be necessary/desireable. A sequence that permits a staged redecking of a bridge while maintaining a given level of both

at some point is considered likely.

Cantilever construction also requires a series of analyses. Temporary fixity must be applied at the pier, and an analysis run for each segment of bridge added to the structure. The pier needs to be checked at each stage to ensure that the load does not become too unbalanced. Displacements determined at each analysis stage can be used to ensure that the final condition matches the desired profile grade.

Bridge girders are sometimes initially simply supported for dead loads, and then made continuous for live loads. Multiple models reflecting the correct connectivity for the loading would be required. Deflections from the simply supported models would be used for determining camber.

Many current software packages have simplified the construction sequence/staged loading process. No longer do separate models need to be analyzed and processed individually. Solving sequentially and “activating”/”deactivating” elements in various load steps, a single model can be utilized for the entire sequence.

For slab-on-girder bridges, the contract plans require camber diagrams or tables for each girder, so that the desired profile grade is achieved under dead loads. In order to calculate the camber, it is necessary to correctly model all sources of stiffness such that deflections are determined with sufficient accuracy. Note there is no such thing as a “conservative” calculation of displacements, since achieving the profile grade within tolerance is the goal. This is particularly important in curved or skewed bridges since transverse load distribution and torsional behavior are significant contributors. Such model(s) may or may not be the same model(s) used for performing strength design.

During steel erection, fit-up of the cross-frames can affect not only the final geometry of the bridge, but can contribute to locked-in stresses as well. There are three possible fit-up

conditions: No load fit (NLF) where the girder webs are vertical and cross-frames are unstressed in the unloaded geometry, steel dead load fit (SDLF) where the girder webs are vertical and the cross-frames are unstressed under steel only load, or total dead load fit (TDLF), where the webs are vertical and the cross-frames unstressed under the full dead load of the bridge.

NLF is the easiest to design and often to construct, but can result in out-of-plumb girders and locked-in forces in curved and skewed bridges. SDLF and TDLF can diminish or eliminate the out-of-plumbness and locked-in forces, but may require the cross-frames to be force-fit initially, in order that they are unstressed at the appropriate displacement. Per AASHTO, the designer is responsible for the fit-up method specified for construction. In some cases a more refined analysis may be warranted, for instance on curved or highly skewed bridges. If a refined analysis of the bridge has been done, modeling the installation of the cross-frames for the desired fit condition is often the simplest and most accurate method of obtaining cross-frame forces. For more information on cross-frame fit-up see G 13.1 Guidelines for Steel Girder

References

AASHTO/NSBA Steel Bridge Collaboration, G 13.1 Guidelines for Steel Girder Bridge Analysis, 2nd _{Edition, 2014.}

Chavel, B., Coletti, D., Frank, K., Grubb, M., McEleney, B., Medlock, R., and White, D., Skewed

and Curved Steel I-Girder Bridge Fit, Fit Task Force, NSBA Subcommittee, August 20, 2014

Coletti, D., Chavel, B., and Gatti, W.J., The Challenges of Skew, TRB 2011 Annual Meeting. FHWA Steel Bridge Design Handbook, Bracing System Design, Publication No. FHWA-IF-12- 052 – Vol. 13, November 2012.

Grubb, Michael A. et. al., Analysis and Design of Skewed and Curved Steel Bridges with LRFD

– Reference Manual, NHI Course No. 130095, Publication No. FHWA-NHI-10-087, February

2011.

Hambly, Edmund C., Bridge Deck Behavior, John Wiley and Sons, Inc., New York, NY, 1976. Krupicka, G., and Poellot, W.N., Nuisance Stiffness, HDR Bridgeline, Vol. 4, No. 1, HDR Inc., Omaha, NE, 1993.

Nutt et. al., “Development of Design Specifications and Commentary for Horizontally Curved Concrete Box-Girder Bridges”, NCHRP Report 620, Transportation Research Board, 2008. O’brien, Eugene K., and Keogh, Damien L., Bridge Deck Analysis, Routledge, New York, NY, 1999.

Wang, W.H., Battistini, A.D., Helwig, T.A., Engelhardt, M.D. and Frank, K.H. (2012), “Cross Frame Stiffness Study by Using Full Size Laboratory Test and Computer Models,” Proceedings of the Annual Stability Conference, Structural Stability Research Council, Grapevine, TX, April 18-21, 11 pp.

White, D. W. et. al., “Guidelines for Analysis Methods and Construction Engineering of Curved and Skewed Steel Girder Bridges”, NCHRP Report 725, Transportation Research Board, 2012. Wong, Wai Tak, Basic Refined Analysis Workshop, Augusta Maine, July 8, 2010.

DRAFT

Section 4 – Verification