FIT-UP This paper summarizes research

supporting the development of

CONSIDERATIONS improved design, detailing and
FOR STEEL I- erection guidelines to ensure
GIRDER BRIDGES reliable fit-up of skewed and/or
curved steel I-girder bridges.
Domenic Coletti, PE, is a Twenty-one bridges, including
Principal Professional Associate multiple framing arrangements
with the Raleigh, NC office of on a number of the bridges, were
HDR. He received his BSCE analyzed. The quantitative data
from Carnegie Mellon of this research support
University and his MSCE from recommended fit conditions as a
the University of Texas at function of the bridge geometry.
Arlington. He is a member of Forces required to assemble the
several Task Groups of the steel during erection were
AASHTO/NSBA Steel Bridge evaluated and difficult cases
Collaboration and is Chair of highlighted. Suggested erection
TRB’s Standing Committee on considerations to facilitate fit-up
Steel Bridges, AFF20. He was a were provided. In addition, the
member of the NCHRP 20-07, research investigated and
DOMENIC A. COLETTI Task 355 Research Team. specified beneficial staggered
Don White is a Professor at the cross-frame arrangements for
School of Civil and straight skewed bridges, as well
Environmental Engineering as framing arrangements around
(CEE). He has been a member of bearing lines at interior piers in
the CEE faculty at Georgia Tech continuous-span bridges. The
since 1997. Dr. White is a research placed an emphasis on
member of the AISC Technical identifying the impacts of the
Committees 4, Member Design, chosen fit conditions on girder
and 3, Loads, Analysis and elevations, girder layovers,
Stability, the AISI Bridge cross-frame forces, girder
Design Advisory Group, the stresses, and vertical reactions in
AISC Specification Committee, completed bridge systems.
and several AASHTO/NSBA Simplified methods of
Steel Bridge Collaboration Task accounting for Steel Dead Load
Groups. He is past Chair of the Fit (SDLF) and Total Dead Load
DONALD W. WHITE SSRC Task Group 29, Second- Fit (TDLF) detailing effects
Order Inelastic Analysis of were provided. In addition,
Frames and currently serves on procedures were developed and
the Executive Committee of the explained for direct calculation
SSRC. He was the Principal of the locked-in forces due to
Investigator for NCHRP SDLF and TDLF detailing in
Research Project 20-07, Task cases where a more precise
355. calculation of these effects may
be beneficial. Lastly,
Thanh Nguyen, PhD, PE is a construction inspection best
Bridge Engineer with the Fort practices were recommended to
Lauderdale, FL office of RS&H. ensure that the erected geometry
He received his BSCE, MSCE, sufficiently meets the specified
and PhD degrees from Georgia fit conditions, and recommended
Institute of Technology. He was design specification provisions
a member of the NCHRP 20-07, were developed that synthesize
THANH V. NGUYEN Task 355 Research Team. the key guidelines.
 FIT-UP
CONSIDERATIONS
FOR STEEL I-
GIRDER BRIDGES
Brandon Chavel, PhD, PE is a
Senior Professional Associate
and Bridge Section Manager
with the Cleveland, Ohio office
of HDR. He received his PhD
from the University of
Pittsburgh. He is a member of
several Task Groups of the
AASHTO/NSBA Steel Bridge
Collaboration and is Chair of TG
11, Steel Bridge Design
Handbook. He was a member of
the NCHRP 20-07, Task 355
Research Team.
BRANDON W. CHAVEL
Michael Grubb, PE is a self-
employed steel-bridge design
consultant with M.A. Grubb &
Associates, LLC in Wexford,
PA. He has 39 years of
experience in steel-bridge
design, steel-bridge design
specifications, straight and
curved steel-bridge research, the
development and delivery of
training courses on steel-bridge
design, and the development of
comprehensive steel-bridge
design examples. He was a
member of the NCHRP 20-07,
MICHAEL A. GRUBB Task 355 Research Team.
Calvin Boring, Jr. is the
Operations Manager at
Advantage Steel and
Construction located in
Saxonburg, PA. He received his
BSCE at the University of
Pittsburgh. He is a member of
Iron Workers Local # 3, ESWP
where he is on the Executive
Committee of the IBC and
served as General Chair in 2014,
and Board Member of ASHE –
SW PENN. He was a member of
the NCHRP 20-07, Task 355
CALVIN G. BORING, JR. Research Team.
 FIT-UP CONSIDERATIONS
FOR STEEL I-GIRDER BRIDGES
Background span lengths and bridge widths were examined,
along with both parallel and non-parallel support
Achieving reliable fit-up of steel girder bridges with conditions. Examples of framing plans for bridges
sharp curvature and/or significant skew is inherently studied are illustrated in Figure 1. In some cases, the
challenging. To help inform the steel bridge design configuration of the framing was varied for the same
and construction industry, an ad-hoc task group overall bridge geometry, typically to investigate the
affiliated with the National Steel Bridge Alliance effects of contiguous versus staggered cross-frame
(NSBA) published guidelines featuring fit-up patterns. The sequence of erection was considered in
considerations and design, detailing, and erection the analysis of each bridge, and in some cases more
recommendations (1). The basis for the than one erection scheme per bridge was evaluated
recommendations was largely qualitative, being to investigate the effects of the various erection
derived primarily from anecdotal accounts and the schemes on the difficulty of fit-up and the magnitude
professional experiences of the authors. Further of locked-in stresses. The difficulty of fit-up was
guidance, again largely qualitatively derived, was evaluated in terms of the magnitude of the “fit-up
also presented in Reference (2). While the guidance forces,” i.e., the forces required to physically bring
provided by these documents was helpful, it was together a cross-frame and a girder to which the
recognized that conclusive, authoritative cross-frame is being connected.
recommendations could only result from quantitative
research. To this end, research was funded as A summary of key findings and recommendations of
NCHRP Research Project 20-07, Task 355, this research is presented in this paper. For more
Guidelines for Reliable Fit-Up of Steel I-Girder detailed discussion, the reader is encouraged to
Bridges, and reported in Reference (3). consult Reference (3).

The research team first conducted a survey of Cross-Frame Fit

current industry practice with regard to cross-frame
The “fit” or “fit condition” of a skewed and/or
framing arrangements, fit conditions / cross-frame
curved I-girder bridge refers to the geometry in
detailing methods, erection procedures, and
which the cross-frames are detailed to attach to the
construction inspection practices; the results were
girders. The fit condition is selected for a given
synthesized and general trends were discussed. The
bridge to facilitate erection by offsetting, or
survey revealed a wide range of practices and a
compensating for (to different extents), the tendency
similarly broad range of understanding of the key
of the I-girders in these bridge types to twist due to
issues associated with fit-up of steel girder bridges.
differential deflections. The selected fit condition
Next, twenty-one steel I-girder bridges were corresponds to a specific targeted outcome of when
analyzed to investigate the effects of structural steel the girder webs will be approximately plumb
framing arrangements, specified fit conditions, (vertical) in the field. “Fit-up” refers to the
construction tolerances, and construction actions on assembly of the structural steel during the bridge
ease of fit-up and locked-in stresses. The bridges erection. It is desirable that the “fit-up” of the
investigated included radially supported curved structural steel should be manageable, without the
girder bridges, straight girder bridges with skewed need for excessive jacking or pulling forces from the
supports, and curved girder bridges with skewed erector. The “fit condition” and the “fit-up” of the
supports. Both single span and multiple-span structural steel are interrelated, but these terms refer
continuous bridges were investigated. A range of to different attributes of the construction.

Page 1 of 16
Table 1 summarizes the three most common fit The term “Total Dead Load,” typically is assumed to
conditions considered in skewed and/or curved I- include either all dead loads that are present when
girder bridges. Alternate names for each potential fit the bridge is opened to traffic, or the as-constructed
condition, which are generally more familiar to dead loads, taken as the weight of the structural steel
fabricators and steel detailers, are also provided in plus the weight of the concrete deck, but not
the table; the names are used interchangeably in including the weight of barrier rails, sidewalks, etc.
practice.

TABLE 1 Common Fit Conditions

Condition Alternate Name Description
No-Load Fit (NLF) Fully-Cambered Fit The cross-frames are detailed to fit to the girders in their
fabricated, fully-cambered and plumb position under zero
dead load.
Steel Dead Load Fit (SDLF) Erected Fit The cross-frames are detailed to fit to the girders in their
ideally plumb as-deflected positions under bridge steel
dead load at the completion of the erection.

Total Dead Load Fit (TDLF) Final Fit The cross-frames are detailed to fit to the girders in their
ideally plumb as-deflected positions under the bridge total
dead load.

FIGURE 1 Framing plans for representative bridges studied in this research.

Page 2 of 16
Forces Required to Assemble the and TDLF) were analyzed. The fit-up forces were
recorded and the ease of assembly was characterized
Steel During Erection as a function of the magnitude of the fit-up forces.
A major focus of the NCHRP 20-07, Task 355 Erectors commonly use come-alongs and other local
research was the ease of fit-up of the cross-frames equipment, as necessary, to make the connections
during erection. In this work, cross-frame fit-up was between the cross-frames and the girders. A typical
estimated by calculating the forces induced at the come-along capacity was taken as 20 kips (some
cross-frame top and bottom connections, for the erectors indicate that 12 kips is more typical). For
second girder to which the cross-frame is connected, the purposes of evaluating various erection schemes
as the cross-frame is installed. The fit-up force and fit-up conditions, a calculated fit-up force
calculations performed in this research are accurate significantly more than 40 kips was considered
to the extent that the nominal assumptions generally “difficult.”
employed in bridge design are satisfied. That is, the A full presentation of the specific findings regarding
simulations to determine fit-up forces are based on fit-up forces for each of the studied bridges is
the following assumptions: beyond the scope of this summary paper; the reader
• No yielding of the steel, is directed to Reference (3) for presentation of all
• No incidental restraint from friction, etc. at results. A sample of the results is presented in Table
temporary or permanent supports, 2. Note the focus on correlation of difficulty of
• The girder geometries, support elevations, erection versus specific bridge geometries (e.g., span
etc. are as specified in the bridge plans, and length, curvature, bridge width, erection scheme,
etc.) and specific behavior (i.e., magnitude of
• Negligible “play” in the connections.
differential deflections).
There are various factors that can influence the
actual bridge erection but cannot be accounted for in Suggested Erection Considerations
any detailed way within a practical engineering In addition to choosing an appropriate fit condition,
erection analysis, such as: determining an effective erection scheme is critical
• Tolerances and the associated play at bolted to ensure that a curved and/or skewed bridge is
connections, constructible and the maximum fit-up forces are
• Adjustments of the crane and support maintained in a reasonable range. In some cases, site
elevations by the erector, constraints such as a waterway, or availability,
• Tolerances on support elevations, and capacity, and allowed erection duration and location
• Changes in the geometry of the steel due to of cranes and shoring towers, can dictate the erection
thermal movements, etc. schemes.

These factors can cause differences between the A full discussion of erection considerations is
actual fit-up forces encountered in the field provided in Reference (3). A summary of selected
compared to the erection analysis estimates. key considerations is provided here.
Connection tolerances and adjustment of crane and Lifting Cranes, Hold Cranes, Shoring
temporary support elevations can indeed make the
Towers, Tie-Downs
fit-up forces somewhat smaller than the calculated
estimates, as discussed in more detail in Reference The lifting scheme for each girder (e.g., two-point
(3). However, the calculated fit-up forces pick, two-point pick with spreader beam, etc.)
determined in this research are forwarded as affects the orientation, deflection, and stresses in the
reasonable engineering estimates associated with the girder during lifting, and can thus affect fit-up (as
nominal design representation of the structures. well as affecting stability of the girder during
lifting). Hold cranes are often used during early
The fit-up forces required to assemble each bridge
stages of erection to reduce deflections and major-
were evaluated at various stages of erection. In some
axis bending moments and facilitate fit-up of girders
cases, more than one erection scheme was
and cross-frames, especially in curved girder
investigated. For each bridge (and each erection
bridges. Shoring towers are often needed in the
scheme), typically three fit conditions (NLF, SDLF,

Page 3 of 16
 TABLE 2 Sample Presentation of Summary of Maximum Cross-Frame Fit-Up Forces for Curved Radially-
Supported Bridges
Differential Cross-Frame Fit-Up
Shoring Ls wg R L s/ Deflections Force
Bridge ng Ls/R
Towers (ft) (ft) (ft) wg (in.) (kip)
SDL TDL NLF SDLF TDLF
(A)
0 90 17.5 200 3 0.45 5.1 0.42 1.67 3.3 7.4 22.3
EISCR1
(B)
NISCR2, 0 150 24.0 438 4 0.34 6.2 0.68 1.83 16.6 28.7 54.0
Scheme 1
(C)
0 150 74.0 280 9 0.54 2.0 0.42 1.19 21.3 35.9 75.3
NISCR7
(D)
1 225 74.0 705 9 0.32 3.0 0.47 0.78 18.6 20.4 21.8
NISCR10
3 322, ∞, 0, 8.0,
(E)
(in curved 417, 40.4 ∞, 4 0, 10.3, 3.10 5.41 37.5 86.3 130.0
EICCR11
span) 322 411 0.80 8.1
Notes:
1. Ls = Span length(s)
2. Wg = Bridge width, measured between exterior girders
3. R = Radius of curvature at centerline of bridge
4. ng = Number of girders in the cross-section
5. Ls/R = Subtended angle between bearing lines
6. Ls/wg = Length to width ratio
7. Color coding of fit-up forces: A typical come-along capacity is taken as 20 kips (some erectors indicate that 12
kips is more typical). Calculated maximum fit-up forces between 30 and 40 kips are shown by light (blue)
shading. Calculated maximum fit-up forces greater than 40 kips are considered difficult and are highlighted by
dark (red) shading.
8. Bridge case (E) EICCR11 involved drop-in segments.
9. NLF = No-Load Fit; SDL = Steel Dead Load; SDLF = Steel Dead Load Fit; TDL = Total Dead Load; TDLF =
Total Dead Load Fit

construction of long-span bridges and curved bridge cross-section, and for drop-in segments
bridges. Multiple field splices may be required installed in continuous spans.
within longer spans. Shoring towers help limit
deflections and facilitate the installation of field Erection Schemes for Curved, Radially-
splices and cross-frames. The shoring towers should Supported Bridges
be used across the full width of the bridge cross- For curved bridges, cranes and/or temporary
section where practicable to best facilitate erection supports are critical for stabilizing the partially
of the structural steel. The number of shoring towers completed systems, as well as for erecting the
and cranes is generally selected to provide for a girders and cross-frames. Individual curved girders
feasible, safe, and economical erection. Furthermore, and narrow partially-erected curved bridge units
tie-downs are typically provided for the girders at have little stability on their own. The bridge cross-
the shoring tower locations and/or the permanent section generally over-rotates until all of its girders
supports to ensure girder stability before and after are installed. For most of the curved radially-
the splices are made within the spans. supported bridges studied in this research, the
The critical stages for fit-up often are stages that bridges are erected from the outside to the inside of
have the highest differential deflections between the the curve. This is for the following reasons:
girders. High differential deflections are indicative • The girder on the inside of the curve on the
of the potential for development of large internal partially completed bridge cross-section
forces between the girders. Fit-up can potentially be deflects less than the outside girder.
the most difficult for the last girders installed in the

Page 4 of 16
 • The girder that is being installed is • For longer spans that require a field splice
supported by a lifting crane, and thus its within the span (because the field sections
deflections are typically small. otherwise become too heavy), and often may
• Erecting from the outside to the inside of the require shoring towers, it is best to install
curve requires smaller fit-up forces due to only a few cross-frames or struts before the
the smaller differential displacements field splice is made, and to install the
between the inside girder and the girder remaining cross-frames after the field splice
being installed. is completed.
• Erecting from the outside to the inside of the • If any temporary supports are still being
curve avoids the need to lift the outside employed when the cross-frames are being
girder on the partially completed bridge installed, positioning the temporary supports
cross-section to achieve fit-up with the next at the final girder steel dead load (SDL)
girder being installed on the outside of the elevations is often a good starting point to
curve, which is typically the case when the alleviate potential large fit-up forces.
bridge is erected from the inside to the • Typically, cranes are only used to lift the
outside of the curve. girders into place and are not critical to the
• For highly curved bridges, the crane and erection of straight skewed bridges
temporary support requirements for erection constructed in the above ways. This is in
from the inside to the outside of the curve contrast to the discussion of curved bridge
can be significantly greater than for erection cases above.
from the outside to the inside of the curve. • When the cross-frames are detailed for
SDLF, their installation using the above type
In many cases, when a bridge is highly curved, a
of erection scheme tends to result in the
holding crane will be required on the girder on the
lowest level of fit-up forces.
outside of the curve until a number of the girders in
the bridge cross-section have been installed. The For continuous-span straight skewed bridges, the
erection schemes employed in this research install erection schemes with the greatest ease of fit-up are
the bearing line cross-frames immediately after the typically similar to those for the simply-supported
girder is placed on its supports, to help provide bridges described above. However, it is impractical
torsional stability to the girder. Then the remaining for the erector to install each girder in all the spans,
intermediate cross-frames are sequentially installed. one at a time throughout the bridge length, to
achieve the girder SDL elevation profiles. Instead,
Erection Schemes for Straight, Skewed all the girders are typically erected in each span
Bridges before moving to the next span. In these bridge
The potential fit-up considerations for straight types, a good option is to:
skewed bridges are somewhat different than those • Install only a minimal number of cross-
discussed above for curved radially-supported frames to keep the bridge stable until all the
bridges. A number of considerations for straight girders are erected.
skewed simply-supported spans are as follows: • Once all the girders in all spans have been
• For short straight skewed simply-supported erected, install the remaining cross-frames
spans that do not require a field splice within span-by-span.
the span, and therefore would rarely require This scheme limits the crane movement along the
shoring towers, the cross-frames can be length of the bridge while keeping the bridge stable
installed sequentially from one abutment to and the SDLF fit-up forces relatively small. In
the other after each girder is lifted onto its addition, this procedure appears to provide the best
vertical supports. option to mitigate large fit-up forces in straight
• Tie downs can be provided at the supports as skewed bridges detailed for TDLF detailing.
necessary to maintain lateral-torsional However, for longer spans with sharp skews, the
stability of the girders. larger fit-up forces associated with TDLF can be
problematic in some cases.

Page 5 of 16
Erection Schemes for Curved and Skewed Detailed Evaluation of Straight
Bridges Skewed Bridge Responses
For the curved and skewed bridges studied in this Associated with the Use of LGA vs.
research, the holding crane, lifting crane and shoring
tower elevations were located at the no-load
3D FEA Camber
elevations. Fit-up forces in curved bridges can be It is common for girder camber profiles to be
reduced by varying the crane and shoring tower calculated from a 1D Line Girder Analysis (LGA)
elevations from the no-load elevations. However, it for some bridges, 2D Grid analysis for others, and in
was shown that the reduction in fit-up forces is some cases from a 3D Finite Element Analysis
relatively small. Also, iteratively adjusting the crane (FEA). For a highly skewed I-girder bridge,
and shoring tower elevations to minimize the fit-up however, the differences in the cambers obtained
forces is not practical in typical erection engineering from LGA versus the other two methods can be
practice. However, in some cases, it can be substantial. Nonetheless, while the camber profiles
beneficial for the erection personnel to install cross- calculated from LGA and 3D FEA for a straight
frames at positions where the deflected geometries sharply-skewed bridge can be substantially different,
are most compatible, and for the crane operator to the final bridge geometries and responses obtained
incrementally raise or lower a girder that is being with either SDLF or TDLF detailing are similar.
installed after successive insertions of cross-frames, The use of cambers from LGA gives the closest
in effect to “button up” the cross-frames between the match to the ideal zero girder layovers and flange
girder that is being installed and the structural steel lateral bending stresses under the targeted dead load
that is already in place. From the studies of the conditions while the use of 3D FEA cambers gives
erection schemes of several curved and skewed girder layovers and internal stresses that are small,
bridge cases, the following conclusions were drawn: but non-zero, compared to the overall dead load
responses under the targeted conditions. The final
• For continuous-span cases, leaving the
girder elevations due to TDLF detailing based on the
shoring towers in place during the erection
LGA cambers closely match with the ideal targeted
of subsequent spans helps to reduce the
girder elevations under total dead load (TDL).
overall deflections, which can facilitate fit-
However, the final girder elevations due to TDLF
up.
based on the 3D FEA cambers deviate only slightly
• Similar to the recommended practice for
from the ideal targeted elevations under TDL.
curved radially-supported bridges, the
Based on the studies synthesized by the research
erection scheme for curved and skewed
team, it was concluded that the 3D FEA results are
bridges should also be from the outside to
close enough to matching the ideal values such that
inside on tightly curved bridges, whenever
it is sufficient to use 3D FEA (or other accurate
practicable, to reduce the maximum fit-up
refined analysis) cambers for detailing of straight
forces.
skewed bridges. For a full discussion of this issue,
• The cross-frames ideally should be installed
along with a comprehensive presentation of
sequentially from the radial bearing line (if analytical studies of a single span, straight, wide
there is a radial bearing line) to the skewed (nine girder lines), severely skewed bridge, please
bearing line. This reduces the deflection see Reference (3).
incompatibilities when installing the cross-
frames near the skewed end of the span. Influence of Framing Arrangements
The cross-frame framing arrangement can have a
significant effect on the overall bridge behavior as
well as the fit-up forces during the steel erection. In
a number of the bridges studied in this research,
specific improvements in the cross-frame framing
arrangements were investigated. These
improvements relate particularly to the alleviation of

Page 6 of 16
significant nuisance transverse stiffness paths slightly larger than 30 ft can be accommodated
associated with skew. These recommended easily in many cases. But at interior pier supports in
improvements are summarized below. multiple-span continuous bridges, where large
negative moments occur, the use of cross-frame
Offsets between Intermediate Cross- spacings larger than 30 ft at acute corners would
Frames and Skewed Supports adversely impact the lateral torsional buckling
References (2) and (4), recommend the use of an capacity of the fascia girders. To address torsional
offset of the intermediate cross-frames from the rotations due to overhang loads and provide lateral
skewed bearing line cross-frames that is the larger of torsional buckling resistance, the first cross-frame in
1.5D or 0.4 Lb wherever practicable, where D is the the exterior bays adjacent to the skewed bearing
girder web depth and Lb is the next or adjacent lines can be framed perpendicular to the girders with
interior unbraced length. The provision of this offset a small offset from the bearing on the interior girder
locates cross-frames where girder differential and then the diagonal members of this cross-frame
displacements between the cross-frame ends are can be removed to reduce the resulting nuisance
significantly reduced, leading to lower cross-frame transverse stiffness, as shown in Figure 2. The cross-
forces. frames highlighted by an oval and labeled on this
plan view as “CO” (for “chords only”) do not
Upon applying these rules to the suite of bridges contain any diagonals. This allows for a small offset
selected for the NCHRP 20-07, Task 355 research, it of these cross-frames relative to the skewed bearing
became apparent that the above 1.5D rule was overly lines without inducing large cross-frame forces from
punitive and difficult to implement in longer-span nuisance transverse stiffness effects, while reducing
highly-skewed bridges. This is because 1.5D is the large unbraced length on the adjacent girder at
commonly a larger fraction of the other unbraced the acute corner of the bridge plan. This scheme may
lengths for longer-span bridges, where the typical be considered as a variant of the lean-on bracing
unbraced lengths of 30 ft or less are a smaller concept proposed by Romage (6) and Zhou (7).
fraction of the overall span length. As such, the
unbraced length on the fascia girders at the acute Cross-Frame Framing and Detailing
corners of the spans tended to be too long. The Considerations for Severely Skewed
research team found that a length of 4bf, where bf is Bridges
the largest girder flange width within the unbraced
lengths on either side of the first cross-frame, serves It is common practice to allow skewed intermediate
as a better minimum limit that should always be met cross-frames where the support lines are skewed by
to ensure that offsets (and stagger distances) actually less than or equal to 20 degrees from normal.
serve their intended purpose. However, where the support lines are skewed more
than 20 degrees from normal, the AASHTO LRFD
For bridges with sharply skewed bearing lines, the Bridge Design Specifications (4) require that the
maximum (4bf , 0.4Lb) offset rule may still result in a cross-frames be framed orthogonal to the girders. In
large Lb on the fascia girder near the acute corners of this case, it may be advantageous to place the
sharply skewed spans. The older AASHTO Standard intermediate cross-frames oriented normal to the
Specifications for Highway Bridges (5) formerly girders in discontinuous lines, to selectively remove
recommended a maximum unbraced length of 25 ft. certain cross-frames, and/or to stagger the cross-
This has been replaced in the more recent AASHTO frames in adjacent bays between the girders, in such
LRFD Bridge Design Specifications (4) by the a manner that the transverse stiffness of the bridge is
requirement for a rational analysis to assess the reduced. Removal of highly stressed cross-frames,
cross-frame spacing, but cross-frame spacings larger particularly in the vicinity of the obtuse corners of a
than 30 ft are still relatively rare in straight I-girder span, reduces the stiffness of the corresponding
bridges, and are not permitted for curved I-girder transverse load path by forcing load transfer via
bridges. girder flange lateral bending.
At the simply-supported ends of a straight I-girder The above practices tend to decrease the cross-frame
bridge, if the overhang loads do not cause excessive forces and increase the girder flange lateral bending.
twisting of the fascia girder, unbraced lengths However, in certain cases involving excessively stiff

Page 7 of 16
transverse load paths, the cross-frame forces may be Effects of Fit Condition on Girder
decreased to the extent that the associated flange
lateral bending stresses are also reduced. The Stresses
unbraced lengths between the cross-frame locations In straight skewed bridges, the influence on the
must still satisfy the flange resistance requirements girder major-axis bending stresses due to SDLF and
of the design specifications. Where the flange sizes TDLF detailing based on refined analysis cambers is
are increased due to the additional flange lateral small and can be neglected, as long as the framing
bending, this increase typically is not significant. In plan is configured in accordance with the
fact, the increased cost resulting from the increased recommendations presented earlier in this paper.
flange sizes is often much less than the increased
cost of providing larger and/or more numerous Effects of Fit Condition on Cross-
cross-frames. Frame Forces
This research recommends framing of the cross- Although the use of refined analysis methods is not
frames within straight skewed spans using required for all curved and/or skewed I-girder
arrangements such as those shown in Figure 3 to bridges, these methods, when utilized, do allow for
both dramatically reduce the number of cross-frames direct consideration of cross-frame forces and girder
within the bridge as well as to reduce the overall flange lateral bending stresses. However, it is
transverse stiffness effects. important to recognize that the dead-load force
effects, when determined from a refined analysis

FIGURE 2 Use of intermediate cross-frames, with chords only, adjacent to skewed

bearing lines.

FIGURE 3 Recommended staggered cross-frame patterns for straight skewed bridges.

Page 8 of 16
model, typically do not include the locked-in force equilibrium with their major-axis bending moments,
effects from SDLF or TDLF detailing of the cross- and to restrain their tendency to twist. SDLF and
frames. That is, the analysis model corresponds to TDLF detailing tends to increase these internal
the assumption of NLF. cross-frame forces, since the cross-frames are used
to twist the girders back in the direction opposite to
In a straight skewed bridge, SDLF or TDLF
the direction that they naturally roll under the dead
detailing twists the girders in the direction opposite
loads; this action effectively increases the restraint
from that which they roll under dead load. However,
provided to the girders from the cross-frames.
in this case, the detailing relieves the dead load force
effects in the cross-frames. This is because the dead It is possible to directly calculate the internal
load twist rotations in a straight skewed bridge are “locked-in forces” associated with SDLF or TDLF
imposed on the girders via the compatibility of detailing directly within either a 2D grid or 3D
deformations with the cross-frames. Finite Element Analysis. The calculations simply
involve the consideration of the initial lack-of-fit
Conversely, in a curved radially-supported bridge,
displacements between the cross-frame connection
the intermediate cross-frames restrain or resist the
work points and the corresponding work points on
tendency of the girders to twist and deflect
the girders in the undeformed No-Load geometry of
excessively, which would occur if they were
the structure. These lack-of-fit displacements are
restrained from twisting only at the bearing lines.
then used to calculate initial strains in the cross-
The intermediate cross-frames tie the girders into the
frame members, or initial fixed-end forces in an
overall structural system, and force the girders to
overall beam element representation of the cross-
work together to resist torsion via differential major-
frames. These initial strains or initial fixed-end
axis bending of the girders across the bridge cross-
forces induce nodal loads in the structural analysis
section. Therefore, the additional pulling or twisting
model that account for the influence of the initial
of the girders in the opposite direction from that
lack-of-fit. The response of the structure to these
which they want to roll adds to the other dead load
nodal loads is added to the above “initial effects” in
cross-frame forces in a curved radially-supported
the undeformed configuration of the structure to
bridge, since the other dead load forces and the
determine the corresponding internal forces and
additional forces associated with the SDLF or TDLF
stresses that are “locked-in” to the structure due to
detailing are both restraining or resisting the
the dead-load fit detailing.
tendency of the individual girders to twist and
deflect excessively. Reference (3), provides a detailed explanation of the
above procedures, complete with benchmark 2D-
As a result, in straight skewed bridges, it is
grid and 3D FEA calculations for a basic straight
conservative to design the cross-frames using the
skewed as well as a curved radially-supported
results from an accurate grid or 3D FEA model and
bridge. It also explains how the results for the
neglecting the SDLF or TDLF effects. This is the
locked-in forces determined from this type of
current common practice when the engineer chooses
analysis may be included within design load
to utilize more than a line girder analysis for the
combinations to properly satisfy AASHTO LRFD
design. In certain I-girder bridges (those with severe
requirements.
skew and large width/span ratio) the cross-frame
forces determined in this manner can be very At the present time, inclusion of the lack-of-fit
conservative. This can lead to excessively large effects from SDLF or TDLF detailing is not well
cross-frames. In lieu of a refined analysis that supported in professional analysis and design
includes the lack-of-fit due to the SDLF or TDLF software. An engineer who wishes to include these
detailing, Reference (3) provides a range of simple effects typically must do significant calculations
reduction factors that may be applied to the cross- outside of the software, then input information such
frame forces and the flange lateral bending stresses as, for example, pseudo temperature changes in the
from a refined analysis that does not otherwise cross-frame members that produce the same initial
account for these effects. strains as the initial lack-of-fit displacements. Until
this situation is improved, and for sanity checking of
In curved girder bridges, the girders require radial
the results from these types of analysis calculations
forces to be introduced by the cross-frames to satisfy

Page 9 of 16
when they are performed, the basic estimates SDLF and TDLF detailing. This table is based on
recommended in Table 3 may be employed to the studies conducted in Reference (3).
estimate the locked-in force effects associated with

Table 3 Recommended estimates of factored dead load bridge responses for curved and/or skewed bridges in
their final constructed condition, in lieu of including lack-of-fit directly within the structural analysis.
Responses (1) Curved (2) Straight Skewed (3) Curved and Skewed
Radially-Supported
Cross-frame γp (2.0 SDL + ADL) for γp TDL for SDLF, Same as (1)
forces a
SDLF , except (γp – 0.4) TDL for TDLF
γp (SDL + ADL) for chords
of X-Type cross-frames
Flange lateral γp (1.2 SDL + ADL) (γp – 0.5) SDL + γp ADL for Same as (1)
bending for SDLF b SDLF
(γp – 0.4) TDL for TDLF
Major-axis γp TDL for SDLF a γp TDL for SDLF b Same as (1)
bending γ TDL for TDLF c
p
Vertical γp TDL for SDLF a
γp TDL for SDLF b, e For simply-supported bridges d, f:
Reactions γp TDL for TDLF c, e
For simply supported Worst-case maximum reactions g:
bridges, SDLF and TDLF For simply-supported bridges • γp (1.2 SDL + ADL) for SDLF a,
tend to increase the the tendency for uplift on the when the length of girder on the
smallest reactions at the girder bearings at the obtuse inside of the curve is increased
girders on the inside of the corners of the bridge plan is by the skew
curve d lessened by the use of SDLF • γp (1.6 SDL + ADL) for SDLF a,
or TDLF detailing based on when the length of girder on the
refined analysis cambers outside of the curve is increased
(compared to the use of LGA by the skew
cambers)
Definitions and Acronyms:
• SDL = Steel Dead Load, SDLF = Steel Dead Load Fit
• TDL = Total Dead Load, TDLF = Total Dead Load Fit
• ADL = Additional Dead Load = TDL – SDL
• LGA = Line Girder Analysis
• γp = Permanent Dead Load Factor
Notes:
a) TDLF detailing is strongly discouraged for curved bridges with Ls/R > 0.03 +, where Ls is the span length
along the centerline of the bridge and R is the radius of the centerline of the bridge cross-section.
b) Contingent on the use of discontinuous cross-frame lines with an unbraced length Lb > max of (4bf,
0.4Lb.adj) for all unbraced lengths within the span, where bf is the largest girder flange width within on
either side of a given cross-frame, and Lb.adj is the smallest adjacent unbraced length.
c) Contingent on Is < 1.0 +, and Lb > max of (4bf, 0.4Lb.adj), where Is is the “skew index” in Eq. 4.6.3.3.2-2
of Reference (4)
d) The influence of SDLF or TDLF detailing on the reactions for curved and skewed continuous-span
bridges is relatively complex; if potential uplift and/or increases in the reactions are a concern, a SDLF or
TDLF refined analysis is recommended.
e) If potential uplift at obtuse corners of the bridge plan is a concern, the uplift condition can be estimated
conservatively by using LGA for the targeted dead load condition and NLF refined analysis for additional
dead and/or live loads.
f) In curved and skewed I-girder bridges, the cross-frame lines need to be contiguous within the spans to
develop the width of the structural system; in some cases, this requirement can exacerbate potential uplift
conditions at obtuse corners of the bridge plan that are on the inside of the curve.
g) If potential uplift at obtuse corners of the bridge plan is a concern, a SDLF or TDLF refined analysis
should be considered.

Page 10 of 16
In curved I-girder bridges, the locked-in force effects Common Items
from SDLF and TDLF detailing tend to be additive
with the corresponding dead load effects. Here are a few items which are common to any
Calculation of adjustments to force effects is curved and/or skewed steel I-girder bridge,
recommended for curved, radially-supported bridges regardless of geometric configuration or specified
with a maximum Ls/R greater than or equal to 0.2. detailing method:
The additional forces associated with TDLF 1. Web Plumbness /Girder Layover Tolerance
detailing tend to be prohibitive for highly-curved I-
girder bridges, and thus TDLF detailing of these Tolerances for girder layover are specified in the
types of structures is strongly discouraged. AASHTO/NSBA Guide Specification S10.1-
Therefore, Table 3 does not address estimates for 2014, Steel Bridge Erection Guide Specification
curved bridges detailed for TDLF. (8).
2. Effect of Girder Layover on Girder Stresses and
Construction Inspection Best Strength
Practices Multiple studies have demonstrated that the
As can be seen from the discussions in this research, effects of girder layover on girder stresses and
the behavior of curved and/or skewed steel I-girder girder strength are negligible, including (2), (9),
bridges can be quite complicated, and the (10), and (11).
constructed geometry can change significantly Inspectors should not be concerned about the
through the various stages of construction. strength or stresses in girders which are out of
However, this research has also shown that this plumb.
behavior is also predictable within reasonable
accuracy, and that properly designed, detailed, and 3. Girder Camber at End of Steel Erection
fabricated bridges, when properly assembled, can Most owners require that the tops of girders be
achieve their constructed geometry at all significant surveyed in the as-erected position, prior to
milestones in the construction sequence. installing deck formwork, and the contractor use
Due to the complex nature of the behavior of these this survey information to determine the correct
types of structures, it is advisable that construction position of the deck forms.
inspectors have some knowledge of that behavior, The surveyed profiles of the girder top flanges
and some understanding of the significance of the are compared to the camber profiles on the plans
various notes and information presented on the to check for general conformance. The surveyed
plans. Inspectors should have a clear understanding profile information is also used to determine the
of the meaning of, and differences between, NLF, appropriate position of the deck formwork
SDLF, and TDLF detailing. They should also relative to the girder top flanges; the anticipated
understand the various synonymous terms such as dead load deflection is subtracted from the
Fully Cambered Fit, Erected Fit, and Final Fit. They surveyed elevation of the top of the girder and
should know how to evaluate the constructed then compared to the desired final roadway
geometry profile and deck thickness to determine the
It is critical that inspectors be able to properly assess correct position of the deck formwork relative to
the constructed geometry of a bridge at two key the top flange.
stages of construction: at the completion of steel Generally, if the top flange is a little higher or a
erection, and at the completion of deck placement. little lower than anticipated, the contractor can
Properly assessing the constructed geometry at these compensate by setting the deck formwork a little
key stages, and taking proper action (or properly lower or a little higher respectively. If the
taking no action) will help ensure successful needed adjustments appear to be excessive, i.e.,
construction and minimize problems, delays, and if the haunch will be too deep or too shallow,
unnecessary costs. With a small amount of other actions may be required, such as providing
instruction, inspectors can achieve this goal. haunch reinforcing (for an excessively deep
haunch), adjusting the final roadway profile (for

Page 11 of 16
 an excessively over-cambered girder with a should not undertake remedial action to
“negative” haunch, i.e., girder flange would be “correct” what may be perceived to be a
embedded in the deck), or other actions. “problem” with uplift. For example, if uplift is
anticipated at some interim stage of construction
Owners should clearly specify the required field
and if the designer evaluated this condition and
survey and calculation procedures, and should
found no long-term problems associated with it,
have clearly identified minimum and maximum
the inspector should not attempt to remediate the
haunch values so that inspectors can easily
uplift by means of shims, counterweights, etc.,
review this information and make appropriate
as these actions would interfere with the
decisions on whether to allow construction to
subsequent behavior of the structure and may
continue, to require adjustments to deck forms,
cause long-term problems.
or to contact the Engineer to discuss more
significant remedial actions. 5. Effects of Deviations from Anticipated Web
Position or other Anticipated Constructed
4. Uplift at Bearings
Geometry Measurements
Uplift at bearings may or may not represent a
Layover and web position for various bridge
problem; inspectors should be provided with
geometries and detailing methods will be
sufficient information in the plans to assess the
discussed further later in this section. The
nature of any observed uplift, and should be
possible consequences of unintended layover or
sufficiently informed about this issue so as to
deviations from anticipated web position are
know if and when to involve the engineer in
discussed here in general terms. Inspectors
discussions about possible remedial actions.
should be familiar with these possible
Generally, uplift is considered undesirable by consequences so that they can have informed
most owners, under any conditions. However, discussions with the contractor and the engineer
some leeway is generally given in allowing as appropriate. The possible consequences of
temporary uplift during construction, provided unintended layover or deviations from
that in the final condition there is no uplift. anticipated web position, and some possible
If temporary uplift is anticipated at some interim remedial actions, are listed below. The list of
stage of erection or deck placement it should be possible remedial actions is not meant to be
clearly indicated in the plans or specifications, comprehensive; other actions may be warranted
or clearly communicated at a preconstruction or necessary in specific situations.
meeting or by other means. The locations where • Increased Rotational Demand on Bearings:
uplift is anticipated, and the specific conditions In some cases this may be a minor effect,
under which uplift is anticipated, should be especially if it is determined that the effects
clearly presented. If feasible and appropriate, are temporary (occurring only during an
some measure of anticipated uplift might also be interim stage of construction). For cases of
presented. This information will allow the temporary increased rotational demand on
inspector to compare the as-built condition of bearings, one possible solution might be to
the bridge under those same stages of erection or temporarily support the girders on blocking
construction to the anticipated conditions. If the (removing all load from the bearings), or
observed behavior of the structure is otherwise providing additional support to
significantly different from the anticipated reduce demand on the bearings in the
behavior, the engineer should be contacted and interim condition.
an investigation undertaken to determine the • Girder/Cross-Frame Fit-Up Problems:
causes and possible consequences of this Unintended layover or other deviations from
behavior, and to determine what, if any, the anticipated constructed geometry (such
remedial actions may be necessary. as excessive deflection, particularly
Inspectors should understand that anticipated excessive differential deflection between
uplift during interim stages of construction is not adjacent girders) at interim stages of steel
necessarily a sign of a problem. The inspector erection may be a sign that the contractor is

Page 12 of 16
 losing control of the constructed geometry. both at the end of steel erection (prior to deck
This problem is sometimes difficult to placement) and also after deck placement.
recognize since specific constructed
Most straight, skewed steel I-girder bridges will be
geometry information at each and every
detailed for one of two possible types of fit:
stage of erection typically does not exist.
However, if such information is available, • Steel Dead Load Fit (SDLF, also known as
the inspector should evaluate the constructed Erected Fit): For bridges which are detailed
geometry at interim stages of erection. If for SDLF the girder webs should be plumb
significant deviations from constructed (within reasonable construction tolerance) at
geometry are observed, the inspector and the the end of steel erection, prior to deck
contractor should discuss the matter and placement. If they are not plumb at the end
verify that the problems can be corrected in of steel erection (prior to deck placement),
the next stage of erection. If the structure the engineer should be consulted and
continues to deviate further from its remedial action should be considered. Later,
anticipated constructed position in the next when the deck is placed, the webs will lay
stage of erection that could be a sign that over and be out of plumb. This sequence of
eventually the contractor will be unable to webs being plumb prior to deck placement
fit-up the remainder of the structural steel. and out of plumb after deck placement is
Inspectors should evaluate compliance with normal and generally does not represent a
the anticipated constructed geometry problem.
throughout the erection of the structural • Total Dead Load Fit (TDLF, also known as
steel. The sooner issues are identified and Final Fit): For bridges which are detailed
diagnosed, the better the chances that for TDLF the girder webs should be plumb
simpler, easier actions will be able to correct (within reasonable construction tolerance) at
the problem. the end of deck placement. The webs will
• Misaligned Joints and Barriers: Unintended be out of plumb at the end of steel erection,
layover or deviations from anticipated web prior to deck placement. If the webs are
position at supports under TDL conditions plumb at the end of steel erection (prior to
can result in misaligned joints or barriers. deck placement), or are out of plumb in the
The best time to assess the position of the wrong direction or beyond reasonable
web is at the end of steel erection, prior to construction tolerances, remedial action
deck placement, since there is still a should be considered. If the webs are in
reasonable opportunity to take remedial their correct, anticipated out of plumb
actions at that time. If problems with web position prior to deck placement, then when
position are not identified until after deck the deck is placed the webs will rotate
placement, the range of possible remedial (twist) to a plumb position (within
actions is very limited and generally very reasonable construction tolerance), at least at
costly. Inspectors should carefully evaluate the supports. This sequence of webs being
the position of the webs at supports at the out of plumb prior to deck placement and
end of steel erection, prior to deck plumb after deck placement is normal and
placement. generally does not represent a problem.
Items Related to Straight Skewed Bridges Some owners/designers may present web orientation
information on the plans; if so, the inspector can use
Straight, skewed steel I-girder bridges will often this data to evaluate the positions of the webs at the
exhibit noticeable changes in their web position (i.e., end of steel erection (prior to deck placement). If
noticeable layover) throughout construction. Girder this information is not on the plans, the web
webs will be plumb under only one loading orientation (out of plumbness) at the end of steel
condition. Girder webs that are plumb at the end of erection (prior to deck placement) can be estimated
erection (prior to deck placement) will not be plumb using a simple geometric formula commonly used
after deck placement, and vice versa. It is important by steel detailers. Depending on the owner’s
that inspectors evaluate girder layover at supports

Page 13 of 16
specification requirements, the inspector may be engineer should be consulted and remedial
able to request this information from the contractor, action should be considered. Girder layover
or may only be able to encourage the contractor to in the span at the end of construction is
perform their own evaluation at the end of steel normal in a curved, radially supported
erection. In either case, both the magnitude and bridge and generally does not represent a
direction of out-of-plumbness of the webs at the end problem.
of steel erection should be considered. • Steel Dead Load Fit (SDLF, also known as
Erected Fit): For bridges which are detailed
Items Related to Curved Radially- for SDLF the girder webs should be plumb
Supported Bridges (within reasonable construction tolerance) at
Curved, radially supported steel I-girder bridges will the end of steel erection, prior to deck
exhibit noticeable changes in their web position (i.e., placement, throughout the length of the
noticeable layover) throughout construction, but bridge. If they are not plumb at the end of
only within the span. At the supports the girders steel erection (prior to deck placement), the
will be plumb both at the end of steel erection (prior engineer should be consulted and remedial
to deck placement) and after deck placement. Out in action should be considered. Later, when
the span, the girder webs will be plumb under only the deck is placed, the webs should still be
one loading condition. Girder webs may be plumb plumb at the supports, but will be further out
when shored, or they may be plumb at the end of of plumb in the span. Again, the girders
erection (after shoring is removed but prior to deck should be expected to twist so that the top
placement). It is highly unlikely that the webs will flange is deflected toward the outside of the
be plumb after deck placement. It is important that curve. Girder layover in the span at the end
inspectors evaluate web plumbness at supports at all of construction is normal in a curved,
stages of the construction process, including under radially supported bridge and generally does
shored conditions (if shoring is used), at the end of not represent a problem.
steel erection (prior to deck placement), and after The use of Total Dead Load Fit detailing (TDLF,
deck placement. also known as Final Fit) for curved, radially
Most curved, radially supported steel I-girder supported steel I-girder bridges is strongly
bridges will be detailed for one of two possible types discouraged as its use in these types of bridges
of fit: generally results in excessive fit-up forces.

• No-Load Fit (NLF, also known as Fully Items Related to Curved and Skewed
Cambered Fit): For bridges which are Bridges
detailed for NLF, the girder webs should be
Curved and skewed steel I-girder bridges are very
plumb under shored conditions throughout
complicated structures. They will exhibit noticeable
the length of the bridge. Later, when the
changes in their web position (i.e., noticeable
shoring is removed at the end of steel
layover) throughout construction. Girder webs may
erection (prior to deck placement) the webs
be plumb when shored, or they may be plumb at the
should still be plumb at the supports, but
end of erection (after shoring is removed but prior to
will be out of plumb in the span. Generally
deck placement). It is highly unlikely that the webs
the girders should be expected to twist so
will be plumb after deck placement. It is important
that the top flange is deflected toward the
that inspectors evaluate web plumbness at supports
outside of the curve. Later, when the deck is
at all stages of the construction process, including
placed, the webs should still be plumb at the
under shored conditions (if shoring is used), at the
supports, but will be further out of plumb in
end of steel erection (prior to deck placement), and
the span. Again, the girders should be
after deck placement.
expected to twist so that the top flange is
deflected toward the outside of the curve. If Most curved and skewed steel I-girder bridges will
the girder webs are out of plumb at the be detailed for one of two possible types of fit:
supports at any stage of construction the

Page 14 of 16
 • No-Load Fit (NLF, also known as Fully I-girder bridges is generally discouraged unless the
Cambered Fit): For bridges which are degree of curvature is very small.
detailed for NLF, the girder webs should be
plumb under shored conditions throughout Conclusions
the length of the bridge. Later, when the Improved design, detailing and erection guidelines
shoring is removed at the end of steel to ensure reliable fit-up of skewed and/or curved
erection (prior to deck placement) the webs steel I-girder bridges, based on detailed analytical
will be out of plumb in the span, and studies of twenty-one bridges, including multiple
possibly also at the supports, particularly at framing arrangements on a number of the bridges,
any and all skewed supports. Generally the are provided in the full report which is summarized
girders should be expected to twist so that in this paper. The report provides quantitatively-
the top flange is deflected toward the outside based recommendations regarding the choice of fit
of the curve, but this may not be true if the condition (aka, cross-frame detailing method) and
geometry is particularly complicated. Later, the selection of erection schemes as a function of the
when the deck is placed, the webs which bridge geometry, based on the goal of minimizing
were plumb at the supports prior to deck fit-up forces and facilitating erection. In addition,
placement will likely still be plumb after the report recommends beneficial staggered cross-
deck placement, but will be further out of frame arrangements for straight skewed bridges, as
plumb in the span. Again, the girders should well as framing arrangements around bearing lines at
be expected to twist so that the top flange is interior piers in continuous-span bridges. Simplified
deflected toward the outside of the curve, methods of accounting for Steel Dead Load Fit
but this may not be true if the geometry is (SDLF) and Total Dead Load Fit (TDLF) detailing
particularly complicated. Girder layover at effects, as well as procedures for direct calculation
the end of construction is normal and of the locked-in forces due to SDLF and TDLF
generally does not represent a problem. detailing, are provided. Lastly, construction
• Steel Dead Load Fit (SDLF, also known as inspection best practices are recommended to ensure
Erected Fit): For bridges which are detailed that the erected geometry sufficiently meets the
for SDLF the girder webs should be plumb specified fit conditions. Recommended design
(within reasonable construction tolerance) at specification provisions have been developed and
the end of steel erection, prior to deck incorporated into the 8th Edition AASHTO LRFD
placement, throughout the length of the Bridge Design Specifications (12) that synthesize the
bridge. If they are not plumb at the end of key guidelines that resulted from this research.
steel erection (prior to deck placement), the
engineer should be consulted and remedial Acknowledgements
action should be considered. Later, when
The research discussed in this paper was funded as
the deck is placed, the webs should still be
NCHRP Research Project 20-07, Task 355. The
plumb at the supports, but will be further out
support and input of Technical Committee T-14
of plumb in the span. Again, the girders
(Structural Steel) of the AASHTO Subcommittee on
should be expected to twist so that the top
Bridges and Structures, the NCHRP 20-07, Task 355
flange is deflected toward the outside of the
Project Panel, and Dr. Waseem Dekalbab of the
curve, but this may not be true if the
Transportation Research Board, are greatly
geometry is particularly complicated.
appreciated. The substantive input on the project
Girder layover in the span at the end of
considerations by Mr. Joshua Orton of Heath &
construction is normal in a curved, radially
Lineback Engineers, Inc., and the prior contributions
supported bridge and generally does not
of Dr. Cagri Ozgur and Dr. Telmo Andres Sanchez
represent a problem.
leading to NCHRP Report 725, are also gratefully
The use of Total Dead Load Fit detailing (TDLF, acknowledged.
also known as Final Fit) for curved and skewed steel

Page 15 of 16
References
1) Chavel, B.W., Coletti, D.A., Frank, K.H., Grubb, M.A., McEleney, W., Medlock, R.D., White,
D.W., Skewed and Curved Steel I-Girder Bridge Fit, White Paper prepared for the National
Steel Bridge Alliance (NSBA), August 2016.
2) White, D. W., D. A. Coletti, B. W. Chavel, T. A. Sanchez, C. Ozgur, J. M. M. Chong, R. T.
Leon, R. D. Medlock, R. A. Cisneros, T. V. Galambos, J. M. Yadlosky, W. J. Gatti, G. T.
Kowatch, NCHRP Report 725 - Guidelines for Analytical Methods and Construction
Engineering of Curved and Skewed Steel Girder Bridges, Prepared for the Transportation
Research Board of the National Academies under the auspices of the National Cooperative
Highway Research Program, February 29, 2012.
3) White, D.W., Nguyen, T.V., Coletti, D.A., Chavel, B.W., Grubb, M.A., Boring, C.G., Guidelines
for Reliable Fit-Up of Steel I-Girder Bridges, Final Report of NCHRP Research Project 20-07,
Task 355, Prepared for the Transportation Research Board of the National Academies under the
auspices of the National Cooperative Highway Research Program, October 26, 2015.
4) American Association of State Highway Transportation Officials (AASHTO), LRFD Bridge
Design Specifications, 7th Edition, 2014, with Interim Revisions through 2016.
5) American Association of State Highway Transportation Officials (AASHTO), Standard
Specifications for Highway Bridges, 17th Edition, 2002.
6) Romage, M.L., “Field Measurements on Lean-On-Bracing for Steel Girder Bridges with Skewed
Supports,” M.S. thesis, University of Texas, Austin, TX, 2008.
7) Zhou, C., “Utilizing Lean-On Cross-Frame Bracing for Steel Bridges,” Ph.D. dissertation,
University of Houston, Houston, TX, 2006.
8) American Association of State Highway Transportation Officials/National Steel Bridge Alliance
(AASHTO/NSBA) Steel Bridge Collaboration, Steel Bridge Erection Guide Specification, 2nd
Edition, 2014.
9) Domalik, D. E, Linzell, D. G, and Shura, J. F, “Design and Field Monitoring of a Horizontally
Curved Steel Plate Girder Bridge,” HDR Bridgeline, Vol.14, No.1, 2005.
10) Domalik, D. E, Shura, J. F, and Linzell, D. G, “The Design and Field Monitoring of a
Horizontally Curved Steel Plate Girder Bridge,” in Proc., 84th Annual Meeting of the
Transportation Research Board, 2005.
11) Howell, T. and Earls, C. ”Curved Steel I-Girder Bridge Response during Construction Loading:
Effects of Web Plumbness,” ASCE Journal of Bridge Engineering, Volume 12, No. 4, July,
2007, pp. 485–493.
12) American Association of State Highway Transportation Officials (AASHTO), AASHTO LRFD
Bridge Design Specifications, 8th Edition, 2017.

Page 16 of 16