The rehabilitation of long span truss bridge

Tomasz W. SIWOWSKI
Professor
Rzeszow University of
Technology
Rzeszow, Poland
siwowski@prz.edu.pl

Tomasz Siwowski, born 1961,

received his civil engineering degree
from the Rzeszow University of
Technology, Poland. He has
established and worked for Promost
Consulting, Poland, before becoming
Professor at the Rzeszow University
of Technology. One of his main area
of research is related to steel bridges
design and assessment.

Summary
The rehabilitation works on the five span continuous Warren type steel truss bridge built in 1961
have been presented in the paper. Because of steel structure deterioration and huge live loads
growth during recent years, the bridge has been selected for the comprehensive rehabilitation with a
main goal to increase its load carrying capacity. Apart from the strengthening of truss girders (f.e.
member sections enlargement, external prestressing, rivet replacement, etc.), the rehabilitation also
included deck grid extension, redecking of deteriorated concrete slab and upgrading of supports.
The repair program for existing steel structure contained among others the replacement of two
diagonals deformed by impact of vehicles. The rehabilitation procedure has been presented in the
paper as well as the proof load test results, which proved rehabilitation efficiency.
Keywords: steel truss bridge, rehabilitation, strengthening, external prestressing, redecking,
member replacement.

1.

Introduction

The bridge over Vistula river in Nagnajow is a key component of the National Road No. 9 in Poland.
It has been built in years 1959-1961 as five span continuous Warren type steel truss with RC deck
slab (Fig.1). The spans of steel truss are 72+3x90+72 m long. The bridge has been selected for the
comprehensive rehabilitation with a main goal to increase its load carrying capacity up to A class
(the highest) according to Polish bridge code, which is similar to the first class load of Eurocode.
Apart from strengthening of truss girders, the rehabilitation also included full replacement of
deteriorated concrete deck, steelwork corrosion protection and upgrading of supports. The repair
program for steel structure comprised also the replacement of two diagonal truss members, which
were deformed by impact of vehicles. The main rehabilitation works as well as the proof load test
results, which proved rehabilitation efficiency, have been presented in the paper.

Fig. 1: The steel truss bridge in Nagnajow before rehabilitation

Two truss girders with the height of 9,0 m and the spacing of 8,40 m are built of welded members
with riveted connections. Upper and lower chords have hollow box section and diagonals have both
box or I sections depending on compression/tension inner force. The main truss girders are
transversally stiffened with portal frames located over each support and horizontal lateral X-bracing
in both chords levels. The rolled I and channel profiles are used for lateral bracing. The deck is
made of RC slab with the thickness of 0,18-0,24 m, supported on steel grid of riveted floor beams
with the height of 1,20 m and three welded stringers with the height of 0,90 m. The slab which is
composite with steel beams is cut every 27,0 m to release temperature stresses. The grade of steel of
the whole steel structure is 355 MPa and the concrete class of slab was C-30.
The bridge superstructure is supported on five massive piers founded on the caissons while
abutments are founded on Franki piles. On the same supports the superstructure of the adjoining
railway bridge is rested. The railway bridge is of the same type as the road one with the exception
of truss members, which are built-up riveted sections instead of welded as for road bridge. The
railway bridge was not rehabilitated and is out of the scope of this paper.

2.

Bridge assessment and evaluation

Before the design work has begun, the comprehensive assessment of the bridge state of repair along
with load carrying capacity evaluation (with the proof test) were undertaken in order to find the
most relevant methods for bridge rehabilitation. The most severe faults discovered during
assessment were mechanical damages (plastic deformations) due to vehicle collision with bridge
diagonals, cracks in one floor beam over a support and loss of member sections due to intensive
corrosion after 50 years of service.
After detailed FEM study (Fig.2) the actual load carrying capacity of the bridge was established on
the D level of Polish bridge code (20 metric tons vehicle). The most critical elements were deck
grid beams (stringers) and some riveted connections, both in truss girders and deck beams joints.
Moreover several diagonals and sections of both chords revealed the lack of carrying capacity when
the required level of A class service load had been considered in calculations.

Fig. 2: FEM models used for analysis: bridge truss and deformed member
The rehabilitation program included repair/replacement of both impact deformed members.
Additionally, in case of tension member it was necessary to strengthen it. The general scheme for
tension members strengthening in the rehabilitation program assumed section enlargement by
welding cover plates on existing webs and flanges. It was not possible for deformed tension
member. The assessment on heat straightening of the member in question revealed it was
impossible. Therefore the decision was taken to replace the damaged member with a new one, with
re-dimensioned cross section to carry increased traffic loads.
The compression member in question had got sufficient load carrying capacity (if undamaged) to
comply with new requirements. Before the decision was taken to replace or to repair/leave it, a
study had been performed to answer following questions:
what is the ultimate load carrying capacity of deformed member;
is it possible to leave the deformed member without repair in the bridge after rehabilitation (i.e.
assuming new loads);
is it technically feasible to repair the deformed member.

In order to answer these questions a comprehensive FEM study on the deformed member has been
performed including detailed damage inspection and inventory, analysis of load carrying capacity in
deformed state and analysis of redistribution of forces in truss girder due to damage of the member
(Fig.2). Taking into account the FEM analysis results it was quite obvious, that the deformed
compression member had to be replaced [1].
The assessment of actual bridge condition and evaluation of load carrying capacity let to established
the scope of required rehabilitation works, which were necessary to fulfill road administration
expectations. The scope of works (limited in this paper to bridge superstructure) included:
repair, strengthening and replacement of relevant members of truss girders;
strengthening of riveted connections;
repair and strengthening of deck grid steel beams along with RC deck slab replacement;
execution of new bridge deck equipment.
The rehabilitation works mentioned above have been described in the following chapters of the
paper.

3.

The rehabilitation works description

3.1
Truss members strengthening
Two direct strengthening methods of truss members have been employed in the rehabilitation:
member section enlargement with additional welded plates (passive strengthening for upper chord
and compression members) and external prestressing (active strengthening for tension members).
The scheme of member section enlargement as well as strengthening works execution are showed
on Fig.3. It was done with simultaneous members relieving after RC deck slab demolition in order
to ensure old and new added parts cooperation in carrying dead load of the deck after rehabilitation.
Using member influence lines the special sequence of old concrete slab demolition was elaborated
to obtain the expected level and of members relieving.

Fig. 3: Scheme of member section enlargement and view of strengthening execution

Fig. 4: External prestressing of tension members

Tension diagonals were strengthened by external prestressing with high strength rods (Fig.4). This
work was undertaken at the final stage of rehabilitation, after execution of the new RC slab together
with full deck equipment. Active prestressing enabled the full utilisation of the member capacity
also taking into account the immediate and rheological losses of prestressing force. The rods of 26
mm and 36 mm diameter were applied along with system nuts and anchorage plates (Fig.4). The
tailor made steel anchorages were located under the deck slab (active) and on the upper chord
(passive). The prestresing begun in the middle of the both truss girders going simultaneously
towards both ends of the bridge. The rods were protected against corrosion with special HDPE
pipes and additionally against vandalism with steel tubes (lower parts of rods till 2 m above deck
level). Active anchor blocks were equipped with easy removable caps with elastic stuff in order to
adjust prestressing force during bridge operation.
3.2

Members replacement

3.2.1 Tension member

Post tensioning steel bars were selected to relieve tension member and stabilize the truss girder
during the replacement procedure. The tension force in the bars was strictly controlled during
tightening. Bars were prestressed up to the value of axial force in the member in question resulting
from dead load. It was assumed that the new RC slab had already been made, all cover plates of
truss members had been welded, and the crane necessary to perform the replacement operation was
standing on the bridge deck. Because the connection of new member was designed with HSFG
bolts, some unexpected slip and relaxation was assumed. Additional safety margin of 50% axial
force was applied due to uncertainty of current state of forces in stage under construction. It was
assumed that temporary structural monitoring system would be deployed during the replacement
and post tension force would be adjusted if necessary. The prestressing of the replaced member was
realized with four high strength M-24 Macalloy bars and special anchor blocks fixed to the upper
and lower chord of the truss (Fig.5). The deformed member was replaced with new one of the
similar shape with webs and flanges re-dimensioned to larger thickness, providing necessary
enlargement of member cross section.

Fig. 5: Tension member replacement

3.2.2 Compression member
Replacing compression member was much more complicated procedure than replacing tension
member. The difficulties was mainly caused (but not only) by a necessity of relieving compression
member and stabilizing the truss girder geometry during the replacement procedure. Three different
approaches were considered:
replacing the member without relieving;
erection of temporary stiffening structure for relieving member and stabilizing its nodes;

relieving the member by means of the temporary support and imposed load.
The first approach was analyzed and discarded due to excessive internal forces in truss members in
question. While the internal forces were close to the ultimate load carrying capacities of truss
members, the process of cutting the compression member would lead to uncontrolled dynamic
impact to the structure. This would certainly have increased forces in members as compared to the
state of its static equilibrium without compression member. The second option required design and
manufacturing of special stabilizing structure, which would act as temporary reinforcement and
stiffening of truss girder. This option was discarded because of complicated execution and need for
purpose-built structure for only one use. Finally the last option was chosen to build a temporary
support under truss girder (Fig.6). Hydraulic jack was used to push the truss node up to the level,
when the axial force in compression member was completely reduced and thus the member was
relieved. After relieving the deformed member was removed and replaced with new element. The
whole procedure of compression member replacement has been described in detail in [1].

Fig. 6: Compression member replacement

3.3
Deck strengthening and slab replacement
After old deck slab demolition the existing floor beams were strengthened with steel plates bolted to
lower flanges with HSFG bolts. The special epoxy glue was also used to create the additional
bonding between old and new steel elements. The deck grid strengthening procedure comprised
also the assembly of three additional stringers in order to relieve two existing ones. The tailor made
sections of new welded stringers were bolted to existing floor beams with HSFG bolts. On the top
flanges of floor beams and stringers shear studs were installed to connect the new RC slab in
composite action with steel grid. When the new steel grid had been executed the concrete slab was
monolithically cast on the formwork rested on the grid. Special sequence of slab concreting was
established to reduce rheological effects (creep, shrinkage). Finally the insulation layer and two
pavement courses were placed on the slab following anticorrosion works undertaken on the whole
steelwork. The new steel sidewalk brackets were also mounted out of truss girders with the
prefabricated concrete slabs laying on them.
3.4
Strengthening of riveted connections
Strengthening of relevant riveted joints was executed by replacing rivets with HSFG bolts with
additional use of epoxy glue. This hybrid bolted/bonded connections were extensively tested in
Poland and proved to be the best method for riveted joints strengthening [2]. The existing riveted
connections of floor beams were partially dismantled (in turn: web, bottom flange and top flange).
Due to very dense grid scheme of deck beams no additional supports were needed during
dismantling. The procedure of execution of hybrid bolted/bonded connections was as follows:
removing old rivets in relevant part of joint;

geometric inventory of existing holes pattern to produce precise new lap plates;
sandblasting of both surfaces;
epoxy glue placement on clean surfaces;
assembly of lap plates in the joint along with temporary bolts mounting;
final HSFG bolt tightening after glue hardening.

Fig. 7: Procedure of execution of hybrid bolted/bonded connection

Strengthening of truss girder joints was also executed with use of the same technology as for the
deck grid with additional enlargement of gusset plates together with supplementary bolts
installation (Fig.7). During joint strengthening the existing connection was partially dismantled by
removing of 50% rivets. The remaining rivets ensured the relevant load carrying capacity required
for dead and technological load on the bridge. The geometrical adjustment of new lap plate was
then done following surface cleaning, glue placing, bolt installation and tightening. In a similar way
the second part of truss connection was strengthened. After the whole joint had been tightened two
parts of additional lap plate were welded together. In case of tension members lower part of lap
plate was welded also to upper part of the member to ensure tension force transfer from member to
joint.

4.

Monitoring and proof test

4.1
Monitoring during members replacement
A temporary structural monitoring system was deployed to assist in operation of replacement of
deformed compression member. The system consisted of:
a) displacement transducers completed with geodesy survey;
b) strain gauges in key structural members;
c) pressure sensor in hydraulic jack.
The continuous monitoring of displacements was set-up in 5 points of each truss girder. Strain
measurements were performed in 3 members, equipped with 4 strain gauges each. The new member
was controlled after its installation, during lowering the jack on the temporary support. The value of
force in hydraulic jack was controlled by monitoring the value of oil pressure. The displacement of
the jack was measured with displacement transducer. The structural monitoring system allowed for
continuous control of replacement operation and on-site decision making. Comparison of recorded
and expected values before and after operation allowed to check the safety of replacement operation
and to verify whether the bridge structure has been repaired properly. More detailed description of
monitoring during members replacement has been presented in [1].
Replacing a member in steel truss girder was a complicated and dangerous task, requiring constant
monitoring during operation. In order to relieve the replaced member a very precise estimation of
force in the member had to be made. A detailed and well calibrated FEM model of the whole
structure was very useful to simulate and predict strains and displacements in all steps of the
replacement operation. The possibility to get immediately measured values of monitored
displacements and strains was the key factor for successful and precise decision making such as
adjusting jack force, releasing post tensioning bars or finally cutting the deformed member.
According to monitoring results it can be stated that the executed replacement operation retained
original geometry of the bridge and existing force distribution.

4.2
Proof test
The proof test carried out after bridge rehabilitation had two main goals: to check the strengthening
efficiency of the applied methods by a comparison of the same parameters (displacements, strains,
accelerations, etc.), measured before and after strengthening works and thus to prove the increased
load carrying capacity of the bridge. As a test load three heavy trucks with the weight of 30 tons
each were used in various placement configuration according to the influence line of measured
parameter. In the tables below these comparisons have been showed, in turns truss girders
displacements (the middle of chosen span), strains in truss members and dynamic coefficient,
calculated as the ratio of average dynamic and static strains in the members in question.
Table 1: Comparison of truss girders displacement
Truss girder
of first span
No. 1
No. 2

Elastic displacement Elastic displacement

Displacement
after strengthening before strengthening
ratio
[mm]
[mm]
13,3
16,4
0,81
13,4
15,9
0,84

According to comparison of displacements in the middle of truss girders (table 1) the global
stiffness of strengthened bridge increased about 18%. It is mainly due to the section enlargement of
many truss members (upper chord, diagonals), riveted joints strengthening and also the stronger
composite action of rehabilitated deck. In table 2 the average strain reduction ratios for main truss
members have been presented. The average strengthening level is between 13% and 45% depending
on a member. The highest level was obtained for floor beams, the smallest one for tension diagonals
with section enlargement. Strains in diagonals strengthened with external prestressing were
obviously the same, because prestressing force was not measured during proof test and the cross
section of member left almost unchanged.
Table 2: Strain reduction ratio for truss members
No.

Member

1
2
3
4
5
6
7

Lower chord
Upper chord
Compression members
Tension members with welded plates
Tension members with external prestressing
Cross-beams
Existing stringers

Strain reduction ratio

Scope
Avarage
0,74-1,07
0,87
0,26-1,20
0,68
0,30-0,91
0,67
0,68-1,08
0,87
0,77-1,27
0,99
0,10-0,85
0,45
0,56-0,71
0,65

Table 2: Dynamic coefficients measured in main members of the bridge

No.

Velocity
[km/h]

Dynamic coefficient

Tension
Compression
diagonal
diagonal
1.03
1.04
1.04
1.02
1
10
(1.05)
(1.01)
(1.04)
(1.03)
1.01
1.03
1.01
1.02
2
20
(1.02)
(1.03)
(1.03)
(1.03)
1.01
1.01
1.03
1.00
3
30
(1.04)
(1.04)
(1.04)
(1.02)
1.03
1.03
1.02
1.03
4
40
(1.04)
(1.08)
(1.03)
(1.05)
1.25
1.39
1.63
1.54
5
10 + impact
(1.54)
(1.57)
(1.84)
(1.55)
Remarks: the values measured before rehabilitation are given in brackets
Lower chord

Upper chord

Floor beam

Stringer

1.01
(1.03)
1.02
(1.05)
1.01
(1.02)
1.04
(1.07)
1.25
(1.29)

1.00
(1.00)
1.02
(1.04)
1.00
(1.00)
1.00
(1.00)
1.02
(1.07)

Dynamic behavior of the bridge after rehabilitation could be evaluated by dynamic coefficients
comparison, measured before and after rehabilitation. The dynamic coefficients were calculated on
the base of strain measurement in several bridge members when loading by heavy truck passing the
bridge with relevant velocity. In table 3 the dynamic coefficients have been presented. The change
of average dynamic coefficient for the bridge is rather small, only 3%. The biggest reduction (about
5%) was measured for lower chord, the smallest for compression members (about 1,3%). The
dynamic test revealed, that dynamic behavior of the bridge did not change so much due to
rehabilitation and strengthening works. The lesson learned is, that the chosen methods of truss
bridge strengthening do not cause significant improvement of dynamic characteristics and thus do
not considerably decrease a fatigue danger.

5.

Conclusions

The proof test results revealed the high effectiveness of all strengthening and repair methods used
on the bridge. Its load carrying capacity and stiffness were considerably enhanced and the dynamic
behaviour did not change. The actual level of stresses under proof load (which was comparable to
design load of A class) was not large, only 15% of Re for both chords and 20% of Re for diagonals
(Re yield strength of existing steel). The global safety factors estimated on the strain measurement
basis were quite high and amount about 5 for chords, 8 for diagonals and 10 for steel deck members.
All these values revealed, that rehabilitation works carried out on the bridge allowed to gain
expected level of strengthening and reliability. The present load carrying capacity fulfils
requirements of the road administration for bridges located on international road network in Poland.
It seemed that the bridge would serve the next 50 years without the need of extensive and costly
maintenance works. But it was wrong supposition. A year after rehabilitated bridge was open to
traffic the heavy truck hit the portal end post of the truss girder causing extensive deformation of
steelwork. Several riveted joints were also influenced by this struck. The detailed inspection just
after accident discovered several cracks under the gusset plates of riveted joints. Some fatigue
calculations based on the European methodology proved that the remaining fatigue life of the
bridge is very short [3]. Therefore it was decided to implement structural health monitoring system
which ensures the required level of bridge safety. The lesson learnt from the case of Nagnajow
bridge is that even very efficient rehabilitation methods used for old steel riveted bridges do not
restore the required level of safety without structural health monitoring. Therefore SHM
implementation should be one of the most important part in rehabilitation strategy for such kind of
bridges.

6.

References

[1]

SIWOWSKI T., TOWSKI P., Replacing members in steel truss bridge a case study,
The proceedings of 12th International Conference on Structural Faults & Repair-2008,
Engineering Technics Press, Edinburgh, June 2008.
AGODA M., The reinforcing of steel bridges with heterogeneous joint and rope
construction applications, The proceedings of 5th International Conference on Structural
Faults & Repair-1993, Engineering Technics Press, Edinburgh, June 1993.
KULPA M., SIWOWSKI T., Fatigue life assessment of steel truss bridge after failure,
The proceedings of XXVI Scientific Conference on Structural Failures -2013, Miedzyzdroje,
May 2013 (in press).

[2]

[3]