Engineering Failure Analysis 62 (2016) 49–57

Contents lists available at ScienceDirect

Engineering Failure Analysis

journal homepage: www.elsevier.com/locate/engfailanal

Strengthening of Garudchatti bridge after failure of Chauras

bridge
Harshad Subhashrao Birajdar, Pabitra Ranjan Maiti, Pramod Kumar Singh ⁎
Department of Civil Engineering, IIT(BHU), Varanasi 221005, UP, India

a r t i c l e i n f o a b s t r a c t

Article history: Two continuous lattice steel truss bridges spanning 190 m with subdivided top chord members
Received 16 June 2015 were constructed on the same design across river Alaknanda at Dugadda and Srinagar, respec-
Received in revised form 6 December 2015 tively, in Uttarakhand, India. Both bridges were similar in design and geometry having middle
Accepted 10 December 2015
span of 110.0 m and two end spans of 40.0 m. Garudchatti bridge at Dugadda was constructed
Available online 11 December 2015
first and opened to traffic, but excessive vibrations and lifting of end supports under live load
condition were observed. Construction of Chauras bridge at Srinagar started later, but it col-
Keywords: lapsed during casting of the deck slab due to buckling of one of its top chord compression
Steel truss bridge
members. Failure of Chauras bridge led to the serious concerns in the minds of people and
Bridge failure
technocrats about the safety of Garudchatti bridge also, because it was constructed using the
Buckling
Bridge strengthening same design of Chauras bridge. Therefore, it was decided to strengthen and carry out load test-
Load testing ing of Garudchatti bridge before reopening to the traffic.
In the present work analyses of Garudchatti bridge are presented to identify structurally unsafe
members for most severe live loads given in IRC:6-2010 code, and recommendations for
strengthening of the critical members. Finite element space frame analyses of the bridge
were carried out using STAAD Pro. v8i software to find excessively stressed members beyond
their permissible stress limit. The critical compression members were strengthened by welding
additional channel sections. RCC anchor blocks were constructed to restrict lifting of end sup-
ports under live load. After strengthening, load testing of the bridge was performed to ensure
the safety of the bridge. The bridge is now reopened for traffic.
In the past a number of bridges have collapsed during load testing. In case of any shortcoming
in the design or overloading during load testing, compression members of the truss may sud-
denly buckle and cause collapse of the bridge. Therefore, it is not advisable to load test steel
truss bridges which do not have adequate margin beyond the service load condition.
© 2015 Elsevier Ltd. All rights reserved.

1. Introduction

Lessons from bridge failures may be treated as learning experiences, because when a bridge collapses it has certainly been
pushed to the limit in some way [1]. Therefore, structural collapses in general, and particularly bridge collapses, which are
often most spectacular, have a significant effect on the development of the knowledge of structural action and material behavior
and have spurred research into particular fields.
Nearly 40% bridges in U.S. are structurally or geometrically deficient (M. E. Thiel, et. all, 2001) [2]. Some of the deficient bridges
are in service with speed and/or load restrictions and some are out of service. Deficiencies that may be found in bridges are

⁎ Corresponding author.
E-mail addresses: h_birajdar@yahoo.co.in (H.S. Birajdar), prmaiti.civ@itbhu.ac.in (P.R. Maiti), prof_pks@yahoo.co.in (P.K. Singh).

http://dx.doi.org/10.1016/j.engfailanal.2015.12.002
1350-6307/© 2015 Elsevier Ltd. All rights reserved.
50 H.S. Birajdar et al. / Engineering Failure Analysis 62 (2016) 49–57

numerous, including uncertainty in load carrying capacity, damage to bridge members due to accidents, excessive loss of the
member cross-sectional area due to corrosion, inadequate geometrical clearances, foundation deficiencies, etc. Options available
for addressing the problems associated with a deficient bridge include both rehabilitation and replacement.
On the average one bridge every month has failed during 2010 to 2012 in USA [3]. In the past, a number of steel truss bridges
have failed during various stages of construction or service. The failures have been partial, or total collapses have taken place. The
most common causes of bridge failure include: design defects and deficiencies, failures during construction due to lack of proper
supervision, and construction error. Failure may also be due to long-term fatigue and fracture, overstress of structural elements,
accidental impacts from ships, earthquakes and lack of inspection, and unforeseen events. In the case of truss bridges, failure of
gusset plates connecting members of truss, and buckling failure of compression members are the most happening failures [1].
Daniel Imhof [4] presented 347 bridge failure case studies from 1444 to 2004 and he reported that out of 347 bridges 36 (10%)
bridges failed due to overloading. In 1881, steel truss bridge near Miramont in France completely failed due to buckling of com-
pression member. In 1883, truss bridge over Töss river in Switzerland failed due to buckling of compression member leading to
death of one person and injury to five others. In 1891, truss bridge near Bergbucke in Austria failed due to buckling and lack of
lateral supports. In 1907, Quebec truss bridge over river St. Lawrence in Canada failed due to buckling of bottom compression
member during cantilevered construction phase leading to death of 74 people. In many cases failure during construction and ser-
vice is due to unexpected increased load on the bridge which many times might be beyond the scope of structural designers
knowledge.
In 1892, the semi-parabolic truss arch bridge near Ljubičevo over river Morava in Serbia failed during load testing. The cause of
failure was buckling of compression chord due to defective connection of two part compression members [1].
Imam and Chryssanthopoulos [5] reviewed case studies of 164 failed metallic bridges. Of these 53% were highway bridges, 34%
were railway bridges and a small percentage comprised foot bridges. Out of the 164 reported cases, 87 bridges were classified as
‘collapse’, 73 bridges as ‘no collapse’ and 4 bridges as failure unknown. The majority of the 87 cases of metallic bridge collapses
occurred in the USA (36%) and UK (20%), partly due to the large number of such bridges in these countries. The most important
factors contributing to collapse were design errors (22%), limited knowledge (22%) and natural hazards (21%). For metallic bridges
the most frequently encountered modes of failure are scour of piers/foundations (17%), buckling (16%), fatigue (13%), impact
(13%) and fracture (9%). Fatigue and fracture taken in combination appeared to be the most critical failure mode, closely followed
by buckling.
Collapse of I-35W bridge over the Mississippi River in Minneapolis, Minnesota on August 1, 2007 resulted in deaths of 13 peo-
ple and injury to more than 100 others [6]. The bridge underwent a number of repair and modifications during its service life. The
reconstructions most significant to the collapse were conducted in 1977 and 1998, and involved increasing the thickness of the
concrete deck from 6.5 to 8.5 in., and the addition of new concrete parapets and guard rails. When the bridge was first opened
for traffic, the concrete deck comprised 70% of the total bridge weight and the concrete added in later years increased the weight
of the bridge by more than 20% and thus represented a significant increase in demand on the structural components. In terms of a
mental image, the addition of 2.0 in. of concrete to the deck was equivalent to doubling the weight of the steel [6]. Structural
analysis of the I-35W bridge determined that the bridge collapsed due to 1/2 inch thick gusset plates at joint U10 in place of re-
quired 1 inch thick gusset plate [6].
Truss bridges are more vulnerable to failure as compared to other bridges. A study based on statistical analysis at Buffalo Uni-
versity has been carried out with support of Federal Highway Administration on ‘A Study of U.S. Bridge Failures (1980–2012)’ [7].
Out of 607,375 existing bridges in U.S., 1062 bridges failed during 1980 to 2012, of which 29% are truss bridge failures, although
they occupied only less than 1% of total population of existing bridges in the U.S.

Fig. 1. Failure of Chauras bridge during casting of deck slab.

 H.S. Birajdar et al. / Engineering Failure Analysis 62 (2016) 49–57 51

Failure of 190 m long Chauras Bridge in Uttarakhand, India, which was a three span (40 m + 110 m + 40 m) continuous deck
type truss bridge, took place during casting of the deck slab (Fig. 1). After launching of the steel truss on two piers and two abut-
ments, casting of deck slab was initiated from midpoint of the 110 m span of the bridge towards right pier. During deck slab con-
creting, when concrete was placed in 52.5 m length from middle of the 110 m span towards right pier, bridge suddenly collapsed
due to lifting at the abutments and buckling of one of the top chord members U13U14, claiming six lives with it [8].
Joints U13 and U14 after collapse of the bridge and buckled view of member U13U14 are seen in Fig. 2.
Exact estimation of varying dead loads of the bridge due to regular repair and modification works and day to day increase in
live load on the bridge is beyond the knowledge of a design engineer. Therefore, suitable load factor against collapse at the initial
design stage of a bridge is of vital importance. It is desirable that the bridge gives sufficient warning by deflection before collapse
of the bridge takes place.
Garudchatti bridge over river Alaknanda was similar to Chauras bridge in all aspects of design and construction. After failure of
Chauras bridge, safety of Garudchatti bridge was questionable and it was a matter of anxiety to the concerned technocrats. In the
present work, reanalysis of the Garudchatti bridge identifying critical members and recommendations for strengthening of the
bridge are presented. Finite element three dimensional space frame analysis of the bridge was performed using standard
STAAD Pro. v8i software. Critical compression members were strengthened by welding additional channel sections to these and
thereafter load testing of the bridge was carried out and the bridge was reopened for traffic.

2. Brief description of Garudchatti bridge

Garudchatti bridge, a three span continuous lattice truss bridge of span 190 m, was constructed over river Alaknanda at
Dugadda in Uttarakhand state of India. Design and construction methodology of Garudchatti bridge were similar with the col-
lapsed Chauras bridge. Garudchatti bridge was successfully completed and opened to traffic, but excessive vibrations in the bridge
were observed during movement of heavy vehicles.
Had the Chauras bridge not collapsed during construction due to marginal design, it might have collapsed during service con-
dition with more fatal outcome. After collapse of Chauras bridge and due to unacceptable vibrations in the Garudchatti bridge,
questions were raised over safety of Garudchatti bridge in long term service condition. Therefore, the bridge was closed to traffic
and its design was reviewed for strengthening.

2.1. Analysis of Garudchatti bridge

Three dimensional space frame finite element analyses of Garudchatti bridge truss using STAAD Pro V8i software were carried
out under self weight of steel truss and deck slab for dead load (DL) condition, superimposed dead load (SIDL), foot path live load
(FPLL) and live load (LL) due to four trains of 2-Lane Class-A vehicles.
Fe 250 (fy = 250 N/mm2) grade steel was used in the steel truss. As per IRC: 24-2010 [9], permissible stress in tension over
 
f cb  f y
the net cross sectional area was adopted as 0.6fy, and in compression it was, σ cr ¼ 0:6 n n
1 .
ðð f cb Þ þð f y Þ Þn
The bridge model adopted for the analyses is based on Fig. 3. At the ends of the bridge compression only spring supports are
used to permit lifting of the truss.
The bridge is analyzed with the following loadings, for DL + LL case and 1.2 × (DL + LL) case.
Weight of steel truss = 10,000 kN
Weight of 220 mm thick deck slab = 5.5 kN/m2

Fig. 2. Joints U13 and U14 and buckled member U13U14.

52 H.S. Birajdar et al. / Engineering Failure Analysis 62 (2016) 49–57

Fig. 3. Half elevation of Garudchatti bridge.

Weight of 150 mm thick cantilever foot path slab = 3.75 kN/m2

Weight of 65 mm thick wearing coat = 1.43 kN/m2
Foot path live load (FPLL) = 1.65 kN/m2
Live load due to four trains of 2-lanes of Class-A vehicle as per IRC: 6-2010 [10].
For the critical compression members (Fig. 3), member stresses for the (DL + LL) cases with lifting at the end support and
without lifting, and 1.2 × (DL + LL) case with lifting prevented at the end supports and before and after strengthening are
given in Table 1.
For comprehensive comparison member stresses for the above different cases, with maximum value of stresses marked, are
again shown in Fig. 4.
It is seen from Table 1 and Fig. 4a that compressive stress is maximum (240.8 N/mm2) in member U13U14 under critical
(DL + LL) case with girder lifting at the end support. This stress is in far excess of the permissible stress of 149.8 N/mm2.
Therefore, it was necessary to provide suitable anchorages at the end supports to prevent their lifting.

2.2. Support reaction and anchorage block arrangement

Maximum support reactions under the most critical live loads on the existing bridge under (DL + LL) case are shown in Fig. 5.
RCC anchorages with adequate factor of safety were designed and provided for maximum lifting force of 278.0 kN at the supports,
details for which are given in Fig. 6.
In new continuous span bridges, it is necessary to adopt proper side to main span ratio so as to avoid lifting of end supports. In
Chauras and Garudchatti bridges, the side to main span ratio was 0.364 which caused lifting at the end supports, leading to
Chauras bridge failure during casting of the deck slab.

2.3. Strengthening details of critical members

Top chord compression members U13U14 to U18U19, bottom chord compression members L7L8 to L9L10 and diagonal mem-
ber L8U9 (Fig. 3) were found to be over stressed in (DL + LL) case (Table 1). In order to prevent collapse of the bridge due to
sudden buckling of the compression members during load testing, strengthening of the compression members was carried out
for 1.2 × (DL + LL) load case. Additional area required for strengthening was calculated corresponding to the excess stress beyond
the permissible stress in the members.
The overstressed top and bottom chord compression members for 1.2 × (DL + LL) case were strengthened by welding four
ISMC 400 channel sections, and diagonal member L8U9 was strengthened by welding two ISMC 350 channel sections to the
existing sections as shown in Fig. 7.

Table 1
Cross sectional area of critical members (cm2) and member stresses (N/mm2) in the bridge.

Member Cross sectional area Member stresses due to

no.
For (DL + LL) case without lifting (DL + LL) case 1.2 × (DL + LL) case without lifting

Before After Self FPLL Maximum LL With Without Before After

strengthening strengthening weight+deck+SIDL condition lifting lifting strengthening strengthening

U13U14 215.96 471.16 141.2 3.7 42.2 240.7 187.1 224.5 110.8
U14U15 360.68 615.88 117.0 3.0 32.0 183.9 152.0 182.4 110.8
U15U16 367.4 622.60 137.8 3.6 36.2 208.4 177.6 213.1 129.7
U16U17 436.4 691.60 131.5 3.4 33.6 197.1 168.5 202.2 131.5
U17U18 465.08 720.28 132.9 3.5 33.6 198.1 170.0 204.0 135.1
U18U19 510.08 765.28 125.6 3.3 31.5 186.0 160.4 192.5 131.4
L7L8 587.4 842.6 111.2 2.9 29.9 130.6 144.0 172.8 94.8
L8L9 488.0 743.2 109.2 2.8 30.2 123.3 142.2 170.6 94.3
L9L10 389.8 645.0 104.6 2.7 29.6 130.4 136.9 164.3 78.2
L8U9 294.92 402.2 94.8 2.5 33.1 163.9 130.4 156.5 137.2
H.S. Birajdar et al. / Engineering Failure Analysis 62 (2016) 49–57 53

Fig. 4. Member stresses in the bridge.

54 H.S. Birajdar et al. / Engineering Failure Analysis 62 (2016) 49–57

Fig. 5. Maximum support reactions.

Fig. 6. Arrangement of anchorage blocks at abutment locations.

Common hot rolled and built-up steel members used for carrying axial compression usually fail by buckling. Buckling strength
of these members is affected by residual stresses, initial bow and accidental eccentricities of load, for which depending on their
shape, different buckling classes of members are defined in the Indian Standard code IS: 800-2007 [11] and IRC: 24-2010 [9].
For built up box section thw b 30 and tbf b 30 are specified (Fig. 8).
At all M-Joints of the trusses, all meeting members were cut and riveted to the gusset plate (Fig. 9). Diagonal members of the bridge
were discontinuous at the M-joints (Fig. 9a). Due to this discontinuity major portion of gusset plate connecting the diagonal members
was laterally unsupported and susceptible to buckling. Therefore, these members were made continuous (Fig. 9c).

Fig. 7. Strengthening details.

 H.S. Birajdar et al. / Engineering Failure Analysis 62 (2016) 49–57 55

Fig. 8. Built up box section (IS:800-2007).

3. Load testing

After strengthening of the bridge, its load testing was carried out as per IRC: SP: 37-1991 [12] and IRC: SP: 51-1999 [13] guide-
lines. Fifteen loaded trucks of requisite weights as per Fig. 10 were deployed for gradual loading of the bridge.
The bridge with loaded trucks is shown in Fig. 11.
Deflection of the bridge at the middle of the 110 m span was recorded with the water column arrangement shown in Fig. 12.
The water column arrangement comprised of a water tank of capacity 1000 l and water column measurement assembly
consisting 0.5 inch transparent pipe, 0.5 inch glass tube, graduated scale, steel base stand and clamps for fitting. Water tank
was placed on deck slab at pier location and water column measurement assembly was placed on the deck slab at mid-span lo-
cation where deflection was measured. Difference between water levels in the transparent glass tube, before loading and after
loading, directly gave the deflection at the mid-span.
The bridge was loaded on 31/05/2013 at 11.00 AM and deflections were recorded as per IRC: SP: 37-1991 [12] and IRC: SP: 51-
1999 [13] for which the summary is given in Table 2.
Recovery in deflection was checked as per acceptance criteria given under cl.5.2 of IRC: SP: 51-1999 [13]. Recovery in deflec-
tion at 24 h after removal of the load was 92.5%, which was more than the required 85%, and therefore, performance of the bridge
after strengthening was satisfactory.

3.1. Precaution in load testing

During load testing, a steel truss bridge is loaded to full design live load with impact as per clause 5.2 of IRC: SP: 51-1999 [13].
Therefore, if there is any shortcoming in the design, construction or overloading during load testing, compression members of the
truss may suddenly buckle and collapse without warning causing loss of life and property. Therefore, it is necessary to ensure that
sufficient reserve strength beyond the design strength exists in the bridge truss, for which necessary strengthening of the bridge
before load testing must be carried out.

Fig. 9. Discontinuity of diagonal members at M-joints.

56 H.S. Birajdar et al. / Engineering Failure Analysis 62 (2016) 49–57

Fig. 10. Loading arrangement.

Fig. 11. Garudchatti bridge over Alaknanda river, India.

4. Conclusions

Garudchatti bridge of 190 m span was constructed on the same design as Chauras bridge on the downstream of Alaknanda
river. Whereas, Chauras bridge failed during casting of the deck slab, Garudchatti bridge was successfully completed but needed

Fig. 12. Water column arrangement for deflection measurement at the midspan.
 H.S. Birajdar et al. / Engineering Failure Analysis 62 (2016) 49–57 57

Table 2
Summary of loading and measured deflections.

Date Time Load at LHS (Garudchatti side) Load at middle location. Load at RHS (Brahmapuri side) Deflection

31.5.2013 11.00 A.M. Initial deflection 0 mm

11.35 A.M. 127.20 T 17 mm
11.55 A.M. 128.90 T 127.20 T 130.10 T 30 mm
1.6.2013 12.00 A.M.(mid-night) 128.90 T 127.20 T 130.10 T 34 mm
10.00 A.M. 128.90 T 127.20 T 130.10 T 40 mm
10.30 A.M. After unloading 8 mm
10.30 P.M. 6 mm
2.6.2013 12.00 P.M. After 24h of unloading 3 mm

strengthening. Therefore, it may be concluded that the design of the two bridges was only marginally unsafe leading to collapse of
one bridge and successful completion of the other.
From the presented details and discussions the following conclusions can be drawn.

1. In Chauras and Garudchatti bridges, the side to main span ratio was 0.364 which caused lifting at the end supports, leading to
Chauras bridge failure during casting of the deck slab. In Garudchatti bridge while strengthening the ends were prevented from
lifting by providing additional RCC counterweights at the supports. Therefore, for new continuous span bridges, it is recom-
mended to adopt proper side to main span ratio so as to avoid girder lifting at the end supports.
2. 3-D finite element space frame analyses of the bridge were carried out to access weaknesses of Garudchatti bridge truss and to
carry out the strengthening. After strengthening, load testing was carried out to ensure safety of bridge in service condition as
per IRC: SP: 51-1999 and IRC: SP: 37-1991.
3. For built up box section thw b30 and tbf b30 are specified in Indian Standard code IS: 800-2007 and IRC: 24-2010, which have to be
strictly followed in the design and construction to avoid premature sudden failure of compression members.
4. A number of steel truss bridges have failed in the past during load testing due to sudden buckling of compression members
causing loss to life and property. For Garudchatti bridge, safety of the bridge during load testing was insured by welding re-
quired additional sections to the compression members for 1.2 (DL + LL) case. The additional welded sections would not be
fully effective during elastic condition, but in the plastic condition both sections would be fully effective. Thus, at the Limit
State of Strength, requisite factor of safety against sudden collapse due to buckling of compression members during load testing
was ensured.

References

[1] Z. Šavor, M. Šavor, J. Gao, M. Franetović, Failures of arch bridges — causes, lessons learned and prevention, Sustainable Arch Bridges, 3rd Chinese–Croatian Joint
Colloquium, 2011 (Zagreb, Croatia).
[2] M. E. Thiel, K. Zulfiqar, and M. D. Engelhardt, "Evaluation and rehabilitation of historic metal truss bridges: survey of literature and current practices", Report No.
FHWA/TX-0-1741-1, Texas Department of Transportation, Research and Technology Implementation Office, P.O. Box 5080, Austin, TX 78763-5080.
[3] T. Galarnyk, Listing of USA bridge failures, http://blogsdir.cms.rrcdn.com/31/files/2012/08/U.S.- Bridge-Failure-Listing-from Timothy-G.-Galarnyk-CEO-of-Con-
struction-Risk-Management.pdf2012.
[4] D. Imhof, Risk Assessment of Existing Bridge StructuresPh.D. dissertation submitted to the University of Cambridge, 2004.
[5] B.M. Imam, M.K. Chryssanthopoulos, A review of metallic bridge failure statistics, IABMAS 2010, 2010-07-11–2010-07-15, Philadelphia, USA, 2010.
[6] M. Liao, T. Okazaki, R. Ballarini, A.E. Schultz, T.V. Galambos, Nonlinear finite element analysis of critical gusset plates in the I-35W bridge in Minnesota, J. Struct.
Eng. ASCE (July 15, 2010).
[7] G.C. Lee, S.B. Mohan, C. Huang, B.N. Fard, A study of U.S. bridge failures (1980–2012), Technical Report MCEER-13-0008, June, 2013.
[8] H.S. Birajdar, P.R. Maiti, P.K. Singh, Failure of Chauras bridge, Engineering Failure Analysis, 45, Elsevier 2014, pp. 339–346.
[9] IRC:24-2010, ‘Standard specifications and code of practice for road bridges’, Section V, Steel Road Bridges (Limit State Method).
[10] IRC: 6-2010, Standard Specifications and Code of Practice for Road Bridges, Section II, Loads and Stresses (Fifth Revision), India, 2010.
[11] IS: 800-2007, Indian Standard, Code of Practice for General Construction in Steel (Based on Limit State Method), India, 2007.
[12] IRC: SP: 37-1991, Guidelines for Evaluation of Load Carrying Capacity of Bridges, India, 1991.
[13] IRC: SP: 51-1999, Guidelines for Load Testing of Bridges, India, 1999.