Engineering Structures 35 (2012) 55–70

Contents lists available at SciVerse ScienceDirect

Engineering Structures
journal homepage: www.elsevier.com/locate/engstruct

Studies on failure of transmission line towers in testing

N. Prasad Rao b,⇑,1, G.M. Samuel Knight a, S.J. Mohan b, N. Lakshmanan b
a
College of Engineering Guindy, Anna University, Chennai 600 025, India
b
Structural Engineering Research Centre, Chennai 600 113, India

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

Article history: The towers are vital components of the transmission lines and hence, accurate prediction of their failure
Received 19 February 2010 is very important for the reliability and safety of the transmission system. When failure occurs, direct and
Revised 22 September 2011 indirect losses are high, leaving aside other costs associated with power disruption and litigation. Differ-
Accepted 10 October 2011
ent types of premature failures observed during full scale testing of transmission line towers at Tower
Available online 7 January 2012
Testing and Research Station, Structural Engineering Research Centre, Chennai are presented. Failures
that have been observed during testing are studied and the reasons discussed in detail. The effect of
Keywords:
non-triangulated hip bracing pattern and isolated hip bracings connected to elevation redundant in ‘K’
Transmission line towers
Non-linear analysis
and ‘X’ braced panels on tower behaviour are studied. The tower members are modeled as beam column
Tower testing and plate elements. Different types of failures are modeled using finite element software and the analyt-
Redundant ical and the test results are compared with various codal provisions. The general purpose finite element
Hip bracing analysis program NE-NASTRAN is used to model the elasto-plastic behaviour of towers. Importance of
Bearing failure redundant member design and connection details in over all performance of the tower is discussed.
Ó 2011 Elsevier Ltd. All rights reserved.

1. Introduction Al-Bermani and Kitipornchai [1] modeled the angle members in

The design of transmission line (TL) tower is generally based on
minimum weight basis. The towers, in general, are of lattice type
consisting of legs, primary/secondary bracings and cross arm
members. Structural design of the tower is mainly governed by
wind loads acting on conductor/tower body, self weight of
conductor/tower and other loads due to icing, line deviation,
broken wire condition, cascading, erection, maintenance, etc.
Generally the tower is modeled as a pin jointed space truss. Only
leg and primary bracings are considered in the analysis and the
secondary bracings (redundants) are ignored. Transmission line
towers are generally analyzed by linear static analysis methods
and the maximum member forces are arrived at assuming that
all members are subjected to axial forces only and the deforma-
tions are small. The final member sizes are determined by assum-
ing the effective lengths. The members are designed based on the
prevailing codes of practice. Bearing type bolted connections are
used to connect the tower members with nominal bolt hole clear-
ance of 1.5 mm. Steel equal angle sections with different grades are
generally used in TL towers. Prototype testing of transmission line
tower is recommended to verify the design and detailing. Most of
the power transmission industries all over the world have made
proto type testing of TL towers mandatory.
⇑ Corresponding author. Tel.: +91 044 22641206; fax: +91 044 22671734.
E-mail address: nprao@sercm.org (N. Prasad Rao).
1
Scientist at CSIR-SERC.
the tower as general asymmetrical thin-walled beam column
elements. The material non-linearity is modeled based on the
assumption of lumped plasticity coupled with the concept of a
yield surface in force space. Formex formulation is used for auto-
matic generation of topology, geometry, load and constraint condi-
tions of TL towers. The effect of joint flexibility can be incorporated
by modifying the tangent stiffness of the element using an appro-
priate moment–rotation (M–hr) relation for the joint [2]. The effect
of nonlinearity is treated as an effective load in conjunction with
the applied loads for general equilibrium [3]. The load is modified
by a residual force to maintain equilibrium instead of modifying
the global tangent stiffness matrix to reflect the extent of nonlin-
earity in the system. Al-Bermani and Kitipornchai [4] reviewed
the current design practices of TL towers and developed a software
which includes the geometric and material nonlinearities, joint
flexibility and the effects of large deflection for the prediction of
the ultimate strength and behaviour of TL towers under static load
conditions. Kemp and Behncke [5] conducted a series of tests on
cross bracing system widely used in TL towers and concluded that
due to dominant effect of end eccentricity, the secant formulation
can be used to provide a design formulation for cross bracing. Sim-
ple diaphragm bracing systems can be very effectively used in
upgrading of old towers. Analytical and experimental studies
showed considerable strength improvements using diaphragm
bracings [6]. The most efficient diaphragm bracing system pre-
dicted from the experimental results was implemented on an
existing 105 m high TV tower. Al-Bermani and Kitipornchai [7]

0141-0296/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.engstruct.2011.10.017
56 N. Prasad Rao et al. / Engineering Structures 35 (2012) 55–70

presented a non-linear analytical technique accounting for both

Second panel Leg 4 parts: Leg and brace failure

material and geometric non-linearity to predict TL tower failure

Horizontal redundant member bearing failure

and calibrated the same with results from full scale tower test con-
ducted on 275 kV double circuit TL tower.

2. Design practice

Bracing failure major axis

Bracing in second panel

Leg failure (Leg 5 parts)

Bracing in first panel

Bracing in first panel

The members in TL tower are generally subjected to tension or

Leg in second panel

Leg in second panel

Leg in second panel

Leg in fourth panel
Leg in third panel
compression forces due to external loads. The design of leg mem-

Leg in first panel

Leg in first panel

Failed members
ber is reliable, since the assumption of concentric load at both ends
of the member is achieved in a real structure due to geometric
symmetry of the structure. The eccentricities involved in the brac-
ing member connections are accounted in the form of end-restraint
coefficients and hence bracing members design do have certain
approximations built into them.

BS in kN
The tower members are generally made of steel equal angle sec-

1744

1939

2465

2534
476

1980

813

253
351
470

36

170
108
807

62
tions. Generally buckling strength of the member about VV-axis
(minor axis) is considered in the design. If the main member is re-
strained by a redundant member connected to a relatively rigid

ASCE 10-97/IS:802 in kN
member at its other end, then it can prevent VV-axis buckling of
full member and increase its buckling strength. If buckling about

Member capacity
VV-axis is prevented using redundant member, then the member
has to buckle about its rectangular axis (XX-axis) for the same
length. This principle is used in the general design practice of TL

65
1644
40
1943
1882

2378

2485
441
433

192
133
774
785

291
375
towers. The capacity of members given in Table 1 are determined
based on the buckling formulae given in IS: 802 (Part1/Sec. 2)-
1992 [8] and ASCE 10-97 [9] which are formulated in accordance Non-linear analysis force
at test failure load in kN
to ASCE manual 52.

3. Failure of transmission line towers

424

1308

1011

267
213
1930
402

26

970
62

520
505
35
50

1962
Even though transmission line towers are designed based on
the codal provisions, some of them may fail during testing due to
many reasons such as incorrect design assumptions, improper
detailing, material defects, fabrication errors, force fitting during
FE NL analysis
Failure load in percentage

erection, and variation in bolt strength. The failures occurred at

Tower Testing and Research Station, Chennai are classified as
94.4

96.6

structural failure like leg, bracing and redundant failures, detailing

106

103

106

101
and connection failures, material defects and fabrication errors. 103
Generally leg members are designed with slenderness ratio varying
from 35 to 50 and at this range the compression capacity is always
62.3
Test

more or less equal to the net tension capacity of the member. Many
75

95
100

90

100

80

failures of the towers are caused by buckling of compression brac-

ing and leg members. The redundant members may also cause fail-
Comparison of member capacities, analytical and test failure loads.

350
350
350

410
410

350
350
350
350
255

275
275
255
255
255
Fy

ure of main members if it is not properly designed.

50

40
128

182
62
64
61
150
36
43
150

61
58

105
76

TC, MC, BC: top, middle and bottom conductors broken.

L/r

4. Present study
150  150  18
150  150  18

150  150  12
150  150  12
150  150  20

200  200  20

ISA 200  200  20

120  120  8
100  100  8
100  100  8

80  80  6
90  90  6
Member details

55  55  4

90  90  6

80  80  6

Out of 145 full-scale tower tests conducted, 33 towers experi-

enced different types of premature failures. For the present study,
five towers are investigated for their failure during testing. Differ-
ent types of failures that occurred in these five towers are investi-
Size
ISA
ISA
ISA
ISA
ISA
ISA
ISA
ISA
ISA
ISA
ISA
ISA

gated. The member capacities have been worked out for individual
cases following ASCE 10-97/IS: 802 provisions and are given in
(b) TC & MC (Trial 1)
(c) TC & MC (Trial 2)
400 kV D/C (15–30o)

Table 1. The capacity of member calculated based on IS: 802 stan-

275 kV D/C ‘H’ type
400 kV D/C (0–15 )
o

dard is same as ASCE. To study the failures in detail, nonlinear

132 kV (0–2o)

400 kV (0–2o)

finite element analysis using NE-NASTRAN was carried out. The

(a) TC & BC
Tower type

member capacities and forces at the time of failure obtained using

beam column elements in finite element models are given in Table
Table 1

1. Member capacities calculated based on BS-8100 [10] standard is

compared with ASCE 10-97 and test results in the same table.
 N. Prasad Rao et al. / Engineering Structures 35 (2012) 55–70 57

10.34
3830
10.34
1250 1.36
1.36 650
1250 L70x6
24.17 L50x5
3830 24.17 1200
1380 L90x7H
5.84 5.84 L60x5
1700
25.59 1151
3875 25.59 L60x5
1350 L90x7H
5.84 1450 5.84 L60x6TF
1800
30.33 L60x5LF
4180 30.33 900
2000
L60x5LF
5.84 5.84
2580
L 50x5TF

3900 L 60x6 LF
L 60x5 TF

2530
19.28
5010
L50x5
14.9

IR L100x8H
HR 6000
L60x5 Fig. 3. Failed leg member at 12 m level.
L50x5
18.22
BRACES

6000 model each angle member of the tower. A minimum four beam
LEGS

column elements are to be used to model a member length ‘L’ for

8714
H HIGH STRENGTH STEEL Fy = 350MPa
DIMENSIONS IN mm: LOADS IN KILO NEWTONS
FAILED LOCATION OTHER MEMBERS Fy= 255MPa
Fig. 1. Configuration of 132 kV D/C tower and loads.
5. Numerical method
The nonlinear finite element analysis methods are effective for
evaluating the behaviour and strength of compression members in
space frames. The non-linear finite element analysis program
NE-NASTRAN is used in this study to model the elasto-plastic
behaviour of tower material. Beam column elements are used to
convergence. So each member of length ‘L’ is subdivided into four
elements. Even though, the beam column element is a line ele-
ment, shape of the section can be defined. The stresses can be ob-
tained at different points on the cross section called stress recovery
points. The nonlinear analysis capability of the software account-
ing for geometric and material non-linearity is used in the analysis
of idealized configuration models to obtain their pre-ultimate
behaviour and the limit loads. The elastic and plastic material
property of steel was represented by an elasto-plastic bi-linear
model, with a modulus of elasticity as 2E5 MPa up to yield and
2000 MPa above yield stress. The stress strain behaviour was given
as input for modeling the material properties for some towers for
which the tensile test data was available. The incremental load
and predictor–corrector iteration scheme was used under each
load increment in the nonlinear range. The isotropic hardening

Fig. 2. Rigging arrangement (typical).

58 N. Prasad Rao et al. / Engineering Structures 35 (2012) 55–70

Fig. 4. FE model beam elements and plate elements.

Fig. 5. FE model stress pattern: beam and plate elements.

model was used in the post yield range. Load increments were car-
ried out in 30–35 steps, until the limit point was reached in the
load deformation behaviour. The yielding is modeled by the von
Mises criterion. Arc-length method in conjunction with modified
Newton Raphson method is used to further accelerate numerical
solution convergence.
NE-NASTRAN and the advanced non-linear finite element analy-
sis by Al-Bermani et al. (1993) for transmission line towers consider
all the members in the tower, account for geometric and material
non-linearities, and use beam column/truss/plate elements. The
effect of combined stresses on general thin walled beam-column
element is taken care off by using yield surface and lumped plasticity
approach in Al-Bermani’s model. In NE-NASTRAN, the limit load is
reached when the pre defined ultimate stress is reached at stress
recovery points on the element whose shape is given as an input.
In-plane joint eccentricities can be adequately modeled in both
the approaches. Out-of -plane joint eccentricities have been ac-
counted for using an analytical model proposed by Prasad Rao and
Kalyanaraman (2001) [11] wherein the bolts are modeled as beam
elements with and without rotational rigidities depending on single
or multiple bolts.
6. Towers studied
For the present study, five vertical configuration towers are con-
sidered. This includes three 400 kV, one 275 kV and one 132 kV
double circuit towers.
 N. Prasad Rao et al. / Engineering Structures 35 (2012) 55–70 59

16
4350 5050
16 4
65x5

x6
4

75

80
258

x5
78 8260 L75x5
3200
135 3
T55x4HTF
18
4 8 L100x10H T 75x6HTF
5300 6150 47 T 75x5HLF
258 8660
63 1 T 75x5HTF
2 78 L110x10H
26 T 70x5HLF
6 7
135 L130x12H T 90x6H
5300 6150
47 8340
63
L150x15H
3
5300 T 90x6HTF
135 39 4 5 135
5770 6620 5350 L150x20H T 80x6HLF
63 63

8000
37
T 75x5H
T80x6H
7500
31 T 80x6 L200x16H

9000 T 75x6
16180

L EQUAL ANGLE SECTION FAILURE LOCATION

1. L90x90x7H 2. L100x100x6H 3. L100x100x7H 4. L110x110x8H
5. L110x110x10H 6. L120x120x8H 7. L120x120x10H 8. L130x130x10H
H HIGH STRENGTH STEEL Fy =350 MPa OTHERS Fy =255 MPa
DIMENSIONS IN mm : LOADS IN KILO NEWTONS

Fig. 6. 400 kV D/C ‘DB’ type tower details and loads.

6.1. 132 kV double circuit tower
The 132 kV double circuit tower [12] is designed for 0–2° line
deviation. The tower has been designed for 11 different load cases
under reliability, security and safety conditions such as normal,
one conductor broken and ground wire broken conditions. High ten-
sile steel of 350 MPa yield stress for leg members and mild steel of
255 MPa for all other members are used. The 39.5 m high square
tower is 8.7 m wide at base and tapers to 1.45 m at 29 m level. Tower
configuration, cross-arm details, secondary bracing pattern, dimen-
sions and loads corresponding to failure loading condition are
shown in Fig. 1. High tensile steel angle of ISA 100  100  8 mm
is used for leg members from ground to bottom cross arm level.
During testing, the transverse horizontal, vertical and longitudi-
nal loads are applied separately. Load and angle sensors are fixed at
cross arm tips to measure the loads. Wind loads on tower body in
between cross arms are combined with the conductor loads at each
cross arm level. The wind on tower body below waist level was
combined and applied at 6, 12, and 17 m levels from ground. The
loads are applied through centralized servo controlled hydraulic
actuators simultaneously at all loading points. Typical steel wire
rope arrangement (referred as rigging) for load application on
tower is shown in Fig. 2.
The testing was conducted based on IEC [13] specifications. The
deflection measurements were taken at each load level in the
transverse and longitudinal directions at top of the tower. On
applying the loads corresponding to normal condition with mini-
mum vertical load case as given in Fig. 1, at 100% load, the leg
member at 12 m level failed as shown in Fig. 3. The non-linear
analysis force in the failed leg member was 4% more as compared
to conventional linear static analysis force. The capacity of the
failed leg member predicted based on ASCE standards was 4% more
compared to the non-linear analysis force at 100% load. The test
deformation measured at top of the tower was 1.6 times more than
the analytical deformation.
The entire tower was modeled using beam-column elements in
NE-NASTRAN. From finite element non-linear analysis the failure
load is predicted as 104% of normal condition. In order to check
the stress distribution in the failed leg member, all four leg mem-
bers in the failed ‘K’ braced 6 m panel and some portion in third pa-
nel are modeled using plate shell elements. At the transition
between beam and plate elements, a rigid element was used to
connect the beam element node with the nodes of the angle sec-
tion modeled as plate elements. The finite element analysis failure
pattern with all members modeled as beam elements and another
with plate elements for leg member alone in the failed panel are
shown in Fig. 4. The non-linear finite element analysis shows that
the stresses have reached yield level only in some portion of leg
member as shown in Fig. 5.
The leg member in the failed ‘K’ braced 6 m panel was divided
into five equal parts and triangulated hip bracing pattern was used.
The horizontal (secondary) redundant bracing member marked as
60 N. Prasad Rao et al. / Engineering Structures 35 (2012) 55–70
Fig. 7. Failure of leg and redundant members.
‘HR’ in the fourth sub-panel at 11.8 m level was connected to leg
member at one end and to an inclined redundant member at the
other end. As the leg member deformed inward about its minor
axis, the horizontal redundant member connected to it also
deformed along its axis since it is connected at mid length of
another flexible redundant. The failure shows the in plane defor-
mation of the inclined redundant ‘IR’ about its full length. The
redundants have not offered the required restraint to the leg mem-
ber. It is apparent that for the flexible redundant member to be
effective, a minimum stiffness provision needs to be advocated
possibly by restricting the flexural deformation of the flexible
redundant.
6.2. 400 kV double circuit medium angle deviation tower
The 400 kV double circuit tower [14] was designed for 0–15°
line deviation. The tower was designed for different load cases
such as normal, simultaneously two conductors broken at a time
and ground wire plus one conductor broken conditions. High ten-
sile steel of 350 MPa yield stress for leg members and mild steel for
all other members are used. Tower configuration, cross-arm de-
tails, secondary bracing pattern, dimensions and loads correspond-
ing to failure loading condition are shown in Fig. 6.
The transverse horizontal, vertical and longitudinal loads are
applied as mentioned earlier in Section 6.1. Wind loads on tower
body in between cross arms are combined and applied separately
Fig. 8. Finite element failure model.
at each cross arm level. The wind on tower body below waist level
was combined and applied at 16.5 and 9 m levels.
The testing was conducted based on IS: 802 part III [15] speci-
fications. The deflection measurements were taken at each load le-
vel in the transverse and longitudinal directions at top of the
tower. On applying the loads corresponding to top and middle con-
ductor broken condition test, at 90% load, the leg and redundant
members in first and second sub-panels of ‘X-braced’ fourth panel
at 25 m height failed as shown in Fig. 7. In this ‘X-braced’ panel the
leg member was divided into four equal parts. The bottom and top
half of X-bracing was divided into two equal parts and fully trian-
gulated hip bracing pattern was provided. Back to back connected
ISA 90  90  6 mm angles for main bracing, ISA 55  55  4 mm
for redundant member and high tensile steel angle of ISA
150  150  20 mm for leg members were used. Horizontal redun-
dant member in the first sub-panel was connected to the leg mem-
ber by two bolts and the other end was connected to main bracing
by single bolt through a gusset plate.
The entire tower was modeled using beam-column elements in
NE-NASTRAN. The non-linear analysis failure pattern is shown in
Fig. 8 and the failure load is 103% of top and middle conductor bro-
ken condition. The predicted capacity of leg, bracing and redundant
member based on ASCE/IS standards was 13%, 17% and 54% more
than the forces obtained from non-linear analysis. The finite ele-
ment analysis failure pattern shows that the leg member was
deforming outward. In order to check the stress pattern, the failed
 N. Prasad Rao et al. / Engineering Structures 35 (2012) 55–70 61

Fig. 9. FE analysis stresses in failed leg and redundant. Fig. 11. Failure at second panel.

7400 43
43 1 2750 12
1 4650 L80x6 60x5TF:50x5LF
12 2 3 105 60x5TF:55x5LF
209 5 8 105 41 3400 L120x8H H90x6TF:90x6LF
81 2 4
41
106 3300 90x6TF:75x5LF

212
5 7
41 3400 L130x12H H100x7TF:90x6LF
L90x6 106
81 2 4 3300 L150x12H H100x7TF:90x6LF
41
4 6 107 3100 L150x12H H100x6TF:90x6LF
215 4800
VIEW D-D 107 41
81 90x6TF:80x6LF
L100x6TF L90x6LF 41 9200
100x6TF:90x6LF
25 D D L150x18H
VIEW A-A 6000 90x6TF:80x6LF
C
28 A
A
BRACES

VIEW C
9000 120x8TF:100x6LF
LEG

B
VIEW B 16580
LF LONGITUDINAL FACE
H HIGH STRENGTH STEEL = 410MPa : OTHER MEMBERS = 275MPa : L ANGLE SECTION
ALL DIMENSIONS ARE IN mm: LOADS ARE IN KILO NEWTONS TF TRANSVERSE FACE
1.L65x6 2.L65x6H 3.L70x6H 4.L75x6H 5.L80x6H 6.L90x6H 7.L100x6H 8.L100x7

Fig. 10. Configuration of 275 kV dead end ‘H’ type tower.

62 N. Prasad Rao et al. / Engineering Structures 35 (2012) 55–70

Fig. 12. Finite element failure pattern at second panel.

3450
1 1
1500
12 12
L60x5 2000 L45x4
L60x5
2100
52 52
3000
L90x6 L60x4
6100 1825
10 10
56 3175 L110x8
56 L65x5
2000
6450 2772 L70x5
10 L130x10
10 L75x5
71 3228
71 1700 L65x5
3000 7250
4000
10 10
D L60x5
5500
L75x6 D
VIEW D L60x5
5600
L150x12

D L65x5

3840 L70x5

L75x6 VIEW C
PLAN A-A 5159
C L75x5
81 A
A
9000
L80x6
B
VIEW B
13066
'L' EQUAL ANGLE ANGLE SECTION : YIELD STRESS 255MPa
DIMENSIONS ARE IN mm : LOADS IN KILO NEWTONS

Fig. 13. Configuration of 400 kV tangent ‘DA’ type tower.

leg and redundant members are modeled using plate elements
keeping all other members as beam-column elements. The finite
element failure load was same as that of beam column model.
The non-linear analysis stresses in some portion of leg member
and in first sub-panel horizontal redundant member crossed the
yield stress as shown in Fig. 9.
The test failure shows that the leg member buckled about its min-
or axis over two sub-panels (0.5 L) instead of assumed one sub-panel
(0.25 L). Horizontal redundant member in first sub-panel connected
to main bracing in transverse and longitudinal faces failed by end
tearing. The redundant member capacity in compression is 54%
more as compared to the nonlinear analysis force and 11% higher
as compared to the 2.5% of axial force in the main member. Bearing
capacity of redundant member is 11% less as compared to the force
predicted based on 2.5% of main member force. From this it is clear
that the bearing failure in redundant member has caused the leg
buckling and shows the importance of detailing. This is an extremely
minor detail in the whole design process, which could not be easily
detected. Also, assuming that two bolts had been provided in the
design, there is every chance that during erection one bolt may be
 N. Prasad Rao et al. / Engineering Structures 35 (2012) 55–70
Fig. 14. Collapsed tower in testing.
missed. This clearly indicates the importance of detailing as also
through inspection at site is essential. Absence of single bolt may
lead to premature tower failure.
6.3. 275 kV double circuit dead end type tower
The 275 kV double circuit tower [16] was designed for maxi-
mum deviation locations. The tower was designed for different
Fig. 15. Finite element analysis failure pattern.
63
load cases such as normal, two conductors broken at a time and
ground wire alone broken condition. High tensile steel of
410 MPa yield stress for leg members and mild steel of 275 MPa
for all other members are used. The tower configuration, cross-
arm details, secondary bracing pattern, dimensions and loads cor-
responding to failure loading condition are shown in Fig. 10.
In testing, the transverse horizontal, vertical and longitudinal
loads are applied separately. The testing was conducted based on
the IEC specifications. The deflection measurements were taken
at each load level in the transverse and longitudinal directions at
top of the tower. The tower failed in normal condition loading in
which it was subjected to transverse and vertical loads. The failure
has occurred in the second level ‘K’ braced 6 m panel. The leg and
bracing members in this panel was divided into five equal parts.
The bracing member was designed for two-fifth length over major
axis (0.4 L/rxx). As per the design, based on ASCE and IS standards,
the leg and ‘K’ bracing members in the failed panel were having a
safety margin of 1.21 and 1.62 respectively. But the failure oc-
curred at 63% of normal loading condition as shown in Fig. 11.
In the bottom 9 m ‘K’ braced panel the leg and bracing members
are equally divided into nine parts and fully triangulated hip brac-
ing pattern was provided as shown in Fig. 10 and this has behaved
as assumed and hence there was no failure in the bottom panel.
The bottom 9 and 6 m ‘K’ braced panels were having same safety
margins. The only difference being non-triangulated hip bracing
pattern was provided in the 6 m ‘K’ braced panel (Fig. 10). Isolated
hip bracing members were provided in the first and third sub-pan-
els. These hip bracings were not connected to the main K-bracing
directly but it was connected to a redundant in elevation away
from the main bracing and elevation redundant intersecting node.
Full tower was modeled using beam-column elements. In the fi-
nite element model the isolated hip bracing connected to elevation
redundant member was also modeled and the failure pattern is
shown in Fig. 12. Analysis failure model shows the inward deforma-
tion of main bracing over full length of the member. This clearly
shows that the non-triangulated hip bracing pattern provided in
the form of an isolated members were not effective and have not of-
fered required restraint to the main bracing to behave as it was as-
sumed. The main K-bracing has buckled over three to four-fifth
length (0.6–0.8 L) instead of assumed 0.4 L and subsequently lead
to leg failure. FE analysis shows that the behaviour of 9 m ‘K’ braced
64 N. Prasad Rao et al. / Engineering Structures 35 (2012) 55–70

4420 6120 23 23
23 23
b x5 L90x6T TF
ANELS

d g 75 x6 5705
d 80 L80x6 LF 4 4 4 251
b a b 3200 4
f 2 251
b b 1 2555 H L75x6T TF 201
d g 201
5 SUB P

b L110x8 H 111
d 4 6 3610
R b 7080 8770
IR

d H3 H L75x6T TF 18 111 18 47
b 63 47 63
4900 L110x10H
a 251
DETAIL AT 'P':CASE II (TRIAL 2) 201 201
a 3 6 3410
a 7574 9264 L150x15H H L90x7T TF 201 111
h
ANELS

d a 27 27
c b 2425 H L90x7T TF 63 63 47
5500
abf L150x16H H L100x6T LF 251 63
d ad VIEW 'R' 2875
d HR3
5 SU B P

e g 8620 10320 5 5124 H L100x6T TF 201

H3

L200x16H 42
H2

42 111
c

bd H L90x6T LF 201
IR

201
H1

a 63
63 47 63
VIEW A-A H L80x6T TF
DETAIL AT 'P' :CASE II (TRIAL 1) 8400 H L75x5T LF
c d A L80x6T A
a d 39 39
c a
ELS

SP4 a L 200x20H
g
e H3 P 7500 H L80x6T
PAN

SP3 h e c
32 32
b c
4 SUB
IR

SP2 Q d H L90x6T
SP1 a 9000 LOADS IN KILO NEWTONS
DETAIL AT 'P': CASE I VIEW 'Q' 17210
a) L45x4 b) L50x4 c) L55x4 1) H90x6 H HIGH STRENGTH STEEL Fy=350MPa DIMENSIONS IN mm (a) TOP AND BOTTOM (b) TOP AND MIDDLE
d) L60x4 e) L65x4 f) L70x5 2) H100x6 OTHER MEMBERS Fy=255MPa 4) H120x8 5) H120x10 6) H130x10 CONDUCTORS BROKEN CONDUCTORS BROKEN
g) L75x5 h) L75x6 i ) L80x6 3) H110x8 'L' EQUAL ANGLE SECTION 'T' ANGLE CONNECTED BACK BACK TRIAL 1 & TRIAL 2
IR INCLINED REDUNDANT: HR HORIZONTAL REDUNDANT: H HIP BRACING:
TF TRANSVERSE FACE : LF LONGITUDINAL FACE:

Fig. 16. Configuration of 400 kV D/C 15–30° tower and loads.

Fig. 17. View of test tower (Case I): failure in top and bottom conductor broken.

panel was as per design assumptions and there was no failure of members. Tower configuration, cross-arm details, secondary brac-
bracing and leg members in that panel. Finite element analysis ing pattern, dimensions and loads corresponding to failure loading
stresses in leg member is less than 370 MPa which is well below condition are shown in Fig. 13. The tower was tested with +9 m
the member yield stress. Specific recommendations for hip bracing body extension. Steel angle of ISA 150  150  12 mm was used
pattern may be incorporated in the codes. Unconventional bracing for leg members from ground to bottom cross arm level.
patterns need to be validated by separate sub-assembly tests. In testing, the transverse horizontal, vertical and longitudinal
loads are applied separately. The testing was conducted based on
6.4. 400 kV double circuit tangent type tower IS: 802 specifications. The tower collapsed at 75%, corresponding
to normal condition loading as given in Fig. 13. In this load condi-
The 400 kV double circuit tower [17] was designed for 0–2° line tion, the tower was subjected to transverse and vertical loads only.
deviation. Mild steel of 250 MPa yield stress was used for all In the bottom 9 m panel, the ‘K’ bracing and the leg member were
 N. Prasad Rao et al. / Engineering Structures 35 (2012) 55–70 65

Fig. 18. Leg and bracing failure in FE model: top and bottom conductor broken.

Fig. 19. View of test tower (Case II: trial 1): bracing failure in top and middle conductor broken.

divided into five equal parts. The leg member was designed for (rvv)
minimum axis buckling about one fifth (0.2 L) lengths. Where as
the ‘K’ bracing member was designed for (rxx) major axis buckling
about two-fifth length (0.4 L) by providing isolated hip bracing at
alternate sub-panels. The capacity of the leg member predicted
based on ASCE and IS standards in the bottom most two panels
were 12–16% more compared to analytical force.
Fig. 14 shows that the entire tower is intact above two panels
from ground level. It shows that the failure has occurred in bottom
‘K’ and ‘X’ braced panels. The capacity of K-bracing in the bottom
9 m panel predicted based on the assumption of major axis buck-
ling about two-fifth length (0.4 L) was 38% more as compared to
the analytical member force at 100% load. The ‘K’ bracing capacity
calculated based on the assumption that the isolated hip bracing is
ineffective and buckling about its major axis on two-third length
(0.6 L/rxx) is closer to the member force determined based on test
failure load.
From finite element non-linear analysis the failure load is pre-
dicted as 96% of normal condition. The FE failure model is shown
in Fig. 15. The hip bracing has not offered sufficient restraint to
66 N. Prasad Rao et al. / Engineering Structures 35 (2012) 55–70

Fig. 20. Bracing failure in FE model: top and middle conductor broken.

Fig. 21. View of test tower (Case II: Trial 2): leg failure in top and middle conductor broken.

the main K-bracing member and hence it has buckled in 0.6–0.7 L/
rxx mode, which is correlating well with FE failure pattern and the
capacity calculated based on this assumption. Non-triangulated
hip bracing pattern has caused the geometric instability in the
structure leading to premature failure. The failure of the tower
reinforces the conclusions drawn under Section 6.3.
6.5. 400 kV double circuit (15°–30°) angle deviation tower
The 400 kV double circuit tower [18] was designed for 15°–30°
line deviation locations. The tower was designed for different load
cases such as normal, simultaneously two conductors broken and
ground wire along with one conductor broken conditions. High
 N. Prasad Rao et al. / Engineering Structures 35 (2012) 55–70 67

failure occurred once in top and bottom conductor broken condi-

tion loading and twice in top and middle conductor broken condi-
tion loading. Studies are conducted on these two cases of failure.

6.5.1. Top and bottom conductor broken condition

Fig. 22. Leg failure in FE analysis.
tensile steel of 350 MPa yield stress for leg members and mild steel
of 255 MPa for all other members are used. Tower configuration,
cross-arm details, secondary bracing pattern, dimensions and loads
corresponding to failure loading condition are shown in Fig. 16.
In testing, the transverse horizontal and vertical loads are com-
bined and applied as resultant load at a resultant angle. The verti-
cal and longitudinal loads are applied separately. The testing was
conducted based on IS: 802 part III specifications. During testing,
In top and bottom conductor broken load condition, the tower
was subjected to transverse and vertical loads in addition to the
longitudinal loads at top and bottom right side cross arms as given
in Fig. 16. At 100% load, the leg and bracing members in the second
‘K’ braced 7.5 m panel failed as shown in Fig. 17. In this panel, the
leg and bracing member was divided into four equal parts and fully
triangulated hip bracing system was adopted. ISA 200  200 
20 mm for leg members and back to back connected ISA
80  80  6 mm high tensile steel angles for main bracing was
used.
The entire tower was modeled using beam-column elements.
The non-linear analysis failure pattern is shown in Fig. 18 and
the predicted failure load is 106%. The predicted capacity of leg
and bracing members based on ASCE/IS standards was 21% and
8% more than the forces obtained from non-linear finite element
analysis. The bracing member capacity predicted based on BS stan-
dard is 5% lesser than the analytical force. The analytical member
forces and capacities are given in Table 1.
The finite element analysis failure pattern shows that the leg
member was failing at second sub-panel and the bracing member
in fourth sub-panel near horizontal belt. The slenderness ratios of
all inclined redundant members and the horizontal hip bracing in
third sub-panel (H3) are close to 250 limits. The nonlinear analysis
force in the inclined redundant member (IR) (Fig. 16 Case I) in the
second sub-panel has reached its capacity and in all other inclined
redundant members the forces are nearer to their capacity.
BS standard procedure is adopted for checking the forces in
the redundant members. According to this procedure, a hypothet-
ical force shall be applied transverse to the leg member being

2800
FEM TEST ASCE and IS BS
2600

2400

2200

2000

1800
CAPACITY ( kN)

1600

1400

1200

1000

800

600

400

200

0
TOWER 1 TOWER 2 TOWER 3 : PANELS 1 & 2 TOWER 4 : PANELS 1 & 2 TOWER 5: TC & BC and TC & MC (TRIAL 2)

Fig. 23. Comparison of leg members.

68 N. Prasad Rao et al. / Engineering Structures 35 (2012) 55–70
400
FEM TEST
375
350
325
300
275
250
CAPACITY ( kN )
225
200
175
150
125
100
75
50
25
0
TOWER 3: 2nd PANEL TOWER 4: PANEL 1
Fig. 24. Comparison of bracing members.
stabilized at the node point of attachment of the secondary mem-
ber. This force varies with the slenderness ratio of the leg mem-
ber being stabilized and is expressed as a percentage of the leg
load. The hypothetical force shall be applied on one node at a
time and the effect on the redundant members is calculated by
performing linear elastic analysis as mentioned in BS Standard.
Since there is more than one intermediate node in the failed ‘K’
braced panel, the secondary bracing system is checked separately
for 2.5% of the leg load applied equally between all the interme-
diate node points at right angles to the leg member and in the
plane of bracing system, as per BS procedure. Redundant mem-
bers shall be designed for the maximum force obtained from
the above two procedures. The forces obtained by this procedure
in the inclined redundant members in the failed panel, exceeds
the respective member capacities and it is lesser than the 2.5%
force in the leg member.
As a remedial measure, the redundant member pattern in the
failed ‘K’ braced panel was changed by dividing the leg and bracing
member into five sub-panels and the tower was tested to its full
capacity in top and bottom conductor broken condition.
6.5.2. Top and middle conductor broken condition (Trial 1)
In the top and middle conductor broken condition, the tower
was subjected to the loads as shown in Fig. 16. The hip bracings
were staggered in such away that the main bracing member under-
goes major axis buckling about one-fourth length. The revised
redundant system in elevation and sizes are shown in ‘‘Detail at
P’’ in Fig. 16 Case II (Trial 1).
In top and middle conductor broken condition, at 80% load, the
bracing member in the second ‘K’ braced 7.5 m panel failed as
shown in Fig. 19. The test failure pattern shows the buckling of
main bracing member along with the hip bracings in second and
third sub-panels.
The failure load obtained through non-linear analysis is 101% of
top and middle conductor broken condition load and the analysis
ASCE and IS BS
TOWER 5: TC & BC and TC & MC BROKEN (TRIAL 1)
failure pattern is shown in Fig. 20. The finite element analysis fail-
ure pattern shows that the main K-bracing member was deforming
out of plane between first horizontal hip (H1) and second horizon-
tal hip (H2) bracings in second and third sub-panels causing the
failure about major axis on 0.30–0.35 L instead of assumed
0.25 L. The nonlinear analysis force in the inclined redundant
member (IR) in second sub-panel crosses its capacity.
The predicted capacity of bracing members based on ASCE/IS
standards was 40% more than the force obtained from finite ele-
ment analysis. The bracing member capacity predicted based on
BS standard is 32% more than the analytical force. The slender-
ness ratio of top horizontal hip bracing (H3) is close to 250 limit.
The test failure pattern clearly shows the bracing failure on its
major axis about 0.35 L between first horizontal hip bracing
(H1) in second sub-panel and horizontal redundant (HR3) in ele-
vation in third sub-panel. The middle horizontal hip bracing (H2)
and top horizontal hip bracing (H3) in second and fourth sub-
panels buckled in the out of plane mode causing the failure of
main bracing. The FE model failure pattern also shows the out
of plane deformation of these hip bracings. The capacity of the
main bracing member evaluated based on the effective length
based on the failure pattern matches with the analysis force. This
shows the insufficient restraint offered by the hip bracings to the
main K-bracing.
6.5.3. Top and middle conductor broken condition (Trial 2)
The main bracing member in the failed second ‘K’ braced panel
was changed to ISA 90  90  6 mm high tensile steel angle back
to back connected and the redundant member pattern in the failed
panel was changed in fourth and fifth sub-panels as shown in ‘‘De-
tail at P’’ in Fig. 16 Case II (Trial 2). Staggering in the hip bracing
pattern was removed and one additional hip bracing member
was connected to the main bracing member in fourth sub-panel.
The main K-bracing was designed for minor axis buckling about
0.2 L.
 N. Prasad Rao et al. / Engineering Structures 35 (2012) 55–70 69

During testing at 95% load, the leg member in the second ‘K’
braced panel failed as shown in Fig. 21. The capacity of the leg
and bracing member was 24% and 50% more as compared to anal-
ysis force. The failure load obtained through finite element non-
linear analysis is 103% and the analysis failure pattern is shown
in Fig. 22. The analysis force in the inclined redundant member
(IR) in second sub-panel exceeded its capacity. The slenderness ra-
tio of horizontal hip bracing (H3) in the third sub-panel was close
to 250 limit. As a remedial measure, the secondary bracing sizes
are increased and provided with two bolt connections on both
the ends. The tower was then tested to its full capacity in top
and middle conductor broken condition. The failure clearly indi-
cate the need for caution when assuming reduced effective lengths
coupled with maximum allowable slenderness ratios. Introducing
one more sub-panel in such cases obviates the risk involved.
Though it is a minor aspect in the design, ignoring the same can
cause failure.
6.6. Codal comparison and discussion
The member capacities are calculated based on British, ASCE
and Indian standards and given in Table 1. British standard consid-
ers full width of angle section for width to thickness (b/t) ratio cal-
culation and in ASCE standard, flat width from edge of fillet to the
extreme fiber is considered. For compound members made of two
angles connected back to back as ‘T’ section and designed for major
axis buckling, the possible additional deformations due to shear is
taken into account in British code by modifying the slenderness
ratio. In Table 1, the capacity of bracing member is maximum cal-
culated based on ASCE and IS standards. The capacity of leg mem-
ber predicted based on British standard is higher than other codes.
In Figs. 23 and 24, bar charts are drawn to compare the capacities
of members calculated by different codes with finite element and
test forces in the failed members. In the bar chart, the capacity of
bracing members predicted from ASCE standard is always higher
than the test values for all slenderness ratios. In Figs. 23 and 24
there is a wide discrepancy between code prediction and experi-
mental values since the premature failures in some of the towers
studied are mainly due to non-triangulated hip bracing pattern
and inadequate redundant member sizes and incorrect design
assumptions and over estimation of member capacities. The
275 kV dead end and 400 kV tangent towers showed failure capac-
ities far less than that predicted by NE-NASTRAN. While it may not
be possible to say with certainty this is the probable cause for
failure, the following observations tend to justify the lower failure
loads observed as compared to the value obtained by NE-NASTRAN
analysis.
(i) The force in the hip bracing at failure as per analysis was
only 1.8 kN for 275 kV tower which meant an axial deforma-
tion of only 0.1 mm. Since the bolt hole clearances at either
end is one millimeter, the deformation in the K-brace has to
increase several folds before the hip bracing becomes effec-
tive, possibly leading to reduced capacity.
(ii) The hip bracing can be considered as simply supported
between the K-bracings. The deflection of the hip bracing
is approximately 16 mm under its own weight. If this is con-
sidered as sag, then a tension force on the element works out
approximately to 7 kN. During erection the hip bracing
member could have been force fitted because of the sag.
Assuming that about 2 kN is the applied force due to force
fitting, the failure load drops to around 65–70% in both the
towers. The tension force in the hip bracing element leads
to initial imperfection on the K-bracing leading to its
reduced capacity. Triangulation would have vertical ele-
ments supporting the rather long and slender hip bracing
elements at number of points. This would lead to far less fit-
ting force.
(iii) The load deflection diagrams of 275 and 400 kV towers are
shown in Fig. 25a and b. It may not be just a coincidence that
the failure loads occur at around the same failure deforma-
tion predicted by NE-NASTRAN. Since the problem of insta-
bility is closely linked to deformation rather than forces, it
can be argued that the K-bracing panel became ineffective
beyond a deformation of 325 and 450 mm respectively lead-
ing to failure loads which are far lower than that predicted
by NE-NASTRAN.
(iv) It becomes increasingly clear that detailing deficiencies may
have significant influence on the performance/failure loads
of the towers – non-triangulation of hip bracing in this case.
Also it is clear that introduction of force fitting construction
deficiencies, imperfections, bolt hole clearances, etc., is diffi-
cult to be incorporated even in a refined model.
Finally without the test, if analysis is the only tool available,
over prediction of strength is seen in these two towers. Only full
scale testing of the towers could show such un-conservatism. Also
it is pertinent to point out that the towers withstood the full

100 100

90 90 FE
PERCENTAGE OF LOAD

FE
PERCENTAGE OF LOAD

80 80

70 70

60 60
TEST
50 50
TEST
40 40
30 30
20 20
10 10
0 0
0 25 50 75 100 125 150 175 200 225 250 275 300 325 350 0 50 100 150 200 250 300 350 400 450 500
DEFLECTION ( mm ) DEFLECTION (mm)
,
(a) 275 kV Dead end tower ( ‘H type) (b) 400 kV Tangent tower (‘DA’ type)
Fig. 25. Load deflection behaviour.
70 N. Prasad Rao et al. / Engineering Structures 35 (2012) 55–70
capacity with the triangulated hip bracing system. May be there
are other possible explanations. But the conclusive evidence based
on the tests is that non-triangulated hip bracing is ineffective.
7. Conclusion
Different types of premature failures and the reasons are dis-
cussed in detail. The study indicated that the nonlinear static anal-
ysis is essential for understanding the behaviour, probable load
carrying capacity, design deficiencies and instability in the struc-
ture. Based on the experimental, analytical and theoretical investi-
gation carried on five different types of towers, the following
conclusions are drawn.
 The capacity of bracing member predicted from ASCE standard
is always higher than the test values for all slenderness ratios.
Where as the British Standard predicts the capacity of bracing
members closer to the experimental values. The predicted
capacity of single angle leg member based on British standard
is higher than the experimental values.
 The member forces predicted from nonlinear analysis are higher
as compared to the conventional linear static analysis forces
with truss elements.
 Fully triangulated hip bracing system is effective in ‘K’ and X-
braced panels at bottom level of towers where width to height
ratio of the panels are above 1.5 and subdivided into two or
more sub-panels. The isolated hip bracings shall be avoided.
 Redundant member shall be checked based on BS code provi-
sions or it shall be designed for 2.5% of the axial force in the
main member.
 The capacity of a compound member consisting of pair of angles
connected back to back and designed for major axis buckling
predicted based on British Standard is closer to test results com-
pared to the capacity calculated based on American Standards.
This may be due to consideration of possible additional defor-
mation due to shear.
 The deficiencies in connections and member detailing is pre-
dicted in testing.
References
[1] Al-Bermani FGA, Kitipornchai S. Nonlinear analysis of transmission towers. J
Eng Struct 1992;14(3):139–50.
[2] Al-Bermani FGA, Kitipornchai S. Elasto-plastic nonlinear analysis of flexibly-
jointed space frames. J Struct Eng ASCE 1992;118:108–27.
[3] Kitipornchai S, Al-Bermani FGA. Nonlinear analysis of lattice structures. J
Constr Steel Res 1992;23:209–25.
[4] Al-Bermani FGA, Kitipornchai S. Nonlinear finite element analysis of latticed
transmission towers. J Eng Struct 1993;15(4):259–69.
[5] Kemp Alan R, Behncke Roberto H. Behaviour of cross bracing in latticed towers.
J Struct Eng 1998;124(4):360–7.
[6] Al-Bermani F, Mahendran M, Kitipornchai S. Upgrading of transmission towers
using a diaphragm bracing system. J Eng Struct 2004;26:735–44.
[7] Al-Bermani FGA, Kitipornchai S. Failure analysis of transmission towers. J Eng
Failure Anal 2009;16(6):1922–8.
[8] IS: 802 (Part 1/Sec1:1995 & Sec 2: 1992) Use of structural steel in over head
transmission line towers- code of practice (Materials, loads and permissible
stresses), Bureau of Indian Standards, New Delhi, India.
[9] Design of Latticed Steel Transmission Structures, ASCE Standard, ASCE 10-97,
Reston, Virginia.
[10] Lattice Towers and Masts: Part 3 Code of practice for strength assessment of
members of lattice towers and masts: British Standard BS:8100 – 3:1999,
British Standards Institution, London.
[11] Prasad Rao N, Kalyanaraman V. Nonlinear behaviour of lattice panel of angle
towers. J Constr Steel Res 2001;57:1337–57.
[12] Test Report of 132 kV D/C TDL, tower +12m Extn., CNP 1090 41, March, 2004,
SERC, Chennai.
[13] Loading tests on overhead lines structures, IEC 60 652, 2002, Geneva.
[14] Test Report of 400 kV D/C Type ‘DB+9’ tower, CNP 1101 41, October, 2004,
SERC, Chennai.
[15] IS: 802 (Part III)-1978, Use of Structural Steel in Over Head Transmission Line
Towers-Code of Practice, Part III Testing, Bureau of Indian Standards, New
Delhi, India.
[16] Test report on 275kV D/C type ‘275H +9 m‘tower CNP.1094 41, SERC, Chennai,
December, 2003.
[17] Test report on 400kV D/C type ‘DA +9 m‘(0–20) tower CNP. 1033 41, SERC,
Chennai, September, 2000.
[18] Test Report of 400 kV D/C Type ‘DC+9’ tower, CNP 1107 41, October, 2005,
SERC, Chennai.