Home Search Collections Journals About Contact us My IOPscience

Failure analysis of lattice tower like structures

This content has been downloaded from IOPscience. Please scroll down to see the full text.

2017 IOP Conf. Ser.: Earth Environ. Sci. 80 012024

(http://iopscience.iop.org/1755-1315/80/1/012024)

View the table of contents for this issue, or go to the journal homepage for more

Download details:

IP Address: 185.158.104.208
This content was downloaded on 09/08/2017 at 13:36

Please note that terms and conditions apply.

ICCIEE 2017 IOP Publishing
IOP Conf. Series: Earth and Environmental Science1234567890
80 (2017) 012024 doi:10.1088/1755-1315/80/1/012024

Failure analysis of lattice tower like structures

RaghavanRamalingam
Civil Engineering Department, SASTRA University, Thanjavur, Tamil Nadu, India

*Email: [email protected]

Abstract. The Experimental investigations have revealed significant mismatches between

analytical estimates and experimentally measured deflections of tower structures. Conductor
and groundwire tension and sag are influenced by the peak and cross-arm deflections and
hence accurate estimates of the tower deflections at service loads are of key interest. This paper
presents a nonlinear formulation for analysis of tower structures in an attempt to close the
mismatch. The analysis includes geometric nonlinear behaviour of the tower, material
nonlinearity as well as leg member buckling in compression. The nonlinear analysis reveals
that the ultimate collapse loads of the towers tested to failure can be predicted. However the,
deflections continue to have a mismatch despite nonlinear analysis. This demonstrates the need
for inclusion of other parameters to be able to reduce the error of analytical deflection
estimates.

1. Introduction
Lattice tower like structures are commonly encountered in the form of communication tower masts
and transmission line towers. The property of tall heights compared with their base widths, while
having limited bending action due to the lattice or truss form, make these structures unique in
behaviour. The various loading conditions on towers cause specific deflection shapes and failure
locations which are further dependent on properties of both members and connections. Towers are
designed traditionally for reliability and security requirements by ensuring the strength limit states for
individual members and connections. The designs however cannot ensure the safety against ultimate
collapse of the tower as a whole. The ultimate collapse of towers are mandatorily studied through full-
scale tests by clients as specified in the testing codes (IEC 60652: 2002 and IS 802:1995 – part 3).
Active research on the ultimate collapse and deflection performance of transmission line towers has
aimed to address several mismatches existing between test observations, analytical procedures and
design provisions.
Initial research on the behaviour of transmission line towers (Al-Bermani and Kitipornchai, 1992)
employed a nonlinear formulation that considered the members as thin-walled beam column elements
to capture the geometric nonlinear effects. Satisfactory predictions of collapse loads were not matched
by satisfactory deflection predictions, which the authors attributed due to the joints. A similar
procedure was proposed to replicate and reduce the need for full-scale tests (Al-Bermani et al, 2009).
The effect of hip bracing patterns on the failure of transmission towers was studied analytically with
additional suggestions for appropriate bracings to adjust slenderness of leg members in different types
of towers (Rao et al, 2010). A common discrepancy highlighted in the above studies is the inability to

Content from this work may be used under the terms of the Creative Commons Attribution 3.0 licence. Any further distribution
of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.
Published under licence by IOP Publishing Ltd 1
ICCIEE 2017 IOP Publishing
IOP Conf. Series: Earth and Environmental Science1234567890
80 (2017) 012024 doi:10.1088/1755-1315/80/1/012024

match deflection predictions with values measured from the full-scale tests. The primary cause is
stated to be the occurrence of bolt slippage in the splices provided for joints of leg members. Two bolt
slippage models were suggested – sudden and continuous slip (Kitipornchai et al, 1994). A detailed
thesis on the analysis of tower structures with consideration of bolt slippage was presented by Kroeker
(2000). Component level studies of bolt slippage in joints are also present
present in literature for further
reference (Ungkurapinan et al, 2003). Modified analytically estimated deflections have also been
compared with full-scale test measurements of deflections (Rao et al, 2012). Provisions for
experimentally obtaining the bolt slip
slip portion of overall deflections are present in the testing codal
standards.

2. Tower modelling and design

In this paper, as truss elements were adopted for the tower models, the secondary bracings were
omitted and checked for 1.5-2% of the axial force in the main leg members. This was to avoid joint
instability which occurs at the joints of secondary bracing members (Fig. 1) due to no out-of plane
restraint with respect to the plane of the secondary bracing. The leg members and main bracings
should comply with the permissible stresses and clauses present in the design codes (IS 802:Part 1 –
Section 2; ASCE Manual 52 – Guide for design of steel transmission towers; IEC 60826, 2003).
Criteria for deflections of towers are not specified in either of the IEC codes or the Indian Standards. It
is reported that erstwhile USSR codes contained provisions to limit tower deflections. This clause has
subsequently been made obsolete given that deflections estimated by structural analysis do not match
actual full-scale test deflections. Thus it would be erroneous for a designer to set deflection limits in
reference to such a clause. This sharply highlights the necessity for improvements in analysis
procedures of transmission line or lattice towers in general. This study aims to attempt to close this
mismatch in deflections at the service load levels using nonlinear analysis of towers with inclusion of
member yielding and buckling.

Figure 1. Joint instability at secondary bracing joints

3. Nonlinear analysis of lattice towers

In this study a formulation based on the corotated (CR) approach (Mattiasson, 1983) was employed
within the updated Lagrangian form of nonlinear finite element equations. This formulation was used
owing to the fact that, for truss elements in a CR aapproach,
pproach, the local tangent stiffness matrix is 1×1

2
ICCIEE 2017 IOP Publishing
IOP Conf. Series: Earth and Environmental Science1234567890
80 (2017) 012024 doi:10.1088/1755-1315/80/1/012024

instead of 6×6. This is a consequence of the separation of strain causing deformations from rigid body
motions of the total deformation in the CR approach. After applying transformations from local to

 =  
′ +
  
global coordinate system, the global tangent stiffness matrix is given by
(1)

where  ,  and  are transformations matrices described in reference (Mattiasson, 1983),

 is
the element force at the single degree of freedom after applying corotation and  is the change in
global displacement vector. The local tangent stiffness
′ (in the corotated coordinate system) is
obtained using nonlinear finite element equations in the updated Lagrangian framework. But the
displacement function in this equation contains just one displacement due to the application of
corotation (instead of displacements at two nodes in traditional displacement function of truss

The local tangent stiffness,
′ can be written as

elements).



′ =  + 

(2)

where
′ is the element local tangent stiffness and is a single value matrix due to the use of the
corotated approach and  is the element stress vector.
The above equations handle geometric nonlinear analysis. Inclusion of material nonlinearity and
member buckling require modifications to the stiffness given by Eq (2). Member buckling is
considered by including an initial imperfection as a small fraction of the member length (0.1%) to
estimate the reduced stiffness of buckled members. This is achieved through the arc length method
(Crisfield, 2000) which traces the load-deflection of members undergoing buckling. The tangent
stiffness expression then includes the form lateral displacement component due to buckling

1
(Jayachandran et al, 2004) as shown below

= 
(3)

 + $
 !∆ #




where, with % as the maximum lateral deflection

1
&∆ = '1 − /
(4)

) , )
1 + * + #- .

Material nonlinearity is taken as the inelastic behaviour of the members and a mixed hardening
model is considered, though for static analysis a flat yielding curve would be sufficient. The state

stress component  in Eq (2) to account for material nonlinearity.

determination procedure (Bhatti, 2000) and von-Mises yield criterion for mixed hardening modify the

01 = 234′
δ5 = 1 − 234 ′
(5)

6 =  − 5 − 01

In the above equations, 01 and δ5 denote the change in subsequent yield stress due to hardening
and the stress shift vector respectively. 2is the mixed hardening parameter (Axelsson and
Samuelsson, 1979), 3 is the hardening slope and 4 ′ is the increment of plastic strain which is
evaluated within the formulation. This combined geometric and material nonlinear analysis procedure
is incremental-iterative. The results obtained using the formulation with elastic analysis and no
member buckling is compared with results from ABAQUS to verify the formulation. However,

3
ICCIEE 2017 IOP Publishing
IOP Conf. Series: Earth and Environmental Science1234567890
80 (2017) 012024 doi:10.1088/1755-1315/80/1/012024

modelling member buckling in commercial FE software is tedious and this is an advantage of the
formulation.

4. Results and discussion

Three examples of towers from literature are included in this study. The first example is the double
circuit tower shown in Fig 2 (units in kN,mm). The tower was modelled and analysed in ABAQUS
(Fig 2) considering only elasticity and this result was used to validate the formulation in the present
study for nonlinear analysis of tower structures. The estimates of deflections at the groundwire (GW)
and bottom panel (BP) joints on the leeward side and the comparisons of deflections with ABAQUS
results are shown in Fig 3(a) and Fig 3(b). Also, the stiffness shown by the current study is 0.93 times
that of commercial FE software.

Figure 2. Double circuit tower

4
ICCIEE 2017 IOP Publishing
IOP Conf. Series: Earth and Environmental Science1234567890
80 (2017) 012024 doi:10.1088/1755-1315/80/1/012024

25 10

20 8

Load Factor
Load Factor

15 6

10 Elastic-GW 4 ABAQUS - GW
Inelastic - GW Present study - GW
5 Elastic - BP 2
ABAQUS - BP
Inelastic-BP
Present study - BP
0 0
-0.2 1E-15 0.2 0.4 0.6 0 0.05 0.1 0.15 0.2
Deflection (m) Deflection (m)
Figure 3. (a) Load-deflection comparisons of Figure 3. (b) Comparison of present study with
double circuit tower at groundwire and bottom ABAQUS results
panel

The second example is the communication mast (Tanaka et al, 1985) in Fig 4(a) studied under
vertical loads on the top nodes and an imperfection load of 0.1% of any one vertical load along the
transverse direction. The axial (lb) and flexural rigidities (lb-in2) for the members are: 7.08 × 106 and
2.16×106 for longerons, 2.7×105 and 6.43×104 for diagonals and battens, 1.65×108 and 2.2×108 for
short longerons, 1.37×108 and 1.52×108 for plan diagonals. Fig 4(b) shows the load-deflection
comparison with consideration of both member inelasticity and buckling. The softening part of the
curve of elastic analysis in the study is due to the imperfection load which makes the overall tower
resemble a slender column and this is not reflected in the reference curve. Member buckling is governs
the ultimate collapse of the tower.
10000

8000

6000
Load (lb)

4000
Ref (Elastic)
Elastic
2000
Inelastic
Ref (buckling)
0 Buckling
0 2 4 6 8
Horizontal deflection (in)

Figure 4. (a) Communication mast Figure 4. (b) load-deflection of communication mast

5
ICCIEE 2017 IOP Publishing
IOP Conf. Series: Earth and Environmental Science1234567890
80 (2017) 012024 doi:10.1088/1755-1315/80/1/012024

The final example is the 33 kV AT type tower (Rao et al, 2012) shown in Fig 5. The linear analysis
from the reference and the ABAQUS results do not capture the ultimate collapse load level which
occurs at a load factor of above 1.0 (GW – ground wire peak, C1 to C3 – top cross arm to bottom cross
arm) from the current study in Fig 6. This is predicted at load factor 1.4 in analysis by buckling of
bottom leg members (at 1.15 load factor in experimental test). Also, the stiffness shown by the current
study is 0.87 times that of commercial FE software. The measured deflection at GW peak reported in
the paper was 350mm, while that obtained from analysis is 257 mm (an error of 26%). This shows that
despite nonlinear analysis, the error in deflection estimates is not bridged.

Figure 5. 33 kV AT type tower.

6
ICCIEE 2017 IOP Publishing
IOP Conf. Series: Earth and Environmental Science1234567890
80 (2017) 012024 doi:10.1088/1755-1315/80/1/012024

0.8
Load factor

0.6

0.4 GW
C1
0.2 C3
C5
0
0 50 100 150 200 250
Deflection (mm)

Figure 6 Load-deflection of 33 kV AT type tower

5. Conclusions
The study presents an efficient formulation for the nonlinear analysis of lattice towers based on the
corotated approach and is verified through comparisons with an elastic analysis of the towers in
commercial FE software ABAQUS. Determination of ultimate collapse loads of the towers is enabled
by inclusion of material nonlinearity and reduced stiffness of buckled members. Experimentally
measured values of ultimate load and deflections in literature showed that the ultimate loads are
predicted with reasonable accuracy through nonlinear analysis. Also the cause and location of failure
are predicted from the analysis. However, the analytically estimated deflections are at considerable
error in relation to the test values inspite of nonlinear analysis. Thus, other effects like bolt slip and
connection stiffness are necessary to enable accurate estimates of deflections through analysis. This
necessitates component level studies for joint effects which can be incorporated into the overall
analysis of the tower.

References
[1] Al-Bermani FGA and Kitipornchai S 1992 Eng. Struct.14(3)139-51.
[2] Albermani F, Kitipornchai S, Chan RWK 2009Eng. Fail. Anal. 161922–8
[3] Rao NP, Knight GMS, Lakshmanan N and Iyer NR 2010Eng. Fail. Anal.171127-41.
[4] Kitipornchai S, Al-Bermani FGA and Peyrot AH 1992J. Struct Eng ASCE120(8)2281-7.
[5] Kroeker D 2000Structural analysis of transmission towers with connection slip modelling MS
thesis, University of Manitoba.
[6] Ungkurapinan N, ChandrakeerthySRDeS, Rajapakse RKND and Yue SB 2003Eng.
Struct.25779-88.
[7] Rao NP, Knight GMS, Lakshmanan N and Iyer NR 2012Pract. Period. Struct. Des. Constr.
ASCE.17(2)65-73.
[8] Mattiasson K 1983On the corotational finite element formulation for large deformation
problemsDrIng thesis, Chalmers university of technology.
[9] Jayachandran SA Kalyanaraman V and Narayanan R 2004 Int. J. Struct. Stab. Dyn.4(1) 1-19.
[10] Crisfield, MA 2000 Non-linear Finite Element Analysis of Solids and Structures, Volume 1:
Essentials John Wiley & Sons, Chichester.

7
ICCIEE 2017 IOP Publishing
IOP Conf. Series: Earth and Environmental Science1234567890
80 (2017) 012024 doi:10.1088/1755-1315/80/1/012024

[11] Bhatti MA 2006 Fundamental finite element analysis and applications: with Mathematica and
Matlab computations John Wiley & Sons.
[12] Axelsson K and Samuelsson A1979Int. J. Numer. Methods Eng.14(2) 211-25.
[13] Tanaka K, Kondoh K and Atluri SN 1985 Finite Elem. Anal. Des., 1, 291-311.