NUMERICAL SIMULATION OF STRUCTURAL

BEHAVIOUR OF TRANSMISSION TOWERS

FGA Albermani1 and S Kitipornchai2

Department of Civil Engineering, University of Queensland,

Department of Building and Construction, City University of Hong Kong,

ABSTRACT

Transmission towers are a vital component and management needs to assess the reliability and safety
of these towers to minimise the risk of disruption to power supply that may result from in-service
tower failure. Latticed transmission towers are constructed using angle section members which are
eccentrically connected. Towers are widely regarded as one of the most difficult form of lattice
structure to analyse. Factors such as fabrication errors, inadequate joint details and variation of
material properties are difficult to quantify. Consequently, proof-loading or full-scale testing of towers
has traditionally formed an integral part of the tower design. Stress calculations in the tower are
normally obtained from a linear elastic analysis where members are assumed to be axially loaded and,
in the majority of cases to have pinned connections. In practice, such conditions do not exist and
members are detailed to minimize bending stresses. Despite this, results from full-scale tower test
often indicated that bending stresses in members could be as high as axial stresses. EPRI (1986)
compared data from full-scale tests with predicted results using current techniques and concluded that

the behaviour of transmission towers under complex loading condition cannot be consistently predicted
using the present techniques. They found that almost 25% of the towers tested failed below the design
loads and often at unexpected locations.

Furthermore, available test data showed considerable

discrepancies between member forces computed from linear elastic truss analysis and the measured
values from full-scale tests. The paper describes a nonlinear analytical technique to simulate and
assess the ultimate structural response of latticed transmission towers. The technique may be used to
verify new tower design and reduce or eliminate the need for full-scale tower testing. The method can
also be used to assess the strength of existing towers, or to upgrade old and aging towers. The method
has been calibrated with results from full-scale tower tests with good accuracy both in terms of the
failure load and the failure mode. The method has been employed by electricity utilities in Australia
and other countries to: (a) verify new tower design; (b) strengthen existing towers, and (c) upgrade old
and aging towers.

KEYWORDS

Angle member, computer simulation, lattice structure, full-scale test, nonlinear analysis, numerical
method, thin-walled structure, transmission tower, ultimate strength

INTRODUCTION

Overhead transmission lines play an important role in the operation of a reliable electrical power
system. Transmission towers are a vital component and management needs to assess the reliability and
safety of these towers to minimise the risk of disruption to power supply that may result from inservice tower failure. One of the problems facing tower designers is the difficulty in estimating wind
loads as they are based on a probabilistic approach. Another is tower strength which in contrast, could
be deterministic provided a proven-reliable analytical tool is available for the specified design load
conditions. In practice, factors such as fabrication errors, inadequate joint details and variation of
material properties are difficult to quantify and they are often used to justify the use of full-scale tower
testing. Strictly speaking, however, test results are only valid for the particular tower under the
particular test loading conditions, and they may not predict exactly how a tower may behave in practice
under different loading conditions.

This paper describes a computer simulation technique for predicting the ultimate structural behaviour
of self-supporting and guyed latticed transmission towers under static loading. The technique can

predict accurately the failure load and the failure mode of towers, and may thus be used to replace or
reduce the need to carry out full-scale tower testing. The method has been employed by electricity
utilities in Australia and other countries to: (a) verify new tower design; (b) strengthen existing towers,
and (c) upgrade old and aging towers.

Three case studies are presented: (i) a calibration case study, (ii) a case study involving the
strengthening of existing towers, and (iii) a case study involving upgrading old towers.

For

commercial reasons, ownership of the towers will not be revealed.

CURRENT ANALYTICAL TECHNIQUES

Latticed transmission towers are constructed using angle section members which are eccentrically
connected. Towers are widely regarded as one of the most difficult form of lattice structure to analyse.
Consequently, proof-loading or full-scale testing of towers has traditionally formed an integral part of
the tower design. Stress calculations in the tower are normally obtained from a linear elastic analysis
where members are assumed to be axially loaded and, in the majority of cases to have pinned
connections. In practice, such conditions do not exist and members are detailed to minimize bending
stresses.

Despite this, results from full-scale tower test often indicated that bending stresses in

members could be as high as axial stresses (Roy et al, 1984). EPRI (1986) compared data from fullscale tests with predicted results using current techniques and concluded that the behaviour of
transmission towers under complex loading condition cannot be consistently predicted using the
present techniques. They found that almost 25% of the towers tested failed below the design loads and
often at unexpected locations. Furthermore, available test data showed considerable discrepancies
between member forces computed from linear elastic truss analysis and the measured values from fullscale tests.

NONLINEAR ANALYTICAL TECHNIQUES

In the proposed nonlinear analytical technique, the tower is modelled as an assembly of beam-column
elements. Linear, geometric and deformation stiffness matrices are used to describe the behaviour of a
general thin-walled beam-column element in an updated Lagrangian framework.

This approach

greatly reduces the number of elements required (Albermani and Kitipornchai, 1990a; 1992) for
accurate modelling of the nonlinear structural response. A lumped plasticity approach coupled with
the concept of a yield surface in force space is adopted (Albermani and Kitipornchai 1990b) for

modelling the material nonlinearity. The formex algebra approach (Albermani et al, 1992) is used for
automatic generation of data necessary for the analysis.

All of the members in the tower are modelled in the analysis, including secondary bracing members.
The technique accounts for both geometric and material nonlinearity. The geometric nonlinearity
accounts for the effects of the accumulated stresses on the structural stiffness of the elements and the
effect of the continuing changes in the geometry as the applied load is increased. Buckling of
structural members can be detected during the load application. The material nonlinearity accounts for
the effect of combined stresses on the plastification of the element cross-section. Stress-resultant yield
surfaces and a lumped plasticity approach are used for this purpose (Albermani and Kitipornchai,
1990b). The analysis can also incorporate other nonlinear effects due to joint flexibility, bolt slippage
(Kitipornchai et al, 1994) and differential support settlement.

In the analysis process, an incremental-iterative predictor-corrector solution strategy is used. Loads are
applied in small increments.

At each load increment several iterations are performed to satisfy

equilibrium and the structural geometry is constantly updated. The solution method is equipped with a
number of numerical strategies that enable prediction of any buckling or instability as well as tracing
the nonlinear load-deflection path.

The described numerical simulation technique has been used to analyse self-supporting and guyed
towers (Albermani, 1997) under specified loading conditions. Some of the towers modelled have
subsequently been tested to failure. Predicted failure loads and failure modes are in good agreement
with those obtained from tests. Over the past 10 years, this technique has been employed by a number
of electricity utilities in Australia and in some other countries.

CASE STUDIES

Case Study 1: Verifying New Tower Design

A new 330 kV double circuit suspension tower was designed and tested to failure in Australia. The
nonlinear analysis was used to verify the design and plan the test sequence prior to the full-scale
testing. The tower is shown in Figure 1. It has a square base of 12.68mx12.68m and a height of
53.4m. The self-weight of the tower was 132 kN. Eight loading conditions were specified for the

tower. The tower response was described in terms of a load factor, , which is the ratio of the applied
load to the specified design load for the particular load case.

The tower was modelled using 1557 elements and 790 nodal points. This gave a total of 4740 degreesof-freedom to model the tower response. The ultimate load factor obtained for the eight loading
conditions varied from 1.10 to 1.78, indicating that the tower would have no difficulty in passing the
full-scale test. Results from the nonlinear analysis, assuming nominal yield stresses, indicated that
failure modes of the tower were due to the spread of plasticity for the loading conditions. The tower
was full-scale tested to 100% ultimate design load, using the recommended test sequence and, as
expected, the tower passed the test under all loading conditions.

Load Case 8A (microburst wind on full tower, loads kept at 100% design loads, except for transverse
conductor and earthwire loads) was chosen as the loading condition to test the tower to failure. The
predicted ultimate load factor for this loading condition was 1.20 with failure due to the spread of
plastic hinges.

Tensile tests were conducted on various members of the tower and actual yield stresses were found to
be at least 10% higher than the assumed nominal yield stresses. The tower was then re-analysed under
Load Case 8A, assuming higher yield stresses. The nonlinear analysis predicted that the tower would
fail by buckling in the compression legs (see Figure 2(a)) at load factor 1.307. The tower was tested to
failure and reported failure at load factor 1.30 by buckling of the compression face of the tower (see
Figure 2(b)), virtually identical to the prediction. This demonstrated that the nonlinear analysis was
capable of predicting accurately both the failure load and the failure mode of the tower.

Case Study 2: Strengthening Existing Towers

A 220 kV transmission line with about 1300 suspension towers was constructed in the early 1980's.
The line experienced two major tower failures at approximately seven years apart. The cause of failure
was thought to be the higher than expected wind loads. The investigation involved analysing the tower
for higher wind loads and identifying weak areas in the tower so that suitable schemes could be
devised to strengthen the tower to reduce the potential risk of a future failure.

The geometry of a typical tower along this line is shown in Figure 3. The tower has a rectangular base
of 6.5m in the transverse direction, 3.55m in the longitudinal direction and a height of about 39.0m.
The self-weight of the tower was 36 kN. Nine new revised loading conditions were used to evaluate
the as-built tower response.

The tower was modelled using 1100 elements and 730 nodal points. This gave a total of 4380 degreesof-freedom. The ultimate load factors obtained for the nine revised loading conditions varied from
0.59 to 1.55 for the as-built tower. Results obtained from the nonlinear analysis revealed that the tower
would collapse under three of the nine loading conditions. The collapse was due to either spread of
plasticity or premature buckling. Figure 4(a) shows the tower response under one such loading
condition where the tower collapsed at a load factor of 0.74. Figure 4(b) shows the magnified
deflected shape of the tower at collapse. Under this loading condition, plastic hinges initiated at a load
factor of 0.69 in the compression leg at the lower part of the Common Body and spread downward in
this leg.

Results from the nonlinear analysis indicated that the as-built tower would fail at loads significantly
below the revised ultimate design loads. A number of possible modification schemes to strengthen the
tower were investigated. The first scheme involved adding a diaphragm at two locations with some
horizontal and secondary bracing members. The addition of these new members was introduced in
stages referred to as Upgrades 1 to 4 as shown in Figure 5. With these upgrades, the tower response
improved but still did not reach the revised ultimate design loads under two loading conditions.

The second modification scheme involved using stay wire to strengthen the as-built tower. Four stay
wires (19/3.25mm) tensioned to 25kN were attached to the tower as shown in Figure 6(a). In addition,
four ring members (75x75x5 MS) were added to the tower at the stay attachment level to help
transferring loads from the stay to the tower as shown in the same figure.

The stayed tower was re-analysed under the same nine loading conditions. A significant improvement
in the tower response was observed. However, the tower still failed to reach the ultimate design loads
under one loading condition where plastic hinges formed at the compression leg just below the stay
attachment level. In order to prevent this failure, further modifications to the stayed tower were made.
These include increasing the pretension force in the stay from 25 to 50kN and the addition of a
horizontal diagonal member at the stay attachment level and a set of ring members just below the
attachment level. These modifications are shown in Figure 6(b). With these modifications, the tower
was re-analysed and passed successfully for all the nine revised ultimate design loading conditions
with the lowest load factor being 1.03.

Case Study 3: Upgrading Old Towers

An existing 400 kV transmission line was designed and constructed almost 45 years ago. The line has
performed its function and suffered no tower failure. The nonlinear analysis was conducted to
determine if the capacity of the towers in this line could be upgraded to carry larger and heavier
conductor loads, and to devise appropriate practical upgrading schemes for the towers.

One of the towers analysed is shown in Figure 7(a). It has a square base of about 9.5mx9.5m and a
height of about 49.0m. The tower self-weight was calculated to be 127kN. Seven loading conditions
were specified based on the revised wind and incorporating larger conductor loads. The tower was
modelled using 1245 elements, 660 nodal points and 3960 degrees-of-freedoms. In the nonlinear
analysis, the vertical loads were applied first up to 100% of their specified values, followed by the
incremental application of the transverse and the longitudinal loading.

Results from the nonlinear analysis indicated that the tower did not reach the new ultimate design loads
in four out of the seven loading conditions. Load factors at failure ranged from 0.87 to 1.32. Based on
results from the nonlinear analysis including the failure pattern, the tower was upgraded by adding a
horizontal diaphragm as shown in Figure 7(b). The upgraded tower was re-analysed using the same
seven loading conditions. With this modification, the tower was able to carry the increased loads
without any difficulty with the lowest load factor being 0.99.

CONCLUSION

The paper describes a nonlinear analytical technique for simulating the ultimate structural response of
latticed transmission towers.

Accurate structural analysis of towers is complicated because the

structure is three-dimensional and comprised of angle section members eccentrically connected. The
influence of geometric and material nonlinearities plays a very important role in determining the
ultimate behaviour of the structure. The proposed technique may be used for verifying new tower
design thereby reducing or eliminating the need for full-scale tower testing, in addition to providing a
degree of design confidence. It can also be used for assessing strength of existing towers, or upgrading
old and aging towers.

Three case studies have been reported. In the first case, the technique was used to verify design prior
to full-scale tower testing which included loading the tower to failure. The method has been shown to
predict accurately both the failure load and the failure mode. In the second case, the technique was
used to strengthen existing towers and in the third case, the technique was employed to upgrade the
capacity of towers that were 45 years old. In Case Studies 2 and 3, proof-loading of the towers would
have been very difficult, if not impossible.

ACKNOWLEDGEMENTS

The paper describes a nonlinear analytical technique for simulating the ultimate structural response of
The work described in this paper has been partially supported by a grant from City University of Hong
Kong (Project No. 9030875).

REFERENCES

Albermani, F.G.A. (1997), Design verification of guyed transmission tower using nonlinear analysis,
International Journal of Space Structures, 12, 1, 43-50.

Albermani, F.G.A. and Kitipornchai, S. (1990a), Nonlinear analysis of thin-walled structures using
least elment/member, Journal of Structural Engineering, ASCE, 116, 1, 215-234.

Albermani, F.G.A. and Kitipornchai, S. (1990b), Elasto-plastic large deformation analysis of thinwalled structures, Engineering Structures, 12, 1, 28-36.

Albermani, F.G.A., Kitipornchai, S. and Chan, S.L. (1992), Formex formulation of transmission tower
structures, International Journal of Space Structures, 7, 1, 1-10.

Albermani, F.G.A. and Kitipornchai, S. (1992), Nonlinear analysis of transmission towers, Engineering
Structures, 14, 3, 139-151.

Electric Power Research Institute (1986), Structural development studies at the EPRI transmission line
mechanical research facility, Interim Report No. 1: EPRI EL-4756, Sverdrup Technology Inc.,
Tullahoma, Tennessee.

Kitipornchai, S., Albermani, F.G.A. and Peyrot, A.H. (1994), Effect of bolt slippage on the ultimate
strength of latticed structures, Journal of Structural Engineering, ASCE, 120, 8, 2281-2287.

Roy, S., Fang, S. and Rossow, E. (1984), Secondary stresses on transmission tower structures, Journal
of Energy Engineering, ASCE, 110, 2, 157-172.

LIST OF FIGURE CAPTIONS

Figure 1: Case Study 1 330 kV DC Suspension Tower

Figure 2: Predicted Failure and Test Failure of 330 kV DC Suspension Tower

Figure 3: Case Study 2 220 kV As-built Tower

Figure 4: Load-Deflection Curve and Tower Deflected Shape of 220 kV As-Built Tower

Figure 5: First Modification Scheme to Strengthen 220 kV Tower

Figure 6: Second Modification Scheme to Strengthen 220 kV Tower

Figure 7: Case Study 3 Upgrading 400 kV Suspension Tower

Fig 1

Fig 2

Fig 3

Fig 4

Fig 5

Fig 6

Fig 7