You are on page 1of 12

Sensitivity of Space Trusses to Sudden Member Loss

~
1.0 Non-composite Edge-supported Space
] Trusses
...R 0.8 \
,, The behaviour of the three space trusses included in
,,
0
-5
~
0.6 this group (Tlledn, T15edn and T21edn) is presented
~ in Figures 11 to 14. The figures show how the three
~ 0.4
...~ trusses responded to cases of gradual and sudden loss
of four internal top and bottom members. The strength
0
-5 0.2 _ _ Behaviour of perfect truss

!:""
c - - - Behaviour associated witha suddenmember failure
-'-'-'- Behaviour associated wiih a gradual member failure reductions caused by these losses are also listed in
fIl 0
1.00 Table 3.
Central displacement (o/<>l!e of longspan) It can be seen from the results presented that the
Figure 11. Effect of losing a bottom internal gradual loss of bottom members in the truss central
member on truss T11edn. region led to reductions in overall strength of up to
12%. When the same members were removed
~ 1.0, - : : : : : _ - - - -_ _
suddenly from the truss model, the maximum strength
i 0.8 reduction escalated to 21%. This could be explained

I
'0
0.6
-- --
by the more severe damage that occurred in the area
surrounding the removed member (including more
0.4 widespread buckling of nearby top chord members)
g~

'0 when sudden losses were implemented.


oS 0.2 - _ Behaviour of perfect truss
The results presented also show that losing internal
I - - - Behaviour associated witha suddenmember failure
-'-'- Behaviour associated witha gradual member failure top chord members inflicted almost the same degree
fIl O+---T"----==r==:::::;:===i==~=;===~~
of damage on the truss, with the maximum strength
reductions reaching 13% and 20%, due to gradual and
sudden top member losses, respectively. Again, when
member removal was imposed suddenly and at a high
load level, more damage resulted in the neighbouring
-:'::::-._._._._.__._._.__.-.-.-._.-._.- area and this affected the ultimate strength of the truss
adversely.
i/'
,: ", ;
./ ; , '"
-- - ......... __
No cases involving losses of diagonal members
II were included for this group of trusses (and also that
/./
of composite edge-supported trusses) since their edge
( supports made the internal force distribution in the
r - _ Behaviour of perfecttruss
- - - Behaviour associated with a suddenmemberfailure diagonal members more even than in comer-supported
_.-._.- Behaviour associated with a gradualmemberfailure trusses. This characteristic resulted in a reduced
0.25 0.50 0.75 1.00 1.25 1.50 sensitivity of space trusses to diagonal member losses,
Centraldisplacement (%geof long span) and for this reason, no single diagonal member could
Figure 13.Effect of losing a top internal member on be seen as particularly critical to truss integrity.
truss T15edn.
g 1.0.+-_ _ Composite Edge-supported Space
] Trusses
i 0.8 Some of the behaviour comparisons obtained for

1 '0
0.6
trusses Tlledy, T15edy and T21edy, all composite
and with edge supports, are presented in Figures 15
~ 0.4 and 16. The figures show how the trusses responded to
B
'0
_ _ Behaviourof perfect truss
___ Behaviourassociated willi cases of individual gradual and sudden member loss.
a suddenmemberfailure

r
oS 0.2
_._._.- Behaviourassociated with Also, see Table 3 for a list of the strength reductions
f Il
a gradualmemberfailure induced by the member loss cases studied.
0.25 0.50 o. s 1.00 1.25 1.50 The results presented herein show clearly that
Centraldisplacement (o/<>l!C of longspan) gradual losses of internal bottom members caused
Figure 14.Effect of losing a top internal member on only a small strength reduction (up to 6%). Even when
truss T21edn. the loss of bottom members was effected suddenly, the

38 International Journal of Space Structures Vol. 12 No. I 1997


A./. El-Sheikh

strength reductions did not exceed 11 %, indicating to the dynamic effects associated with sudden member
that composite edge-supported trusses possessed a low losses, and that these in effect release large forces into
sensitivity to both gradual and sudden bottom member a structure that has already been weakened by the loss
losses. of one of its critical members.

:;·
8 1.0 40o/ll.. - Non-composite comer-supported trusses
t
....0
0.8
- - - - Composite corner-supported trusses
- - - Non-composite edge-supported trusses
oS .5-
.... 30o/ll..
•••••••••. Composite edge-supported trusses

.f .g
0.6

,W
i
...
0.4
....~0
_ _ Behaviourofperfcct truss
___ Behaviour associated with '!~ 20%- I
I
0.2 a suddenmember failure

-
_._._.- Behaviour associated with

~ ~--1
a gradual member failure
i
<II
0 ]0%..
U')
0.25 0.50 0.75 1.00 1. 5 1.50
Central displacement (%geoflong span)

Figure 15.Effect of losing a bottom internal O%,..J--r---T---T----i--~_


i I, I
.............-
member on truss T11edy. 1:1 1.5:1 2:1 I:] 1.5:] 2:]
ij 1.0 _ Trusses witha critical Trusses with a critical
]8. 1
=:...-----------:-- bottom member removed top member removed

'0
0.8
,-".
,.".- , ------
-'-'-'-'-'-'-'-'-'_._'-'-: :':-~-:':-: '-=-'-
Aspect ratio
f
oS
0.6 .... VI ," ,,"
Figure 17.Maximum strength reductions due to
gradual member losses.
.... r ,
i 0.4 ! ".,
....o
oS 0.2
Y _ _ Behaviour of perfecttruss
___ Behaviour associated with
a suddenmember failure
- Non-composite comer-supported trusses
- - - - Composite comer-supported trusses
I
<II
_._._.- Behaviour associated with
a gradual member failure
- - - Non-composite edge-supported trusses
-••••••••. Composite edge-supported trusses
0t--":T:"----:T:----,.--====r===:;:::==::;=
o 0.25 0.50 0.75 1.00 1.25 1.50
Centraldisplacement (%geof long span)

Figure 16.Effect of losing a bottom internal


member on truss T15edy.

The response of the three composite space trusses


to cases of top member losses was better. The trusses
showed no noticeable reduction in strength or an
alteration in stiffness even when the top members O%...J-. .- - - ; - - - i - - -__r---io....-4--
1:1 1.5:1 2:1 1:1 1.5:1 2:1
considered were removed from the truss model
Trusses witha critical Trusses with 8 critical
SUddenly. This finding was thought to be reasonable bottom member removed top member removed
due to the high in-plane stiffness of the concrete slabs
Aspect ratio
of composite trusses, and consequently, their ability to
carry almost all the top chord internal forces without Figure 18.Maximum strength reductions due to
sudden member losses.
any significant contribution from the top chord
members.
Another point that featured clearly in the results
6. DISCUSSION OF RESULTS obtained was the effect of composite action in
The results presented in the previous section reveal a substantially reducing truss sensitivity to both gradual
number of important common trends. These are and sudden member loss. While this point was
discussed in detail in this section, with the main demonstrated in a reduced truss sensitivity to bottom
conclusions briefly presented in the next section. member loss, it became particularly evident as the top
In the cases studied, sudden member losses were chord members of composite trusses turned to be
significantly more serious than gradual losses as can largely uncritical. The numerical analyses showed that
be seen from Figures 17 and 18. This can be attributed with the composite action, these members carried only

International Journal of Space Structures Vol. /2 No.1 1997 39


Sensitivity of Space Trusses to Sudden Member Loss

small internal forces and were effectively prevented sudden member losses. For instance, comer-
from buckling. supported trusses generally suffer more severe
The only case in which the composite action had strength reductions due to member losses than
failed to produce a significant positive effect was that edge-supported trusses.
involving the loss of a comer diagonal in comer-
supported trusses. This however, was because the loss 4. The composite action with top concrete slabs has
of a comer diagonal resulted in a local failure around a positive effect in reducing .space truss
a comer support, leading in effect to the complete loss sensitivity to both gradual and sudden member
of support there. The composite action, which was losses.
primarily a technique to strengthen truss top chord,
could not prevent this local failure (which involved 5. The comer diagonal members in corner-
only diagonal members), and consequently would not supported space trusses are perhaps the most
make a difference if such a vital member was lost. critical members to truss integrity. The loss of, or
The results presented also show that with high damage to, any of them can lead to a local failure
aspect ratios, space trusses became more sensitive to around the support to which they are attached,
the loss of a chord member in the main direction thereby resulting in substantially reduced truss
(which was the long direction in comer-supported strength and stiffness.
cases, and the short direction in edge-supported
cases), than to the loss of an equivalent member in the REFERENCES
perpendicular direction, see Table 3.
I. Affan, A., Collapse of Double-Layer Space Grid Structures,
It can also be seen from Table 3 that edge-supported
PhD thesis, Department of Engineering, Cambridge
space trusses were in general less sensitive to both University, UK, 1987.
gradual and sudden member loss than comer-
supported trusses. This can be related to the many 2. Schmidt, L.C., Morgan, P.R. and Clarkson, J.A., Space
more load paths available in trusses with edge Trusses With Brittle-Type Strut Buckling, Journal of
Structural Division, ASCE, Vol. 102, No.7, 1976, pp. 1479-
supports, and also to their higher degree of statical 1492.
redundancy.
In general, truss stiffness was not notably affected 3. Schmidt, L.C., Morgan, P.R., O'Meagher, AJ. and Cogan,
by chord member losses. When however, a comer K., Ultimate Load Behaviour of a Full-Scale Space Truss,
Proceedings of Institution of Civil Engineers, Vol. 69, No.
diagonal was lost, a complete local failure occurred
2,1980, pp. 97-109.
around the near support, and this was seen to affect
truss stiffness significantly. See for example Figures 8 4. Schmidt, L.C., Morgan, P.R. and Hanaor, A., Ultimate Load
and 10. Testing of Space Trusses, Journal of Structural Division,
ASCE, Vol. 108, No.6, 1982, pp, 1324-1335.

CONCLUSIONS 5. Mezzina, M., Prete, G. and Tosto, A., Automatic and


From the results of the present work, the following Experimental Analysis for a Model of Space Grid in Elasto-
conclusions can be drawn: Plastic Behaviour, Proceedings of Second International
Conference on Space Structures, Surrey, UK, 1975, pp.
570-588.
I. Although space trusses typically contain a large
number of redundant members, significant 6. Hanaor, A. and Ong, A-F., On Structural Redundancy in
reductions in their strength may result due to the Space Trusses, International Journal of Space Structures,
10;5s of a critical chord or diagonal member. Vol. 3, No.4, 1988, pp. 237-241.

7. El-Sheikh, A.I., Sensitivity of Composite and Non-


2. Sudden member losses are generally more Composite Space Trusses to member Loss, International
serious than gradual member losses. They result Journal of Space Structures, Vol. 9, No.2, 1994, pp. 107-
in large forces being shed into the structure in a 119.
dynamic manner, causing in some cases a
8. Smith, E.M., Alternate Path Analysis of Space Trusses for
widespread damage. Progressive Collapse, Journal of Structural Division,
ASCE, Vol. 114, No.9, 1988, pp. 1978-1999.
3. The way space trusses are supported has a major
effect on their sensitivity to both gradual and 9. El-Sheikh, A.I. and McConnel, R.E., Experimental Study of
Behaviour of Composite Space Trusses, Journal of

40 International Journal ofSpace Structures Vol. J2 No. I 1997


A.I. EI-Sheikh

Structural Engineering, ASCE, Vol. 119, No.3, 1993, pp. II. McConnel, R.E. and El-Sheikh, A.I., A Study of Shear
747-766. Interaction Efficiency in Composite Space Truss Systems,
First International Conference on Structural Engineering
10. El-Sheikh, A.I. and McConnel, R.E., Experimental Study of and Geotechnical Problems. Alexandria, Egypt, December
Non-Composite and Composite Space Trusses, Technical 1990, pp. 409-419.
Report CUED/D Struct!fR 137, Department of Engineering,
Cambridge University, 1991.

International Journal of Space Structures Vol. 12 No. I 1997 41


Design Verification of Guyed
Transmission Tower Using
Nonlinear Analysis
F. Albermani
Department of Civil Engineering, University of Queensland, Australia 4072

(Received 28th February 1996)

ABSTRACT: This paper describes the nonlinear analysis technique used for the
design verification of a guyed transmission tower. This tower forms part of a new
transmission line which is currently under construction. The nonlinear analysis
technique is briefly described and the findings from the analysis are highlighted
together with the resulting modifications to the original design. The findings from
the nonlinear analysis predict a catastrophic collapse of the transmission tower
based on the original design. This collapse would has resulted even if a full-scale
testing of the tower was used to verify the design adequacy.

INTRODUCTION tower. The guy cables are 20 mm in diameter and have


In the transmission towers industry, proof-loading or a specified minimum breaking force of .368 leN. A
fUll-scale testing of the tower has traditionally formed pretension force of 25 leN is specified in the design for
an integral part of the development of tower design. these guy cables. All the members in the tower are
This is particularly true when new types of towers are angle sections of either Grade 250 or 350 steel (MS or
introduced by the electricity authorities. With HTS steel).
Continuing economic pressures on electricity Instead of performing full-scale tests on this new
authorities, transmission engineers need to develop tower, which involves shipping the full tower overseas
innovative designs with a minimum of trial-and-error for testing, QEC decided to use the nonlinear static
testing and in-service experimentation I. For these analysis technique developed by the author to verify
reasons, the Queensland Electricity Commission their design. This paper briefly describes the nonlinear
(QEC) in Australia when faced with designing a new analysis used, and then highlights the necessary
275 kV single circuit transmission line chose to use modifications to the original design which resulted
guyed towers. Figure 1 shows the initial design by from the nonlinear modelling of the guyed tower. The
QEC of a 275 kV single circuit lattice guyed tower to transmission line is now under construction following
be used on this new transmission line. The mast height the implementations of all the findings of the
is 53.420 m from the ground. The mast is composed of nonlinear analysis.
a 1.8 m high base pyramid which is pinned to the
grOund and a 1.4 x 1.4 m square shaft. Two crossing NONLINEAR ANALYSIS METHOD
guy cables are attached to the bottom cross-arm on This section describes briefly the nonlinear analysis
each of the front and the back faces of the tower and technique used in simulating the tower response under
are anchored on the ground. The guy cables are not static load. For a more detailed description, the reader
joined together at the crossing points. They are should consult Refs. 2-4.
attached to the cross-arm and not to the mast's legs in The analysis performed is a geometric and material
order to increase the overall torsional resistance of the nonlinear analysis. An updated Lagrangian

International Journal of Space Structures Vol. /2 No. I 1997 43


Design Verification ofGuyed Transmission Tower Using Nonlinear Analysis

formulation is adopted. The geometric nonlinearity of MODELLING OF THE GUYED TOWER


each structural element is accounted for by The guyed tower shown in Figure 1 is modelled using
augmenting the element stiffness matrix, kit by 1300 elements and 926 nodal points. Beam-column
geometric, kg, and deformation, kd, stiffness matrices. elements are used to model the mast's legs and
The geometric stiffness matrix introduces the effect of horizontal diaphragms, truss elements are used to
initial stresses, while the deformation stiffness matrix model the bracing members, and cable elements are
introduces the necessary membrane-flexural coupling used to model the guy wires. The mast is supported by
and reduces the order of discretisation required for the a pin support on the ground.
nonlinear analysis. The resulting tangent stiffness
matrix, k., of the element is

(1)

The geometric nonlinearity of the tower is accounted


for by continuous updating of the tower geometry
following each loading cycle.
In order to account for the nonlinearity due to the
sag of the guy cables, a modified modulus of elasticity,
Ee, is used for the cable elements!

E
s, = AE(wH)2 (2)
1+----
12 T3

in which E, A and w are the elasticity modulus, cross-


sectional area and weight per unit length of the cable
respectively, T is the cable tension and H is the
projected length of the cable on a plane normal to the
gravity direction.
For large-scale structures such as transmission
towers, modelling of material nonlinearity based on
the assumption of lumped plasticity, coupled with the
concept of a yield surface in force space, provides a
compact and practical method for modelling the
nonlinear global structural responses-.
A predictor-corrector incremental-iterative solution
method is employed. A substantial saving in
computing time and a more stable algorithm is
achieved by keeping the predictor part of the
algorithm (i.e. the tangent stiffness) as simple as
possible while a strict force recovery procedure
(corrector part) is enforced. An out-of-balance force
convergence criteria is adopted using the Euclidean
norm measure with a convergence tolerance set to 5%.
The self-weight of the tower and the pretension
forces in the cables are applied in the first loading
cycle prior to the incremental application of the
extemalloads (wind load and conductors load).
The tower's topology, geometry, loading and
supporting conditions are generated automatically
using a formex formulatione-'. Figure 1.The original tower.

44 International Journal of Space Structures Vol. 12 No. I 1997


F.G.A. Albermani

Six loading conditions (T I to T6) are identified for RESULTS FROM THE NONLINEAR
the nonlinear analysis. Figure 2 shows some of these ANALYSIS
loading conditions and the locations of the deflection
stations used to monitor the tower response. (a) Original Tower (Figure 1)
The tower weight, which is automatically generated Using the original tower design as supplied by the
by the program, is 41.160 kN. Different factored self- QEC and loading condition T1 (Figure 2a), the
weights of the tower are used for different loading nonlinear analysis predicted buckling in the lower
conditions as shown in Figure 2. diaphragm at the top of the base pyramid at a load
The tower response is described in terms of a load factor of A = 0.43, as shown in Figure 3a and 3b. The
factor Awhich represents the ratio of the applied load load-deflection curve for LX 1 is shown in Figure 3c.
during the analysis to the design load shown in Figure As seen in Figure lc, excessive longitudinal deflection
2. The following nomenclature is used to describe the started to develop once the load factor reached 0.4.
deflection at various points on the tower (Figure 2d); A closer examination of the results revealed that a
deflection in the transverse direction is designated as large moment had developed at the lower diaphragm
T, in the longitudinal direction as L and in the vertical at this load level (A = 0.43). As shown in Figure 4a.
direction as V. For example, a deflection TCR means the slope change of 21 degrees in the leg members had
deflection at CR in the transverse direction. resulted in a large shear force in the lower diaphragm.
Accordingly, the base pyramid was modified as shown

11.278 16.402
1
6.094
---,.-- 1 11.313

48.543
21.665 1
3200 Pa
148.543 148.543 at~
3200 Pa
21.665 21.665

Deadloadfactor.2

Deadloadfactor.2

(e)
(a) Cl

8.458 12.301
3.047 1 1 5.656 8
---.,r--
27.857
al an
8.192 1
36.4071 ~5-.3-60---1 136.407
Xl
1600 Pa
10.833 10.833

Dead loadfactor .1.5


v~.\
(d)
(b)
Figure 2. Loading Conditions and Nomenclature.

International Journal of Space Structures Vol. 12 No.1 1997 45


Design Verification of Guyed Transmission Tower Using Nonlinear Analysis

The tower geometry and topology were altered


according to this modification. Figure 5 shows the
modified tower. This modified tower is referred to as
Ml.

(b)

Figure 3. The original tower under load condition


T1.
1400mm
./ 1400mm

(c)

2800mm

(a) (b)

Figure 4. The original and modified lower panel.

in Figure 4b. This modification had the effect of


reducing the slope angle from 21 to 14 degrees, which
was within an acceptable limit. Figure 5. The modified tower (M1).

46 International Journal of Space Structures Vol. 12 No. 11997


F.G.A. Albermani

(bJ Modified Tower Ml


A new analysis was performed on the modified tower
MI using Loading Condition T1. The instability
problem experience in the lower diaphragm was
resolved and the tower passed the load factor A. of
0.43. However, soon after this load level, the solution
process diverged at about A. = 0.46. Initial examination
of the results revealed no buckling or yielding
anywhere in the tower. In order to understand the
cause of this divergence in the results, it is necessary
to closely examine the load-deflection curves up to
this load level.
Load-deflection curves for LCR and TCR are
shown in Figures 6a and 6b. Figure 7 shows the right Figure 7. Deflected shape of M1 under Loading
side view of the tower at load factor A. = 0.46. The Condition T1.
Sudden excessive deflection in the longitudinal 4.76KN
direction following A. = 0.4 is seen clearly in Figure 6a,
and as well, twisting of the whole tower is evident in
Figure 7. Figure 8 shows the tension forces in the
cables at A. = 0.46. It is seen from this figure that
tension in these cables was not symmetric, and hence
a twisting moment was induced in the tower which
brought about the tower collapse due to the formation
of a global mechanism.

'\.0.46

3.44KN
Figure 8. Cable tension under Loading Condition T1
at 1..= 0.46.

Another analysis was performed with the initial


pre-tension in the four cables increased from 25 to 50
kN. However, the same collapse mechanism was again
obtained but this time at a slightly higher load factor A.
of 0.54.
In order to avoid such a global mechanism it was
decided to connect together the two guy cables located
on the front and on the back faces of the tower. The
tower with this new modification is refered to as M2.

(c) Modified Tower with Connecting Guys


M2
A new trial analysis was performed using load
condition TI and an initial cable pre-tension of 25 kN.
A load factor A. = 1.145 was obtained from this
analysis. Hence joining the cables together has a very
Figure 6. Load-deflection curves under load significant influences on the tower stability.
condition T1. The load-deflection curves at TCR and LCR are

International Journal of Space Structures Vol. 12 No. I 1997 47


Design Verification ofGuyed Transmission Tower Using Nonlinear Analysis

shown in Figures 9a and 9b, respectively. It is clear


from these figures that the longitudinal deflection was
very small.
1.2

-:
... 0.8
o V
-: V

] 0.6 /" TCR


"8 V
-J 0.4 /
/'
V
0.2

o
o
V 200 400 600 800 1000 1200
Deflection (nvn)
(a)

1.2

~
\ I
Lca

0.2

o
-5 -4 -3 -2 -1 o
DefleetJon (mm)
(b)
Figure 9. Load-deflection curves for tower M2
under Load Condition 11.

This was in complete contrast to the situation when


the cables were not connected.
The results show that plastic hinges formed in the
facial diagonal members of the bottom right arm at A
= 1.05 and spread but were confined only to this arm.
The deflected shape of the tower at the ultimate load
is shown in Figures IDa to IOc. The tension in the
cables at load factor A= 1.145 is depicted in Figure 11.
It is concluded that connecting the cables on the
same faces had a profound effect on the tower
response. This is seen by comparing the side views of
the tower shown in Figures 7 and IOc. Moreover,
when the cables were connected, negligible
Figure 10.Deflected shape of M2 under Loading
longitudinal deflection was obtained.
Loading condition T3 (Figure 2b) has a
=
Condition 11 at A. 1.145.
longitudinal load component applied in addition to the
Figure 12 shows the load-deflection curves at LB.
vertical and transverse components used in T1.
Several plastic hinges started at A:::::: 1.0 in the top
A nonlinear analysis was performed using this
chord of the top arm, followed by other hinges in the
loading condition and the tower response was found to
facial diagonal members of the bottom right arm.
be satisfactory. A load factor A:::::: 1.4 was obtained.
Figures 13a and 13b show the deflected shapes of the

48 InternationalJournal 0/ Space Structures Vol. 12 No. I 1997


F.GA Albermani

1.6

1.2
r-,
- - --"'- _.- r---- -
~iJ=&
~ 1
g
... 0.8
"8
.3 0.6
<,
_. I I LB
<;
0."
<,
0.2 r-...
o I I '\
-1800-1600-1400-1200-1000 -800 -600 -400 -200 0 200
o.rllclion (mm)
Figure 12. Load-deflection curve at LB, tower M2
under Loading Condition T3.

Figure 11. Cable tension under Loading Condition 11

(a)

Figure 14. Cable tension under Loading Condition


(b) T3 at A. 1.4. =
1
0.9
/
0.8
0.7
V
./
20.6
~ 0.5
./
V
! 0." /'
0.3
/
0.2
0.1
I
II
o
-100 o 100 200 300 400 500 600 700
o.rllclion (mm)

Figure 15. Load-deflection curve at LaL, tower M2


under Load Condition T6

Figure 13. Deflected shape of M2 under Loading Pa wind at 45°. The nonlinear analysis revealed that
Condition T3 at A. =1.4. the tower reaches a load factor of A. = 0.77 under this
condition with several plastic hinges spread along the
tower at load factor A. == 1.4. Figure 14 shows the guy mast's leg members. Accordingly, these leg members
cables tension at the ultimate load. were upgraded from HTS angle 90 x 90 x 6 mm to 90
Loading condition T6 (Figure 2c) represents a 3200 x 90 x 8 mm. With this modification, the tower

International Journal of Space Structures Vol. 12 No. I 1997 49


Design Verification ofGuyed Transmission Tower Using Nonlinear Analysis

no further upgrading of the leg members was


introduced.

CONCLUSIONS
The paper shows how a reliable nonlinear analysis
technique is utilised to verify the design of large-scale
structures such as guyed transmission towers.
The nonlinear analysis identified a number of
weaknesses in the structures, which were rectified
accordingly. The results from the nonlinear analysis
prevented a catastrophic collapse of the tower which
could happen even during the full-scale testing. The
Figure 16. Deflected shape of tower M2 under Load savings in resources and time results from the
=
Condition 16 at A. 0.95. nonlinear simulation is quite substantial.

ACKNOVVLEDGEMENT
The author would like to thank the designer, Mr Max
Gardener of Queensland Electricity Commission, for
many interesting and stimulating discussions. Thanks
is also due to Dr F. Gatto of BHP Engineering for
reading the manuscript.

REFERENCES
I. Electric Power Research Institute, Structural Development
Studies at the EPRI Transmission Line Mechanical
Research Facility, EPRI EL-4756, 1986.

2. Albermani, F.G.A. and Kitipomchai, S., Nonlinear Analysis


of Thin-Walled Structures Using Least Element/Member, J.
Structural Engineering, ASCE, 116 (I), pp. 215-234 (1990).

3. Albermani, EG.A. and Kitipomchai, S., Elasto-plastic large


deformation analysis of thin-walled structures, Engng
Struct., 12 (I), pp. 28-36, 1990.
Figure 17.Cable tension under Loading Condition
=
16 at A. 0.95. 4. Albermani, F.G.A. and Kitipomchai, S., Nonlinear Finite
Element Analysis of Latticed Transmission Towers, Special
Issue on Transmission Line Structures, Engng Struct., 15
reached a load factor of A. =0.95 with 27 plastic hinges (4), pp, 259-270, 1993.
developing in the leg members just below the
5. Ernst, J.H., Der E-Modul von Seilen unter Berucksichtigung
upgraded members. The load-deflection curve for LaL
des Durchanges, Der Bauingenieur, 40 (2), 1965.
is shown in Figure 15, and the tower deflected shape
is shown in Figure 16a. An enlarged view of part of 6. Albermani, F.G.A., Kitipomchai, S. and Chan, S.L., Formex
the m~t where plastic hinges have developed is Formulation of Transmission Tower Structures, Int. 1. Space
depicted in Figure 16b. The cable tension at the Structures, 7 (I), pp. 1-10, 1992.
ultimate load factor A. = 0.95 is shown in Figure 17. 7. Nooshin, H., Formex Configuration Processing in Structural
The ultimate load factor of A. = 0.95 rather than 1 was Engineering, Elsevier Applied Science Publishers Ltd.
deemed satisfactory by the electricity authority, hence (1984).

50 International Journal of Space Structures Vol. 12 No.1 1997

You might also like