Documente Academic
Documente Profesional
Documente Cultură
3
4
1234567890
33
32
31
20
19
14
13
12
11
2
34
37
38
42
43
44
45
46
47
b,*
a
Structural Engineering Research Centre, Chennai 600036, India
Department of Civil Engineering, Indian Institute of Technology, Madras, Chennai 600036, India
48
49
50
Abstract
65
Lattice microwave towers and transmission towers are frequently made of angles bolted
together directly or through gussets. Such towers are normally analysed to obtain design forces
using the linear static methods, assuming the members to be subjected to only axial loads and
the deformations to be small. The effects of the end restraints, eccentricity of connections
and secondary bracings (redundants) on the strength of the compression members are usually
accounted for in the codal recommendations by modifying the effective length of the members
and thus the design compressive strength. Hence, forces in the redundants are not known from
the analysis and their design is empirical. In this study, non-linear analysis of angle compression members and the single panel of angle planar as well as three-dimensional lattice
frames, as in typical lattice towers, are carried out using MSC-NASTRAN software. Account
is taken of member eccentricity, local deformation as well as rotational rigidity of joints, beamcolumn effects and material non-linearity. The analytical models are calibrated with test results.
Using this calibrated model, parametric studies are carried out to evaluate the forces in the
redundants. The results are compared with codal provisions and recommendations for the
design of redundants are presented. 2001 Published by Elsevier Science Ltd.
66
51
52
5
6
53
54
55
56
57
58
59
60
61
62
63
64
67
68
1
27
28
30
129
* Corresponding author.
E-mail address: kalyan@civil.iitm.ernet.in (V. Kalyanaraman).
2
3
4
5
21
3
4
1
221
ARTICLE IN PRESS
2
3
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
1. Introduction
Microwave and overhead electric transmission line towers are usually fabricated
using angles for the main legs and the bracing members. The members are bolted
together, either directly or through gusset plates. In order to reduce the unsupported
length and thus increase their buckling strength, the main legs and the bracing members are laterally supported at intervals in between their end nodes, using secondary
bracings or redundants (Fig. 1).
The lattice towers are usually analysed assuming the members to be concentrically
connected using hinged joints so that the forces in the angle members are only axial.
Under this assumption, the forces in the redundants are negligibly small or zero and
hence are not included in the linear analysis models. However, the main legs and
the bracing members are not axially loaded and the redundant forces are not negligibly small, due to the following reasons:
The main legs are usually continuous through the joint.
516
517
5
6
518
520
519
522
521
Fig. 1.
Tower configuration.
21
3
4
1
221
ARTICLE IN PRESS
N. Prasad Rao, V. Kalyanaraman / Journal of Constructional Steel Research (2001)
3
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
5
110
6
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
Usually more than one bolt is used in the connections and hence the joints are
semi-rigid.
The angle members are normally bolted through only one of their legs and hence
the force transfer in the members is eccentric.
The joints are flexible due to the local deformation of the leg of the angles under
the concentrated bolt forces.
The towers with high electric ratings tend to be flexible and hence equilibrium
in the deformed configuration has to be considered (large deformation effects).
The compression member deformation increases the bending moments (Pd
effect).
Therefore, the angle members of the tower experience both axial force and bending
moments, even well before the tower fails. This also produces forces in the redundant
members due to their participation in overall frame action, which are not negligible
as often assumed in designs.
Roy et al. [1] studied the effects of joint rigidity and large deformation of tall
high-power electric transmission towers and concluded that these towers experiencing heavier loads are more flexible and the secondary effects are more pronounced.
Al-Bermani and Kitipornchai [2] evaluated the ultimate strength of towers considering the material (lumped plasticity) and geometric non-linearity, joint flexibility and
large deflection, using an equivalent tangent stiffness matrix for the members. They
concluded that the material and geometric non-linearity have a major effect on the
ultimate strength of towers. They attributed the larger difference between their analysis and experimental results to the bolt slippage, not modeled in the analysis. Hui
et al. [3] presented details of geometric non-linear analysis of transmission towers
to trace the load deformation behaviour, treating the main legs as beam-columns and
the bracings as truss members, using updated Lagrangian formulation.
Chuenmei [6] and Shan et al. [7] used rectangular plate elements to model the
lattice tower members, which is impractical in the analysis of full towers. Rajmane
[8] used the beam-column element with seven degrees of freedom per node
(including the warping deformation) to analyse the braced frames including the
effects of eccentricity. Stoman [9] used minimisation of total potential energy to
study the plastic stability of X-braced systems and demonstrated the restraining
effects of tension diagonals.
Experimental studies have been conducted on concentrically and eccentrically
loaded single angles [6,8,1012], planar and three-dimensional lattice frames made
of angles [9,1315] and full-scale towers [22].
It is seen that the analytical studies reported have not considered all the important
factors that may influence the behaviour of lattice towers before failure, particularly
the eccentricity of connections, and the flexibility of the joints due to the local deformation of the bolted leg of the angles. Rao and Kalyanaraman [18] presented details
of a non-linear analysis of a panel of lattice towers, considering the effects listed
earlier, which affect the tower member forces. In their study, plate elements were
used at joints and at plastic hinge locations, and beam-column elements at the rest
of the locations of members, to model the angle members in the towers.
21
3
4
1
221
ARTICLE IN PRESS
4
141
This paper initially presents details of non-linear analyses of angle members and
lattice towers made of angles, using MSC-NASTRAN. These analysis models consider all the factors listed earlier, which affect the tower behaviour, and the analysis
results are calibrated against test results. Using the model thus developed, a parametric study has been done in order to understand the effects of the various factors
that influence the strength of lattice towers and the design of redundant members.
Finally, the analysis results are compared with the empirical methods recommended
in codes of practice for the design of members. Based on this approach, a method
for designing redundants in lattice towers is recommended.
142
133
134
135
136
137
138
139
140
146
Initially, single angle compression test specimens, loaded through the centroid or
through one of the legs, are modeled and analysed. Subsequently, latticed plane
frame and space frame tests using angle members are also modeled and analysed.
These analyses help to calibrate the models used in the subsequent parametric studies.
147
143
144
145
5
6
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
21
3
4
1
221
ARTICLE IN PRESS
N. Prasad Rao, V. Kalyanaraman / Journal of Constructional Steel Research (2001)
3
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
5
191
6
192
193
194
195
196
197
198
199
200
202
201
203
204
205
206
207
208
209
211
210
212
213
214
21
21
754
748
716
702
50506
576
50506
814
730
50506
960
736
Mean
742
Standard deviation
646
674
660
688
60
85
100
L/r
330
330
330
162
139
132
159.2
131.1
109
Fy
Failure load (kN)
(N/mm2)
Test
Code
5.7
4.0
6.1
1.5
6.3
152
140
130
152
140
133
1.7
5.9
17.3
8.3
8.0
153
145
140
6.1
0.4
1.4
2.3
3.4
M2
M1
Code
M1
M3
Analysis model
M2
5
6
623
629
626
632
614
Angle
section
594
587
601
Length
(mm)
Table 1
Single angles under concentric compression [8]
579
580
578
6.1
0.4
0.8
1.6
3.9
M3
577
576
3
4
1
221
ARTICLE IN PRESS
3573
575
574
21
a
1056
1055
1049
1043
899
885
576
960
1440
625
1690
2000
395
632
790
948
1104
Length
(mm)
60
100
150
72
97
114
51
82
103
123
143
L/r
50506
50506
913
50506
927
45453
941
90906
955
90906
969
40404
983
40404
997
40404
1011
40404
1025
40404
1031
Mean
1037
Standard deviation
829
857
843
871
806
812
809
815
797
330
330
330
255
255
255
270
270
270
270
270
87.3
78.5
50.5
47.3
133.4
122.6
52.2
39.3
39.1
37.4
33.8
107
92
58
62.2
165.9
148.7
59.7
52.4
47.0
40.7
33.6
Fy
Failure load (kN)
(N/mm2)
Test
Code
5
6
Angel section
778
771
785
762
Table 2
Single angles under eccentric compressiona
763
764
761
760
92.0
75.0
47.0
42.4
122.0
111.6
45.0
37.5
35.9
32.8
29.3
94.0
79.0
56.9
45.0
131.6
122.9
45.9
39.6
39.2
34.9
32.3
90.0
78.0
58.6
49.3
135.9
121.9
51.1
43.1
41.2
39.4
33.8
22.5
17.2
14.8
31.5
24.4
21.3
14.4
33.3
20.2
8.8
0.6
18.9
9.7
5.4
4.5
6.9
10.4
8.5
9.0
13.8
4.6
8.1
12.3
13.3
7.8
5.4
M1
7.7
0.6
12.7
4.9
1.3
0.24
12.0
0.8
0.23
6.7
4.4
0.64
6.7
M2
Code
M1
M3
Analysis model
M2
3.1
0.6
16.0
4.2
1.9
0.6
2.1
9.7
5.3
5.3
0.0
3.84
5.3
M3
3
4
1
221
ARTICLE IN PRESS
3757
759
758
3
4
1
221
ARTICLE IN PRESS
8
5324
525
526
528
527
530
529
5
215
6
216
217
218
219
220
221
222
224
223
225
226
227
228
229
230
231
232
233
234
235
236
237
Fig. 2.
elements. This enabled modeling of the progressive yielding at the point of plastic
hinge formation and the subsequent failure by local buckling of the elements. At
the transition between the beam-column element and the flat-shell elements, rigid
elements were used to connect the beam column node to the nodes of the flatshell elements. Whenever the load is transferred through gussets at the ends, the
gussets and the legs of the angle over 0.2 times the length at the ends were
modeled using the flat-shell elements. Beam elements were used to represent the
bolts, connecting the gussets and the beam-columns/flat-shell elements.
Model 3 (Fig. 2(c)). In this case (M3), the entire length of the angle member is
modeled using a number of flat-shell elements. Whenever the load is transferred
through gussets at the ends, the gusset plates also are modeled using the flat-shell
elements and the connections between the gussets and the angles are modeled
using the gap elements available in MSC-NASTRAN. The bolts are modeled using
beam elements.
The non-linear analysis capability of MSC-NASTRAN, accounting for the geometric and material non-linearity, was used to analyse the models and obtain their
pre-ultimate behaviour and the limit loads. The elasticplastic material property of
steel was represented by a bi-linear model, having modulus of elasticity up to a yield
stress equal to 2.0105 MPa and 2000 MPa beyond yield stress. The incremental
load and predictorcorrector iteration under each load increment were used in the
non-linear range. The Von-Mises criterion was used to define yielding. The isotropic
hardening model was used in the post-yield range. The load increments were carried
21
3
4
1
221
ARTICLE IN PRESS
N. Prasad Rao, V. Kalyanaraman / Journal of Constructional Steel Research (2001)
250
out in 2030 steps, until the limit point was reached in the load deformation behaviour.
The test results are compared in Tables 1 and 2 with the strength evaluated based
on
the
three
MSC-NASTRAN
models.
The
percentage
error
((TheoryTest)100/Test), the mean and standard deviation of the errors also are
presented in Tables 1 and 2, corresponding to the three models (M1, M2 and M3,
respectively). It is seen that the MSC-NASTRAN model results compare well with
the test results. The model M3 comparison with the test results is the best of the
three, although model M2 is quite adequate. The model M1 has the largest mean
error among the three models, particularly in the eccentrically compressed cases. For
the further study of lattice frame models M1 and M2 are used, since the model M3
consumes a large amount of time and memory due to the large number of degrees
of freedom.
251
238
239
240
241
242
243
244
245
246
247
248
249
252
253
254
255
256
5
257
6
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
Rajmane [8] tested planar angle lattice frames, and Natarajan [13] tested planar
and three-dimensional angle lattice frames, consisting of X bracings and K bracings.
Details of the test specimens and results are presented in Fig. 3 and Tables 35. The
experimental strengths of these frames are compared with the code based and numerical analysis based strengths as discussed below.
2.2.1. Code equations
These lattice frame test results are compared with the strengths based on code
provisions by the following procedure. The member forces are obtained from a linear
elastic analysis of concentrically connected lattice truss models of the frame, as commonly done in practice. The design strength of the critical angle member as obtained
from the code provisions and linear analysis member forces are used to calculate
the frame strength. It is seen (Tables 35) that the code provisions either under- or
overestimate the actual strength of the lattice frame by as much as 18%
(conservative) to +29% (unconservative). It is clear from this study that the error in
the code based design of members, for forces obtained from the linear elastic analysis
of a concentrically connected truss model, could be high, particularly in the case of
slender bracing members.
2.2.2. Numerical analysis
The conventional assumption of hinged joints does not represent the real joint
behaviour in latticed towers. Two types of joint models given below, to represent
the bolted connections between angles in the frames, were evaluated in the numerical study.
In the rigid joint model, the flexibility of the bolt and the legs of the angle at the
joint were disregarded and the joints were assumed to be rigid by enforcing the
compatibility of translations and rotations in all the members meeting at the joint.
However, the effect of an eccentric bolted connection between members was
accounted for by using rigid elements between the bolt lines and the centroid of the
21
3
4
1
221
ARTICLE IN PRESS
10 N. Prasad Rao, V. Kalyanaraman / Journal of Constructional Steel Research (2001)
5332
533
5
6
534
536
535
Fig. 3.
538
537
Joint models. (a) Rigid bolted joint model. (b) Flexible bolted joint model.
1059
1060
1062
1061
1063
1073
1068
1078
Table 3
X-braced plane framesa
Reference
1089
Test
Code
1103
1123
1113
1133
1143
1153
1163
1168
1173
1183
1178
1188
8
117
8
106
14
177
Mean
Standard deviation
a
157.0
159.0
60.0
144.7
155.5
60.6
Model
M1
Model
M2
154.0
166.0
59.0
155.0
173.0
62.0
Code
7.8
2.2
+1.0
3.0
4.4
Element
Element
model M1 model M2
1.9
4.4
1.7
0.3
3.6
1.3
8.8
3.3
3.6
5.0
21
21
36.8
51.7
52.0
59.1
30.1
52.1
52.1
57.0
Code
37.8
56.0
57.9
PF
36.4
55.7
57.8
59.4
SF
54.7
55.1
PF
52.8
54.3
55.5
SF
55.6
55.6
58.3
SF
18
0.8
0.2
3.5
5.1
8.8
Code
2.7
8.0
11.0
7.2
4.2
PF
1.0
7.7
11.0
0.5
4.5
5.7
SF
Model M1
5.8
6.0
5.9
0.1
PF
2.0
4.0
6.0
0.0
5.3
SF
Model M1
7.5
6.9
1.3
5.2
3.4
SF
Model
M2
1446
1447
PF, plane frame model; SF, space frame model. Panel type: A*: no secondary bracing and with single bolt connection; B: one level secondary bracing;
C: two level secondary bracings; D:two level secondary bracings and corner stays.
deviation
144
105
105
96
Test
31190
1192
1191
a
1445
1437
1429
1373
1357
A*
B
1389
C
1405
D
1413
Mean
1421
Standard
1293
1325
1309
1341
1256
1270
1263
1277
L/r
bracing
5
6
1235
1239
1237
1241
1225
Panel
type
1202
1207
1212
1195
Table 4
K-braced plane and space frames [13]a
1196
1197
1194
1193
3
4
1
221
ARTICLE IN PRESS
3
4
1
221
ARTICLE IN PRESS
12 N. Prasad Rao, V. Kalyanaraman / Journal of Constructional Steel Research (2001)
31450
1451
1453
1452
1454
1458
1462
1466
Table 5
K-braced panel of a 6 m extension of a 220 kV tower [15]
Bracing L/r
1476
Test
Code
Analysis
Code
Member model
1
28.6
5.23
1486
Member model
1
1494
1508
1501
1515
1522
1529
279
280
281
282
283
284
285
286
287
288
5
289
6
290
291
292
293
294
295
296
297
298
299
300
301
191
51.6
66.4
54.3
angle members (Fig. 3, elements 1 and 3 for members A and B, respectively) and
a beam element (element 2) joining these rigid elements was used to represent the
bolts. Freedom of relative rotation of the members about the axis of the single bolt
was modeled by keeping the torsional stiffness of the beam element very low (60
mm4).
The above rigid joint model does not account for the flexibility and the local
deformation of the legs of the angles at the bolted joint. For evaluating these effects,
a finite element analysis of the joint region alone was carried out using the model
shown in Fig. 3(a). In this flexible joint model (FJM) a short segment of angles
joining at the node along with the bolts were studied. The angles were modeled
using flat-shell elements. The contact force transfer between the legs of the angles
was modeled using the gap elements, available in MSC-NASTRAN. The bolts in
the joint were modeled using a rod element.
Static analyses of the joint model were carried out to obtain the joint stiffness
considering the local deformation effects. The analyses were carried out for two
different sets of member sizes to obtain the joint stiffness values in the practical
range of member sizes. These joint analyses results were used to evolve a beam
element connecting the centroidal lines of the two angles, with an equivalent stiffnesses. The flexural stiffnesses of the connecting equivalent beam elements are given
in Table 6. The equivalent link elements were used in the full frame model, to
represent the joint flexibility and eccentricity. Such full frames with equivalent beam
elements corresponding to the flexible joint model are referred to as FJM. The FJM
has been used in the analyses of K-braced frames only.
1537
1538
1540
1539
1541
1549
1545
1553
Table 6
Joint flexibility model results
Angle member
1562
1567
1577
1572
1582
Leg
Bracing
1587
90908
45453
45453
45453
1592
1597
1.78108
1.93107
10,257
621
21
3
4
1
221
ARTICLE IN PRESS
N. Prasad Rao, V. Kalyanaraman / Journal of Constructional Steel Research (2001) 13
5340
541
542
544
543
546
545
302
303
304
305
306
307
308
309
5
6
310
311
312
313
314
315
316
317
318
319
320
321
322
Two types of models were used to represent the angle members in the frame, as
discussed earlier. In one, the entire length of the angle is represented by a number
of beam-column elements (model M1). In the second model 20% of the central
length of the angle compression members and 20% of the length closer to the joints
in the angle tension members were modeled using flat-shell elements (model M2),
as discussed earlier.
The non-linear analyses were carried out assuming an initial bow of member
length/1000 in a few cases, to study their effects. The eccentricity of connections
had greater influence than the initial bow in these frames. The K-braced latticed
space frames were tested for different patterns of secondary bracings and in the
analytical model of these frames, the different secondary bracing patterns were represented.
Only plane frame analyses were carried out in X-braced frames, whereas K-braced
frames, tested as three-dimensional lattices, were analysed as both plane and space
frames. The angle member models (M1) and (M2) were used in the case of the
flexible joint model of space frames and only the angle member model (M1) was
used in the case of the rigid model. Typical analytical models are shown in Figs. 4
and 5. Some of the failure mode shapes are shown in Fig. 6.
The strength of the frames as obtained for different frames tested and different
nonlinear analysis models are presented in Tables 35. These non-linear analysis
results when compared with the test results indicate the following:
548
549
550
552
551
554
553
Fig. 5.
21
3
4
1
221
ARTICLE IN PRESS
14 N. Prasad Rao, V. Kalyanaraman / Journal of Constructional Steel Research (2001)
5356
557
558
560
559
561
562
563
323
324
325
326
327
328
329
330
331
333
332
5
334
6
Fig. 6. Three-dimensional model of 6 m extension of a 220 kV tower [15]. (a) Secondary bracing
pattern (I); (b) secondary bracing pattern (II).
The numerical analyses results for X- and K-braced lattice frames compare well
with the test results. The maximum error is 8%.
The flexible joint frame models generally compare better with test results in the
case of K-braced frames.
There is not much of a difference in the results obtained using the two angle
member element models, M1 and M2.
The mean and the standard deviation of the error in numerical analysis results
are less than 5%.
339
The results of the finite element analysis using member model M1, considering
the eccentricity and flexibility of connection as well as material and geometric nonlinearity, compare fairly well with the test results. Hence this model is used for
parametric studies in the following sections without incurring the high expenses of
experimental studies.
340
335
336
337
338
347
The secondary bracing members are provided to reduce the unsupported length
and thus increase the buckling strength of the main compression members. Linear
elastic analysis of lattice towers with secondary bracings, assuming the member connectivity to be concentric and hinged, would normally indicate zero or near zero
force in the secondary members. Hence no force for the design of secondary bracings
can be obtained from such analyses. However, secondary bracings should have some
minimum strength and stiffness to perform intended functions.
348
341
342
343
344
345
346
349
350
Codes of practice suggest provisions for the design of the secondary bracings as
given below.
21
3
4
1
221
ARTICLE IN PRESS
N. Prasad Rao, V. Kalyanaraman / Journal of Constructional Steel Research (2001) 15
3
352
351
353
354
355
356
357
358
359
360
361
362
364
363
365
366
367
368
370
369
371
British code. The British code prescribes the application of a fictitious load acting
transverse to the main member being stabilized by the secondary member, at the
node of attachment of the secondary member to the main member. This force to
be applied is prescribed as a percentage of the main leg or other main bracing
member force, depending upon the slenderness ratio of the member (Table 7) .
This force should be applied in the plane of the bracings in turn at each node
where the secondary members meet the main member. The secondary bracing
forces should also be analysed separately, by applying 2.5% of the force in the
main leg distributed equally at all the interior nodal points along the length of
the leg excluding the first and the last node. The nodal forces should be applied
transverse to the leg member in the plane of the bracing.
ASCE Manual 52. The maximum slenderness ratio of the secondary bracing members is restricted to be below 330. This manual does not require calculation of
forces for which the secondary bracing members have to be designed. However,
it suggests that the magnitude of the load in the redundant members can vary
from 0.5 to 2.5% of the force in the supported member.
IS: 802 (1992). This standard specifies the maximum limit on the slenderness
ratio of the redundants to be equal to 250.
376
Thus, it is seen that some variations in the design requirements of the secondary
bracings exist in codes. The non-linear finite element analysis method, discussed in
the earlier section, can be used to evaluate the forces in the secondary bracings prior
to failure. The forces in the secondary bracings so evaluated could serve as a guideline for the design of secondary bracing members.
377
372
373
374
5
375
6
378
379
380
381
382
383
384
For this purpose a parametric study was carried out to evaluate the forces in the
secondary bracings in a typical bottom panel of a K-braced three-dimensional latticed
frame (Fig. 6). In a typical tower the force resultants in the form of vertical force
V, the shear force H and the over turning moment M vary over the height of the
tower. In the parametric study of the single panel of the tower, the force resultants
at the top of the panel were applied corresponding to different values of V/H and
M/bH ratios in the practical range, where V, H and M are vertical force, shear force
1603
1604
1606
1605
1607
1611
1609
1613
1615
1619
1617
1621
1638
1656
1671
Table 7
Secondary member forces calculation BSI DD 133
Applied force as percentage of leg load, F
Slenderness 0 to
ratio (L/r)
40
Applied
1.02
force
(percentage
of FL)
45
50
55
60
65
70
75
80
85
90
95
100
1.15
1.28
1.42
1.52
1.60
1.65
1.70
1.75
1.80
1.85
1.92
2.0
21
3
4
1
221
ARTICLE IN PRESS
16 N. Prasad Rao, V. Kalyanaraman / Journal of Constructional Steel Research (2001)
3
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
5
408
6
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
and over turning moment resultants acting at the centre at the top of the panel. The
forces in the four corner nodes in the model were evaluated corresponding to these
force resultant values and were applied in the three orthogonal directions at the four
top nodes, so as to obtain the desired ratio of the force resultants as given in Table 8
The sections of the main bracing, secondary bracing and leg members were kept
constant in most cases. Changes in the size of these members were made in a few
of the analysis cases (Sl. nos. 2, 3, 4, 5 and 7), to understand the impact of such
changes. In all cases secondary ties joining the main bracings on two adjacent faces
of the three-dimensional latticed tower were provided. A typical displacement configuration prior to failure is shown in Fig. 7(b). The shear H, corresponding to the
failure of the structure as obtained from the non-linear analysis is given in Table 8,
in addition to the corresponding maximum compressive forces in the main leg, FL,
main bracing, Fb, and the secondary bracings Fsb. Further, the maximum values of
equivalent panel shear, tmax, corresponding to the secondary bracing forces, Fsb, from
the non-linear analysis at limit load, are also presented in Table 8.
The parametric study results in Table 8 indicate the following:
The leg forces, FL, obtained from linear and non-linear analyses are nearly the
same in all the cases, the maximum difference being 4%.
The non-linear analysis results indicate appreciable increase in the maximum axial
force in the bracing. The increase can be as high as 38%. This is usually more
in cases where secondary bracings are very light or type 2 secondary bracings
are used. It is therefore essential to design the bracing members conservatively
for the force obtained from the linear analysis.
As the size of the secondary bracings decreases from the standard value (45355
having l/r250) to a lesser value (25254 having l/r330), the strength of the
panel is appreciably decreased (Sl. no. 1 versus Sl. nos. 2 and 6 versus Sl. no.7
in Table 8). However, increases in the size of the secondary bracing above the
standard value do not seem to improve the strength of the panel appreciably (Sl.
no. 1 versus Sl. no. 3 in Table 8). This indicates the importance of the minimum
stiffness requirement of secondary bracings.
It is seen from the results of Sl. no.4 in Table 8 that the same secondary bracings
(45355) are able to sustain even a larger panel force without initiating failure
when the other (leg and main bracing) member sizes are increased. Similarly,
reduction in the main leg size (Sl. no. 5 in Table 8) causes reduction in the strength
of the panel, due to the strength being governed by the leg buckling.
The design recommendations of various codes are compared with the parametric
study results in Table 9. The following conclusions can be drawn based on this comparison:
The secondary bracing forces calculated based on BS recommendations, Fsb, are
compared with the secondary bracing forces obtained from the non-linear analysis
results, Fsb,NLA, in terms of their ratios in Table 9. It is seen that the correlation
is very poor, with the mean value of the ratio equal to 0.86 and the coefficient
of variation equal to 0.38.
21
21
M/bH
I
1001008
1001008
1001008
15015012
656510
1001008
1001008
1001008
1001008
1001008
1001008
II
1001008
1001008
1001008
1001008
Leg
Section
75756
75756
75756
1001008
75756
75756
75756
75756
75756
75756
75756
75756
75756
75756
75756
75756
75756
75756
75756
Belt
75756
75756
75756
75756
75756
75756
75756
75756
75756
75756
75756
Main brace
45355
45355
45355
45355
45355
25254
60605
45355
45355
45355
25254
45355
45355
45355
45355
Redundant
259.2
229.8
178.4
195.3
258
204.2
259.2
448
194.5
229
153.6
195.2
178.8
133.2
139.3
438
433
410
409
435
345
438
757
328
433
290
409
410
426
427
FL
16.6
15.8
6.7
7.3
20.9
16.6
21.0
36.3
15.8
18.6
12.5
7.3
6.7
12.8
11.3
Fb
Linear analysis
436
433
408
407
433
343
436
754
327
431
278
407
408
428
426
FL
21.4
20.9
6.9
10.7
21.1
19.4
20.6
35.1
15.8
18.5
17.3
9.3
8.8
14.3
10.9
Fb
Non-linear analysis
4.2
4.1
4.6
4.8
9.0
4.7
11.7
17.0
5.3
6.8
3.7
10.0
8.0
3.7
3.7
Fsb
3.55
3.47
4.55
4.55
4.00
3.55
4.66
9.75
3.60
3.40
2.17
4.60
4.55
3.60
3.65
tmax
31687
1689
1688
2078
2079
a
2077
FL=force in the leg; Fb=force in the main bracing; H=total shear in the structure, Fsb=force in the secondary bracing; tmax=horizontal component of shear
in the secondary bracings.
2062
2047
1820
1805
V/H
5
6
1758
1788
1773
1803
1743
1730
Sl. no.
1702
1710
1718
1692
Table 8
Parametric study resultsa
1693
1694
1691
1690
3
4
1
221
ARTICLE IN PRESS
3
4
1
221
ARTICLE IN PRESS
18 N. Prasad Rao, V. Kalyanaraman / Journal of Constructional Steel Research (2001)
5365
566
5
6
567
569
568
571
570
434
433
435
436
437
438
439
440
441
442
443
444
445
446
Fig. 7.
The ratio of the maximum value of secondary bracing forces obtained from nonlinear analyses to the maximum leg forces, (Fsb/FL) is also presented in Table 9.
The ratio is in the range of 0.92.7%, comparable to the ASCE recommended
range of 0.52.5%. The mean value of the ratio is equal to 1.6% and the coefficient
of variation is equal to 0.375.
The ratio of tmax to leg force, FL, as a percentage is also given in Table 9. Usual
design practice has been to use a value of 2.5%. It is seen that the mean value
of tmax/F expressed as a percentage is equal to 1.01% with a coefficient of variation of 0.16. It is seen that designing the secondary bracings for a characteristic
panel shear of 1.3% of the leg force is the most consistent method for designing
secondary bracings in addition to prescribing a limiting slenderness ratio in the
range of 250330.
21
ARTICLE IN PRESS
3
4
1
221
32082
2083
2085
2084
2086
2092
2098
2104
2112
2118
2115
2121
Table 9
Comparison of non-linear analysis results with code provisions
Sl.
no.
V/H
M/bH
2130
2148
2139
2157
2159
2168
2177
2186
2195
2204
2213
2222
2231
2240
2249
2251
2260
2269
2278
2287
2296
Maximum force in
redundants, FSb (kN)
Comparison
FSb,NAL
FSb,Code
FSb,Code/FSb,NAL
tmax/FL100
Fsb,NLA/FL100
57
6
8
5.4
5.7
3.5
5.7
5.7
5.7
5.7
0.63
0.51
0.47
1.02
0.84
0.95
0.57
0.71
1.54
1.54
0.93
1.07
1.29
1.10
0.80
0.78
1.13
1.11
1.00
0.86
2.07
2.68
2.25
1.62
1.58
1.33
2.45
1.96
0.87
0.87
3.7
3.7
3.7
3.7
0.87
0.89
0.81
0.77
0.81
0.80
1.12
1.12
0.96
0.95
1.13
1.18
I
9.0
11.7
17.0
5.3
6.8
3.7
10.0
8.0
3.7
3.7
II
4.23
4.14
4.55
4.80
5
6
447
459
Non-linear FEM models were developed for the analysis of panels of latticed angle
towers by calibration with test results. It is found that the current methods of design
of main leg members based on the forces obtained from a linear analysis are not
consistent with test results. The results obtained using non-linear analyses compare
well with test results. Using such a model, full tower analysis can be done to obtain
more accurate values of member forces including secondary bracing forces prior to
failure and the strength of a tower.
This analysis model was used to perform a parametric study to obtain forces in
the secondary bracing members prior to failure. Based on this study it is recommended that the secondary bracing member designs should meet both strength
requirements (tmax1.30FL/100) and stiffness requirements (l/r250330) to perform their functions adequately.
460
5. Uncited references
448
449
450
451
452
453
454
455
456
457
458
461
[4,5,17,21].
21
3
4
1
221
ARTICLE IN PRESS
20 N. Prasad Rao, V. Kalyanaraman / Journal of Constructional Steel Research (2001)
3
462
Acknowledgements
466
The authors acknowledge the constant support given by Dr. T.V.S.R. Appa Rao,
Director, Dr. R. Narayanan, DGS, Structural Engineering Research Centre, Madras.
The authors also wish to thank Mr. P.R. Natarajan, former Head, Tower Testing &
Research Station, SERC, Madras for the technical support during the work.
467
References
463
464
465
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
5
6
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
[1] Roy S, Fang S-J, Rossow EC. Secondary stresses on transmission tower structures. J Energy
Engng 1984;110(2).
[2] Al-Bermani FGA, Kitipornchai S. Nonlinear analysis of transmission towers. J Engng Struct
1992;14(3):13951.
[3] Yan H, Liu Y, Zhao D. Geometric nonlinear analysis of transmission tower with continuous legs.
In: Advances in steel structuresvolume I, Proceedings of International Conference on Advances
in Steel Structures, Dec 1114; Hong Kong, 1996:33944.
[4] Kitipornchai S, Lee HW. Inelastic buckling of single-angle, tee and double angle struts. J Construct
Steel Res 1986;(6):321.
[5] Chen WF, Astsuta T. Theory of beam-columns, behaviour and design, vol. 2. New York: McGrawHill, Inc., 1977.
[6] Chuenmei G. Elasto plastic buckling of single angle columns. J Struct Div ASCE 1984;98(6):1391
5 [Pro. paper 18888].
[7] Shan L, Peyrot AH. Plate element modeling of steel angle members. J Struct Engng 1988;114(4).
[8] Rajmane Sanatkumar P. An investigation on the behaviour of X and K-bracings of hot rolled single
angles. Thesis submitted in partial fulfillment of the requirements for a PhD, Indian Institute of
Technology, Madras, Chennai, India; 1992.
[9] Stoman SH. A stability criteria for X bracing system. J Struct Div ASCE 1988;114(ST8):142634.
[10] Usami T, Galambos TV. Eccentrically loaded single angle columns. Publication of the International
Association for Bridge and Structural Engineering, vol. 31-II. Zurich (Switzerland). p. 153184.
[11] Elgaaly M, Davids W, Dagher H. Non slender single angle struts. Engineering Journal, American
Institute of Steel Construction, second quarter 1992:4958.
[12] Natarajan PR, Muralidharan K, Mohan SJ, Raghunathan MD. Buckling of eccentrically loaded single
angle struts. Report no. RD-30/8. Madras (India): Structural Engineering Research Centre; 1992.
[13] Natarajan PR, Muralidharan K, Mohan SJ. Buckling of K-bracing. Report no. RD-30/2. Madras
(India): Structural Engineering Research Centre; 1991.
[14] Natarajan PR, Muralidharan K, Mohan SJ, Raman NV. Studies on X-braced panels. Report no. RD1418. Madras (India): Structural Engineering Research Centre; 1992.
[15] Natarajan PR, Muralidharan K. Behaviour of a 6 m extension portion of a 220 kV transmission
line tower with K-bracing. Report no. RD-14/6. Madras (India): Structural Engineering Research
Centre; 1992.
[16] BS Code of Practice for strength assessment of members of lattice towers and masts. London: British
Standards Institute.
[17] ECCS 1985. Recommendation for angles in lattice transmission towers. European Convention for
Constructional Steel Work; Jan 1985.
[18] Prasad Rao N, Kalyanaraman. Non-linear analysis of lattice panels in transmission line towers. In:
Trans Tower 96: International Seminar on Modern Trends in Design of EHV Transmission Towers,
Nagpur (India): Institute of Engineers; 1997:8995.
[19] ASCE manuals and reports on engineering practice no. 52, Guide for design of steel transmission
towers. 2nd ed. New York: American Society of Civil Engineers.
[20] Use of structural steel in over head transmission line towerscode of practice IS:802 (part 1/set
2): 1992 third revision. New Delhi (India): Bureau of Indian Standards.
21
3
4
1
221
ARTICLE IN PRESS
N. Prasad Rao, V. Kalyanaraman / Journal of Constructional Steel Research (2001) 21
3
510
511
512
513
514
2305
2081
1686
1602
1536
1449
1189
1058
756
572
564
555
547
539
531
523
515
[21] White DW, Chen WF, editors. Plastic hinge based methods for advanced analysis and design of
steel frames, Bethlaham (USA): Structural Stability Research Council; 1993.
[22] Natarajan PR, Muralidharan K, Mohan SJ, Raghunathan MD. Buckling of eccentrically loaded equal
angle struts. In: International Conference on Stability of Structures, ICSS 95; Coimbatore (India),
PSG College of Technology; 1995.
5
6
21