Blucher Mechanical Engineering Proceedings

May 2014, vol. 1 , num. 1

www.proceedings.blucher.com.br/evento/10wccm

DYNAMIC ANALYSIS OF BUILDINGS USING THE FINITE ELEMENT METHOD

C. A. T. Lucena1, P. C. O. Queiroz2, A. L. H. C. El Debs2, A. V. Mendona1
1

Dept. of Mechanical Engineering. University Federal of Paraiba, Brazil

[email protected]
[email protected]
2
Dept. of Structural Engineering. School of Engineering at the University of So Paulo, Brazil
[email protected]
[email protected]

Abstract. This paper deals with finite element vibration analysis of buildings. Each component of
building is discretized by its appropriated finite element, that is, bar and beam element for the frame
sub-structure, plate finite element for the slabs. By applying compatibility and equilibrium conditions,
all sub-structural interactions are incorporated into the system in order to produce a more refined
structural analysis of buildings. Other issues for building vibration such as shear deformation, rotatory
inertia, and plate-beam eccentricity are investigated as well. Numerical examples are presented and
compared with results from commercial numerical packages widespread.
Keywords: FEM, Buildings, Dynamic
1. INTRODUCTION
The Finite Element Method (FEM) is a numerical technique based on continuum discretization so
that the body is divided into finite number of small parts named elements and by expressing the
unknown variable fields in terms of assumed approximating functions within each element. These
functions are expressed in terms of discrete points named nodes. Clough & Wilson [1] give interesting
historical details about first steps to establish of FEM solutions. For example, in 1960 the designation
of finite element method was coined by Ray Willian Clough when a static analysis of stress plane
problem [2] was modeled using that new born technique. Since then FEM solutions have been
received many other contribution and applied to a variety engineering problems. This is the case for
vibration analysis of frame and plate [3]-[6]. Additional details on vibration FEM solutions including
more complex structures such as shells can be found elsewhere, for example [3]-[9].
A specific topic in building structural analysis has been studied by many researchers is associated
with eccentric relative position beam-plate at each floor. One of many strategies has established to
deal with eccentricity problem is the rigid offset approach. Harik et al. [14] presented an analytical
solution, Mukhopadhya [15] proposed a finite difference solution, Harik et al. [14], Arajo [16],
Sapountzakis et al. [17], Deb et al. [18]-[19], gave finite element solutions, Tanaka et al. built
boundary element solutions.
In this paper the influence of effects such as shear deformation, rotatory inertia, and platebeam eccentricity into building vibration responses are analyzed by house-made program called EDF
and commercial packaged Ansys.

2. ASSEMBLY OF MASS AND STIFFNESS MATRICES

In this section the strategy of assembling of each elemental mass and stiffness matrices into global
matrices of the building is briefly discussed. The building analysis can be split into major problems,
namely space frame and plates (in bending and/or in tension), and their interaction effects, see Figure
1.

Plates

Space Frame

Building

Figure 1. Building structural components.

In a space frame problem, the efforts in each member can be simultaneously mobilized due to axial,
bending, and torsional actions. Both mass and stiffness matrices of a 3D frame member are wellknown and they can be found in many works, such as Queiroz [9] and Lucena [10]. If only an isolated
space frame is considered, the assembly of the global matrices can be done as shown in appendix Figure 14(a). In this paper for the stretching and bending plate vibration problems were implemented
respectively two elements namely CST (Constant Stress Triangle) and DST (Discrete Shear Triangle).
Mathematical details for both elements can be found elsewhere, for example Petyt [12], Batoz and
Lardeur [13], Lucena [10]. A relevant topic is the study of influence of beam-plate eccentricity of each
floor to the global response of the building, see Figure 2.

Figure 2. (a) Beam-plate configuration: (a) non-eccentric case; (b) eccentric case.

When a beam and a plate have no eccentricity implies nodal points of beam element coincide with
some nodes of plate element, see Figure 2(a). In this case, no additional step is required and both beam
and plate elemental contributions can be directly and independently assembled in global matrices of
the structure. On other hand when beam-plate eccentricity exists (very usual situation in buildings, see
Figure 2(b) it is necessary to reposition the degrees of freedom of the beam (usually located on its
centroidal axis) onto a plane of the plate (usually median plane, see Figure 3) or vice-versa. Due to
change of nodal location of beam, the energy conservation of system requires mass and stiffness
matrices of the beam are transformed too.

Figure 3. Local displacement vectors of the beam and their eccentricity.

Let there be {u} and {U} the global displacement vectors at beam centroidal axe and at plate
median plane, respectively, and a relation can be written as follows.

1
0

0
[ ] =
0
0

u1
U 1
u
U
2
2
u 3
U 3
= [ ] where
u 4
U 4
u 5
U 5


u 6
U 6

0 0 0 0 e
1 0 0 e 0
0 1 0 0 0

0 0 1 0 0
0 0 0 1 0

0 0 0 0 1

(1)

The final transformed matrices of the beam with respect to plate median plane system are:

[K ] = [ ] [K ][ ]
[M ] = [ ] [M ][ ]
P

(2)

(3)

In addition, the assembly of the transformed matrices given in Eq(2) and Eq(3) can be done as
shown in appendix - Figure 14(b). For the building problem, an assembling strategy is to create three
subsets of typical nodes. The first is associated with (stretching/bending) plate nodes that do not
receive any contribution from space frame nodes, called Plate nodes. The second subset is called
Mixed nodes and it receives both plate and space frame contributions. The third subset called
Frame nodes is formed by off-plate nodes. The subset dimensions are respectively given by
5 N L , 6 L M and 6 L F . Let there be a plate element with node orientation (i, j, k) and interacting with

two (i,j) and (j,l) oriented frame elements, see Figure 4. Then, the assembly of building matrices from
all elemental contribution can be done as shown in appendix - Figure 14(c).

Detail A

Figure 4. (a) Real structure; b) Discretized structure; c) Beam-Plate elements interacting at node j.
3. NUMERICAL RESULTS
All mathematical details of the finite element solution described in this paper were encoded in the
program called EDF.FOR, see Lucena [11]. In order to check EDF.FOR performance, vibration
analysis of two structures (space frame and four-storey building) is presented. The results of this
house-made program are compared to commercial package ANSYS outputs.
3.1 Space Frame
This example was originally proposed by Petyt [12], see Figura 5. The mechanical and dimensions
of the members are: Young modulus, E = 219,9 x10 9 N / m 2 ; density, = 7900kg / m 3 and length,
L = 1m . Both Euler-Bernoulli and Timoshenko models are taken into account and the results from
both programs are shown in Table 1.

Section AA

Section BB

Figura 5. Space frame and its discretization.

Mode
Number

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20

Table 1. Frequency results (Hz)

Ansys
EDF
Model
Model
Model Euler2 Model Euler
Timoshenko 1 Timoshenko 2
11.76402563 11.76387476
11.808643
11.808
11.76402563 11.76387476
11.808643
11.808
15.3966169
15.36295663
15.411955
15.412
33.98476808 33.98170942
34.112377
34.111
33.98476808 33.98170942
34.112377
34.111

EDF
Model
Euler1
11.808796
11.808796
15.4457303
34.1154902
34.1154902
43.433024
123.299665
125.736015
145.00127
154.139833
154.139833
157.719628
169.296732
169.296732
169.626978
190.617268
225.625976
225.625976
239.219681
246.457127

43.330254
123.18396
125.6063
144.88404
153.98871
153.98871
157.54384
169.14853
69.148533
169.32003
190.16343
225.32584
225.32584
238.87411
246.0154

43.329
122.05
124.36
143.99
153.63
153.63
156.15
168.87
168.87
169.31
190.15
223.8
223.8
238.54
245.72

43.29169296
122.9237718

43.18938625
122.8097391

125.3376278
144.3209478

25.20986295
144.2060394

153.4583667
153.4583667

153.3105848
153.3105848

156.8369922
164.9313557
168.3485203
168.3485203
185.4286436
223.4873333
223.4873333
236.8261719
243.4084611

156.6655141
164.6636774
168.2042581
168.2042581
185.0307075
223.2009211
223.2009211
236.4969022
243.0062042

For sake of a better visualization, the results of Table 1 are plotted, see Figure 6.

Figure 6. Space frame results.

1
2

No rotatory inertia
Rotatory inertia

Ansys
Model
Timoshenko

11.774
11.808
15.391
33.992
34.108
43.263
121.99
124.3
143.88
153.45
153.63
156.02
164.81
168.69
168.87
185.12
221.94
223.76
237.25
244.01

3.2 Four-storey building

In this example vibration analysis for variety slender ratios of structural components of a fourstorey building is done, see Figure 7. In addition the influence of plate-beam eccentricity, shear
deformation and rotatory inertia effects in building vibration behavior is also considered. For all
components, the following values for mechanical properties are set: Young modulus,
E = 21x109 N / m 2 ; density, = 2500kg / m 3 . For plate elements, Poissons ratio is = 0,30 and
all results are done using DKT element for bending plate contributions into the problem, except when
is explicitly warned.

Figure 7. Four-storey building: (a) geometry, (b) discretization

The study was divided into three parts: firstly beam slender ratio influence in building vibration is
analyzed under following assumptions. All plates have 10cm constant thickness and all columns have
a typical (20 x20)cm square cross-section. A (20 xh)cm cross-section is set for all beams and two
values are assumed for h.
Table 2 Frequencies (Hz) for non-eccentric case: (plate h=0.10m) (beam h=0.50m)
Euler-Bernoullis beam
Timoshenkos beam
Mode
Number
EDF
ANSYS
EDF
ANSYS
1
1.033634243568042
1.0335
1.026163178509705
1.0261
2
1.051321466365204
1.0512
1.042713789161397
1.0426
5
10
15
20
25
30
35
40
45

3.083254020001557
6.133236349035065
10.699237061042660
17.936212724087930
24.823920795071000
30.606705038030040
40.816869158106850
47.758536892544060
51.998620754883710

3.0829
6.1322
10.651
17.847
24.497
30.556
40.234
46.995
50.961

3.059042393325857
6.090743238372382
10.629603122588180
17.743458077855130
24.609889249841980
30.441299394962270
40.348166803980430
46.961864549614620
51.399784463171740

3.0586
6.0897
10.582
17.655
24.338
30.389
39.775
46.235
50.532

Table 3 Frequencies (Hz) for eccentric case: (plate h=0.10m) (beams h=0.50m)
Euler-Bernoullis beam
Timoshenkos beam
Mode N
EDF
ANSYS
EDF
ANSYS
1
1.061105210654078
1.0600
1.052878883863958
1.0518
2
1.069293282336665
1.0684
1.060060841738016
1.0593
5
3.124465230676489
3.1205
3.098809652840658
3.0951
10
6.143373253738758
6.1371
6.100320485187121
6.0943
15
12.818649259830260
12.742
12.705003588424920
12.630
20
20.492074748603760
20.333
20.289840881110260
20.189
25
27.490250113643460
27.043
27.221579028101290
26.782
30
31.587787232354530
31.561
31.405497422876140
31.376
35
43.533910399307850
42.826
43.291480405625190
42.625
40
50.194220128101630
48.459
50.045250822373090
48.315
45
53.158655994945020
51.930
52.177054714839100
51.022
Table 4 - Non-eccentric case: (plate h=0.10m) (beam h=0.80m)
Euler-Bernoullis beam
Timoshenkos beam
Mode N
EDF
ANSYS
EDF
ANSYS
1
0.9777737465142289
0.97767
0.9698353132263625
0.96973
2
0.9796925061525277
0.97959
0.9722366292403232
0.97213
5
2.845731851293956
2.8454
2.824630045531332
2.8243
10
5.497798210133395
5.4970
5.459845809899770
5.4591
15
14.831850434225120
14.708
14.716596862275180
14.596
20
24.099711540057980
23.971
23.685053731839450
23.524
25
30.405686660826370
29.792
30.182141949107190
29.584
30
35.744411088574980
34.730
35.716495690795860
34.703
35
45.569945054429910
44.477
44.918095831465320
43.615
40
53.179774314214070
52.360
52.789080492217070
51.107
45
59.693230687868560
58.063
58.688231895087200
58.148
Table 5 - Frequencies (Hz) for eccentric case: (plate h=0.10m) (beam h=0.80m)
Euler-Bernoullis beam
Timoshenkos beam
Mode N
EDF
ANSYS
EDF
ANSYS
1
0.9813890571155897
0.98293
0.9732779425289443
0.98125
2
0.9863427704428834
0.98709
0.9786756420889796
0.98592
5
2.861052757951518
2.8606
2.839452174069153
2.8580
10
5.500626179040095
5.4950
5.462595784379647
5.4940
15
16.706018819362260
16.555
16.558035888165120
16.411
20
28.044182252379500
27.728
27.343770282073450
27.046
25
31.864847655130900
31.222
31.620607422652910
30.992
30
36.210165576136050
35.291
35.955307263621120
34.968
35
47.867119090662290
44.218
47.636652351543720
44.003
40
55.012062684831130
52.910
54.902568106283380
51.712
45
58.919546022347830
55.373
58.042801175215840
54.735

Mode
1
2
5
10
15
20
25
30
35
40
45

Variation
(%)
-5.4
-6.8
-7.7
-10.4
38.6
34.4
22.5
16.8
11.6
11.4
14.8

Figure 8. Beam ridigity influence on non-eccentric case

For sake a better visualization, the results of non-eccentric plate-beam case shown in Table 2 and
Table 4 are plotted together in Figure 8. The eccentric plate-beam results in Table 3 and Table 5 are
plotted in Figure 9.

Mode
1
2
5
10
15
20
25
30
35
40
45

Variation
(%)
-7.5
-7.8
-8.4
-10.5
30.3
36.9
15.9
14.6
10.0
9.6
10.8

Figure 9. Beam ridigity influence on eccentric case

If the results of the thinner beams (h=50) in Figure 8 and in Figure 9 are taken as reference values,
the first ten natural frequencies for thicker beams (h=80) decreased respectively between 5.4%and
10.4% for non-eccentric problem and 7.5% and 10.5% for eccentric case. On the other hand when
range between 15th and 45th modes is checked, it can be seen that natural frequencies of the thicker
beam problem are increased between 11.4% and 34.4% for the non-eccentric case and 9.6% to 36,9%
for the eccentric problem. These major discrepancies are certainly due to more severe bending mode
activation for thicker beams. A typical case is observed about 20th mode where mode configurations
for both cases (h=50 and h=80) are shown in Figure 10.

Figure 10. Settings of the 20th mode (a) h=80; (b) h=50
The second part of the building vibration analysis is concerned on plate slender ratio influence in
building responses. Hence some assumption is done: beam-plate problem is always eccentric, all
beams have the same cross-section (20 x80)cm , and all columns cross-section is square with dimension
20cm.
The influence of plate thickness variation in the building vibration behavior is initially done and
shear deformation effects are not included (classical plate model). Analyses for three thicknesses of
the plate 10, 20 and 30 cm are independently done. By selecting DKT element, EDF and Ansys
results are shown in Table 5, Table 6 and Table 7. For sake a better visualization the outputs are
plotted together in Figure 11.
Table 6. Frequencies (Hz) eccentric case: (plate h=0.20m) (beam h=0.80m) - column (0.20x0.20)m
Euler-Bernoulli
Timoshenko
Mode N
EDF
ANSYS
EDF
ANSYS
1
0.8293750409755191
0.83044
0.8227936436689241
0.82388
2
0.8340287604609957
0.83445
0.8276543806135503
0.82810
5
2.417162832657765
2.4167
2.399140468688497
2.3988
10
4.634355366607846
4.6318
4.602253750231750
4.5998
15
18.129607984803820
18.046
17.833315482366810
17.714
20
28.109150010122560
28.145
27.991753951863470
28.027
25
42.387049407920120
41.670
41.322183753097150
36.889
30
45.099012533007870
44.012
43.801345471503300
42.697
35
58.162430748004090
57.335
57.756874147415520
55.909
40
64.49996888117163
64.405
63.066791827934590
62.789
45
65.345883383739720
65.287
63.707594933563410
63.631
Table 7. Frequencies (Hz) eccentric case: (plate h=0.30m) (beam h=0.80m) - column (0.20x0.20)m
Euler-Bernoulli
Timoshenko
Mode N
EDF
ANSYS
EDF
ANSYS
1
0.7314230031823287
0.73224
0.7259813796638660
0.72678
2
0.7357778526412748
0.73612
0.7303252140427594
0.73065
5
2.131274143423030
2.1311
2.115764573823893
2.1156
10
4.079733654343166
4.0784
4.051566408859258
4.0503

15
20
25
30
35
40
45

16.630121151899140
29.002845646654430
44.219052338206040
55.184507852580140
63.994146800298080
65.285823976308460
65.604209407136820

16.612
29.057
43.821
54.130
63.636
65.233
65.584

16.579815076595760
28.904550876622010
43.227278646436850
54.682751123006350
62.540515496259820
63.617427580988740
63.909799670753620

Mode
1
2
5
10
15
20
25
30
35
40
45

16.562
28.960
42.832
53.618
62.281
63.571
63.882

Variation (%)
HL=20 HL=30
-15.5
-25.5
-15.4
-25.4
-15.5
-25.5
-15.7
-25.8
8.5
-0.5
0.2
3.4
33.0
38.8
24.5
52.4
21.5
33.7
17.2
18.7
10.9
11.3

Figure 11. Plate ridigity influence for eccentric case.

By assuming thinner plate results as reference values, it can be seen there is a decreasing of natural
frequencies until 10th mode (15.7% for h=20 and 25.8% h= 30 cm). For higher modes (until 40th) the
thicker plates increase natural frequencies. A complementary study of the earlier problem in order to
check shear deformation effect in all building components is done as well. The vibration analysis is
modeled by EDF (using both DST plate and Timoshenko beam elements) only and the results are
shown in Table 8 and Figure 12.
Table 8. Frequencies (Hz) incorporating shear deformation effects for plate and beam
DST elements for plate and Timoshenkos beam
(beam h=0.80m), column (0.20x0.20)m
Mode
plate (h=10 cm)
Plate (h=20 cm)
Plate(h=30 cm)
Number
e=35cm
e=30cm
e=25cm
1
0.9732744677093985
0.8227266718092883
0.7257943847425563
2
0.9786749015925560
0.8276392615450652
0.7302702382631422
5
2.839449996129212
2.399105589554805
2.115640255869238
10
5.462592105083606
4.602237793494199
4.051521762546975
15
16.526669415551750
17.803260799809390
16.557976877659960
20
27.326817262917330
27.979698423999490
28.856040883028330
25
31.503036906434580
41.135618019665660
42.754293742999600
30
35.822998742607440
43.680109794382580
54.197368729408740
35
47.543719135812010
57.320421256622860
62.500908874808780
40
54.688519389612870
62.974795699346510
63.596349288916920
45
57.948194188817130
63.703264647440940
63.906124089558420

Mode
1
2
5
10
15
20
25
30
35
40
45

Variation (%)
HL=20 HL=30
-15.5
-25.4
-15.4
-25.4
-15.5
-25.5
-15.7
-25.8
7.7
-0.2
2.4
5.6
30.6
35.7
21.9
51.3
20.6
31.5
15.2
16.3
9.9
13.2

Figure 12. Plate rigidity influence and shear deformation effects on eccentric case
The results show there are no significative changes until 45th mode vibration responses when shear
deformation was taken into account.
At last third part of the building vibration analysis in this paper is concerned about influence of
variation of column cross-section dimensions. The assumptions for this analysis are: all plates has
thickness 10cm and a cross-section (20 x80)cm is set for all beams. Two cross-section dimensions are
set to the all columns, ( 20 x 20)cm and (40 x 40)cm .. The EDF and Ansys results for (40 x 40)cm crosssection columns in Table 9 are shown only. In Figure 13 both column cross-sections analyses are
plotted.
Tabela 9. Frequencies (Hz) eccentric case: (plate h=0.10m) (beam h=0.80m) - column (0.40x0.40)m
Euler-Bernoullis beam
Timoshenkos beam
Mode
Number
EDF
ANSYS
EDF
ANSYS
1
2
5
10
15
20
25
30
35
40
45

2.984353648437686
3.054102673539377
9.212390099187937
17.312097285493380
20.308793665539740
32.328691239677820
35.467487874409320
45.352876833737640
50.625448328801750
59.774830750619830
62.338234193366160

2.9881
3.0591
9.2083
17.097
20.198
31.618
34.664
43.417
48.568
53.851
60.002

2.897020117432950
2.942812889441475
8.899009872081662
17.159695509391720
19.716451606859310
31.930651694976390
34.106090408379160
44.819779928222640
49.406304454311320
59.611505646993460
61.645643383805460

2.9021
2.9500
8.9050
16.950
19.628
31.330
33.445
42.518
47.507
53.757
59.406

Mode
1
2
5
10
15
20
25
30
35
40
45

Variation
(%)
204.1
209.6
222.0
214.7
21.6
15.3
11.3
25.2
5.8
8.7
5.8

Figure 13. Column rigidity influence on eccentric case

The results show lower modes are strongly affected when the column cross-section dimensions are
changed.
4. CONCLUSIONS
The numerical results in the paper suggest results from EDF house-made program are in a good
accordance with those run with commercial package Ansys. The changes in four-storey building
vibration responses are analyzed, when slender ratios of each group of structural components are
independently changed. The most severe changes for lower natural frequencies of the building is due
to column cross-section variation. Shear deformation effects produce slight changes to lower natural
frequencies of the building when compared to classical models.
Acknowledgments
The authors thank Capes and CNPq for financial support for this research.
5. REFERENCES

[1]

Clough, R. W.; Wilson E. L. Early Finite Element Research at Berkeley, Present at the Fifth
U.S. National Conference on Computational Mechanics, 1990.

[2]

Clough, R. W., The Finite Element Method in Plane Stress Analysis, Proc. 2nd ASCE Conf.
On Eletronic Computation, Pittsburg, Pa. Sept. 1960.

[3]

Anderson, R.G., Irons, B.M., Zienkiewicz, O.C. Vibration and stability of plates using finite
elements. Int. Journal of Solids and Structures, 4 1031-1005, 1968.

[4]

Dickinson, S.M., Henshel, R.D. Clough-Tocher triangular plate-bending element in vibration,

AIAA Journal 7 560-561, 1969.

[5]

TOCHER, J. L. Analysis of Plate Bending Using Triangular Elements. University of

California, Berkeley, Ph.D. Dissertation, 1962.

[6]

Mackerle, J. Finite element vibration analysis of beams, plates and shells. Shock and
Vibration, 6 97109, 1969.

[7]

Samanta, A. & Mukhopadhyay, M. Finite element static and dynamic analyses of folded
plates. Eng. Struct., 21 277-871 1999.

[8]

Lim, G.H. Vibration of plates and shells using finite elements (1996-1997). Finite Elem. Anal.
Design, 31 223-30, 1999.

[9]

Queiroz, P. C. O. Anlise esttica e dinmica de estruturas reticuladas: ambiente de simulao

em JAVA. Universidade Federal da Paraba, Dissertao de mestrado 2010.

[10] Lucena, C. A. T. Anlise esttica e dinmica de edifcios industriais utilizando-se o mtodo

dos elementos finitos. Universidade Federal da Paraba, Dissertao de Mestrado, 2011.

[11] Viana, H. R. G. Anlise esttica e vibratria de placas utilizando-se o mtodo dos elementos
finitos. Universidade Federal da Paraba, Dissertao de mestrado, 2008.

[12] Petyt, M. Introduction to finite element vibration analysis. Cambridge University Press, 1990.
[13] Batoz, J.L. & Lardeur, P.A. A discrete shear triangular nine dof element for the analysis of tick
to very thin plates. Int. J. Num. Meth. Eng. V.28. p. 533-560, 1989.

[14] Harik, I. E. & Guo, M. Finite Element Analysis of Eccentrically Stiffened Plates in Free
Vibration. Computer and Structures. 6, Vol. 49, pp. 1007-1015, 1993.

[15] Mukhopadhya, M. Stiffened Plates in Bending. Computer and Structures. 4, Vol. 50, pp. 541548, 1994.

[16] Arajo, J. M. Avaliao dos mtodos simplificados para clculo de lajes macias apoiadas em
vigas flexveis. Teoria e Prtica na Engenharia Civil. Vol. 12, PP. 1-11, 2008.

[17] Sapountzakis, E. J. & Katsikadelis, J. T. Analysis of Plates Reinforced with Beams.

Computacional Mechanics. Vol. 26, pp. 66-74, 2000.

[18] Deb, A. & Botton, M. Finite element models for Stiffened plates under transverse loading.
Computer and Structures. 3, Vol. 28, pp. 362-372. 1988.

[19] Deb, A. & Deb, M. K. Analysis of Orthotropically Modeled Stiffened Plates. International
Journal of Solids Structures. 5, Vol. 25, pp. 647-667, 1991.

[20] Tanaka, M. & Bercin, A. N. Static bending analysis of stiffened plates using the boundary
element method. Engineering Analysis with Boundary Elements. Pp. 147-154, 1998.

APPENDIX

Figure 14. Formation of the matrices of the structure: (a)stiffness

matrix of ith member in the global matrix framed structure,
(b)condensed form of the plate structure matrices, (c)expression for
assembly of stiffened plate structure.

The sub matrices in Figure 14, are given by:

[ [ ] [0 ] ]
[k ] = [ [k ] [ 0 ] ]
[k ] = [k ] [0] + [k ]
[0]
0
[k ]

*
ij 5 x 6

= k ijL

*
ik 5 x 6

**
ij

L
ik 5x5

L
jj 5 x 5

(5)

5x1 5 x 6

P
jj 6x6

5 x1

6 x6

1x 5

(4)

5x1

5x6

(6)

[k ]

k Ljk 5 x 5
=
[0]1x 5

[ ]

[0]5 x 5 + [k P ]
[0] 6 x 6 jk 6x6

(7)

[k ]

kL
= kk 5 x 5
[0]1x 5

[ ]

[0]5 x1 + [k P ]
[0] 5 x 6 kk 6x6

(8)

78 678
[0]5 x1 + [6
k Pjj ]6x6 + [k jj ]6x6

[0] 6 x 6

(9)

**
jk 6 x 6

**
kk 6 x 6

[k ]

***
jj 6 x 6

[ ]

k Ljj 5 x 5
=
[0]1x 5

Pilar

barra