INGENIERÍA E INVESTIGACIÓN VOL. 32 No.

3, DECEMBER 2012 (37-41)

A comparative study of two models for the seismic analysis

of buildings
Estudio comparativo de dos modelos para análisis sísmico de edificios
A. Luévanos1

ABSTRACT
This paper presents a model for the seismic analysis of buildings, taking two lumped masses at each level in a structure’s free nodes
and comparing them to the traditional model which considers lumped masses per level, i.e., a mass for each floor of the entire build-
ing. This is usually done in the seismic analysis of buildings; not all values are conservative in the latter, as can be seen in the table of
results. Both models took shear deformations into account. Therefore, the usual practice of considering a lumped mass per each
level would not be a recommended solution; using two lumped masses per level is thus proposed and is also more related to real
conditions.

Keywords: modal analysis, spectral analysis, eigenvalues and eigenvectors, modal participation factor, spectral acceleration, maxi-
mum normal coordinate vector

RESUMEN
En este documento se presenta un modelo para análisis sísmico de edificios en el cual se toman dos masas concentradas por cada
nivel aplicadas en los nodos libres de la estructura y se compara con el modelo tradicional, que considera las masas concentradas
por nivel, es decir, una masa por cada piso de todo el edificio, que es como normalmente se hace; en este último no todos los
valores son conservadores, como se puede notar en la tabla de resultados del problema considerado. Ambos modelos toman en
cuenta las deformaciones por cortante. Por lo tanto, la práctica usual de considerar las masas concentradas, una por cada nivel,
no será una solución recomendable, y se propone el empleo de tomar dos masas concentradas por nivel, que se apega más a la
realidad.

Palabras clave: análisis modal, análisis espectral, valores y vectores característicos, factor de participación modal, aceleración
espectral y vector de coordenadas normales máximas.

Received: June 23th 2011

Accepted: November 9th 2012

specific acceleration registries can be used for the dynamic

Introduction1 2
method (Zárate, Ayala et al., 2003).
Three types of methods can be used for the seismic analysis of
building structures: simplified, static and dynamic methods. The This study was aimed at presenting a model which would take
simplified method is applicable to regular structures standing less into account a building’s two masses per level applied to free
than 13 m high and simultaneously fulfilling all requirements nodes and considering three degrees of freedom at the joints,
indicated by the building regulations. The static method is appli- comparing it to the traditional model taking one mass per level
cable to buildings whose height is less than or equal to 30 m for and considering one degree of freedom per floor (horizontal
regular structures and irregular structures standing less than 20 displacement per level). Both models took shear deformation
m high; these limits increase to 40 m and 30 m, respectively, for into account.
structures sited on rocky terrain. The dynamic method consists
of the same basic steps as that for the static method, with the Analytical development
reservation that applicable lateral forces in the floors’ centre are
determined from a structure’s dynamic response. Modal spectral Equations of motion in a structural dynamic system
analysis and step-by-step analysis or calculating responses having Overall equations of motion in a structural dynamic system,
(Przemieniecki, 1985) without including border conditions, can
be written in matrix form as follows:
1
Arnulfo Luévanos Rojas. Affiliation: Facultad de Ingenieria, Ciencias y Arquitectura,
Universidad Juárez del Estado de Durango, Mexico. Engineering Doctor with
specialty in Planneation and Construction, MSc in Planneation and Construction,
[ ][ ] [ ][ ] [ ][ ] [ ]
Civil engineer, Universidad Juárez del Estado de Durango, Mexico. MSc in Structu-
res, Escuela superior de Ingenieria y Arquitectura del Instituto Politécnico Nacional, where: is a vector of n x 1 generalised absolute displace-
Mexico. MSc in Administration, Facultad de Contaduría y Administración de la
Universidad Autónoma de Coahuila, Mexico. E-mail: [email protected]
ments (unknown) corresponding to non-restricted degrees of
How to cite: A. Luévanos. (2012). A comparative study of two models for the freedom “n”, is a vector of m x 1 generalised absolute dis-
seismic analysis of buildings. Ingeniería e Investigación. Vol. 32, No. 3. December placements (null or known) corresponding to the degrees of
2012, pp. 37-41.

37
A COMPARATIVE STUDY OF TWO MODELS FOR THE SEISMIC ANALYSIS OF BUILDINGS

freedom in supports “m”, are mass matrices, damping The value of the equivalent mechanical elements acting in free
and stiffness are associated with degrees of freedom “n” and/or nodes (McCormac, 2007) is:
“m” respectively, is a vector of n x 1 representing associate
dynamics’ requirements regarding degrees of freedom “n”, is (7)
a vector of m x 1 representing associate reactions (unknown) to
The mechanical element components acting on members
the supports’ degrees of freedom "m".
(Tena Colunga, 2007) are:
Considering linear systems involving orthogonality for stiffness
( ), mass ( ) and damping ( ) matrices, it is convenient to diag- (8)
onalise the system of equations of motion to transform it into a
where and are each bar’s stiffness and displacement.
normal modal coordinate system. A system having free un-
damped vibration, which can exist in the absence of any excita- Application
tion of the supports, would give:
An example of the dynamic seismic design method is presented,
(1) using two different models, considering shear deformation for an
office building built with a steel frame structure. The analysis is
where is a mass matrix corresponding to non-restricted only developed transversally. Figure 1 shows the office building’s
degrees of freedom “n”, is a stiffness matrix corresponding floor-plan and elevation and Figure 2 shows the horizontal re-
to non-restricted degrees of freedom “n”, is a vector of sponse spectrum, representing soil movement where the building
relative displacement and ̈ is a vector of relative acceleration. is supported. Table l shows the steel profile properties.
The solution of equation (1) is defined (by Aguilar Falconi, 1998;
García Reyes, 1998) as:

where is natural vibration frequency, is a modal vector

associated with “ω”, √ and t is time.
The values of and are determined by resolving
eigenproblems as:

( ) (2)

The modal participation factor (Clough, Penzien, 1975;

Bazan, Meli, 1998) can be expressed as:

Poisson ratio = 0.32

(3)
Figure 1. Floor-plan and elevation regarding an office building construct-
ed with steel frame structure.
where is the transposed vector of a modal vector corre-
sponding to the mode "n" is a pseudostatic influence vec-
tor.
The normal maximum coordinates of the system for each mode
( ) are:

( ) (4)
ω
where is spectral acceleration corresponding to mode "n".
The vectors corresponding to maximum relative displacement
vector components for each mode can be defined
as:

( ) (5)
The maximum value of the vector of relative displacements in
structural system is: Figure 2. The horizontal response spectrum

⁄
Table 1. Steel frame properties
{∑( ) } Total area Shear area Moment of inertia
Frames
(cm2) (cm2) (cm4)
(6) W10X60 114.19 27.42 14,318
or: W10X45 85.16 22.83 10,364
W10X21 40.00 15.35 4,454
⁄ W24X94 178.71 80.83 111,966
( ) ( ) ( ) W24X61 116.13 64.06 64,100

38 INGENIERÍA E INVESTIGACIÓN VOL. 32 No. 3, DECEMBER 2012 (37-41)

 LUÉVANOS

The load to be considered in the analysis by level was:

Weight of level 1 = 6,867 N/m2
Weight of level 2 = 5,886 N/m2
Weight of level 3 = 4,905 N/m2
Weight of level 4 = 2,943 N/m2
Elasticity modulus = 20,019,600 N/cm2

Model 1
The beams and columns were considered for analysis in this
model, taking into account two lumped masses per level applied
in free nodes and considering three degrees of freedom at each
joint (horizontal displacement, vertical displacement and rota-
tion). The building was only analysed in a transversal direction.
Figure 3 gives the mathematical model. Figure 4. A lumped mass per building level

MATLAB software was used for solving the determinant and the
roots of the polynomials. Table 2 shows the first four modes of
the sixteen for model 1 (M1) and the four modes for model 2
(M2) are presented in.
Table 3 shows the spectral accelerations of the first four modes
of M1 and the four modes of M2.
Equation (3) gave modal participation factor . The maximum
normal coordinates of the system for each mode ( )
were located by using equation (4). The first four values for the
M1 and the four values for M2 are shown in Table 4. The maxi-
mum relative displacement vector components for each mode
were given by equation (5) and the maximum value
of the structural system relative displacements’ vector for build-
ing was obtained by equation (6). These values are
shown for both models in Table 5.

Table 2. Eigenvalues
Circular frequency Period
Mode (Rad / sec) (Sec)
M1 M2 M1 M2
Figure 3. Two lumped masses per building level
1 5.8479 6.5879 1.0744 0.9537
2 14.8181 15.9421 0.4240 0.3941
Model 2 3 25.7784 27.3889 0.2437 0.2294
This model considered that the beams were rigid compared to 4 32.4698 33.4725 0.1935 0.1877
the columns and, therefore, the beams did not influence dynamic
analysis of the building. It also considered one degree of freedom Table 3. Spectral acceleration
per level, i.e. horizontal displacement (Luévanos Rojas, 2010). Frequency Acceleration
The mathematical model is presented in Figure 4. Mode ωn ( Hz ) san (cm/sec2)
M1 M2 M1 M2
The mass and stiffness matrices for each member were evaluated 1 0.9307 1.0485 0.1772g = 173.7738 0.1920g = 188.2877
(Appendix; Luévanos Rojas et al., 2010) followed by the change
2 2.3584 2.5373 0.2928g = 287.1387 0.3000g = 294.1995
of local system to overall system. The mass and stiffness matrices
3 4.1028 4.3591 0.3000g = 294.1995 0.3000g = 294.1995
in each member’s general system were then coupled and the
system’s general matrix obtained. This general matrix was organ- 4 5.1677 5.3273 0.3000g = 294.1995 0.3000g = 294.1995
ised to separate the degrees of freedom in the structure (M11
and K11) and degrees of freedom in the supports (M22 and Table 4. Participation factors and maximum normal coordinates of
the system for each mode ( )
K22). A similar transformation was applied by exchanging row
and column matrices (permutation matrix). Maximum normal coordi-
Participation factors
nates for the system for each
Mode Ln
mode (Yn)máx
Ignoring the effect of damping in free vibration, as in equation (1),
M1 M2 M1/M2 M1 M2 M1/M2
being a vector (24x1 for model 1 and 4x1 for model 2) of
relative displacements corresponding to the building’s structural 1 +1.3390 +1.3316 1.0056 +6.8038 +5.7770 1.1777
system degrees of freedom, then the eigenvalues and eigenvec- 2 –0.4325 –0.4028 1.0737 –0.5656 –0.4663 1.2130
tors were obtained by solving the determinant resulting from 3 +0.1277 +0.0952 1.3414 +0.0566 +0.0373 1.5174
equation (2). 4 –0.0972 –0.0799 1.2165 –0.0271 –0.0210 1.2905

INGENIERÍA E INVESTIGACIÓN VOL. 32 No. 3, DECEMBER 2012 (37-41) 39

A COMPARATIVE STUDY OF TWO MODELS FOR THE SEISMIC ANALYSIS OF BUILDINGS

Once the deformations were obtained, equation (7) was used to Table 6. Axial forces in the bars for a central frame (N)
find the values of the forces in “X”, the forces in “Y” and mo-
Bar Node Concept Model 1 Model 2 M1/M2
ments; these were applied at the free joints. Such effects were
equivalent to what would have occurred due to a movement in 0 Force “Y” − 199339 − 137801 1.4466
1
the soil where the building was located. The mechanical elements 1 Force “Y” +199339 +137801 1.4466
at the joints on the members of the whole building were deter- 1 Force “Y” − 113011 − 71701 1.5761
2
mined by equation (8) and then obtained for each of the build- 2 Force “Y” +113011 +71701 1.5761
ing’s rigid frames. The axial forces, the shear forces and the
2 Force “Y” − 64059 − 30166 2.1236
moments for a central frame are presented in Tables 6, 7 and 8, 3
3 Force “Y” +64059 +30166 2.1236
respectively.
3 Force “Y” − 35385 − 6975 5.0731
Table 5. Deformation vector 4
4 Force “Y” +35385 +6975 5.0731
Amount 5 Force “Y” +35385 − 6975 5.0731
Node Concept Unit 5
M1 M2 M1/M2 6 Force “Y” − 35385 +6975 5.0731
1 Displacement “X” cm 2.8763 2.6577 1.0823 6 Force “Y” +64059 − 30166 2.1236
6
2 1 Displacement “Y” cm 0.0504 − − 7 Force “Y” − 64059 +30166 2.1236
3 Rotation rad 0.0018 − − 7 Force “Y” +113011 − 71701 1.5761
7
4 Displacement “X” cm 4.3305 3.6668 1.1810 8 Force “Y” − 113011 +71701 1.5761
5 2 Displacement “Y” cm 0.0848 − − 8 Force “Y” +199339 − 137801 1.4466
8
6 Rotation rad 0.0012 − − 9 Force “Y” − 199339 +137801 1.4466
7 Displacement “X” cm 6.0494 5.1725 1.1695 4 Force “X” 0 +18806 −
9
8 3 Displacement “Y” cm 0.1337 − − 5 Force “X” 0 − 18806 −
9 Rotation rad 0.0007 − − 3 Force “X” 0 +27301 −
10
10 Displacement “X” cm 6.8275 5.7959 1.1780 6 Force “X” 0 − 27301 −
11 4 Displacement “Y” cm 0.1555 − − 2 Force “X” 0 +23338 −
11
12 Rotation rad 0.0004 − − 7 Force “X” 0 − 23338 −
13 Displacement “X” cm 6.8275 5.7959 1.1780 1 Force “X” 0 +21464 −
12
14 5 Displacement “Y” cm 0.1555 − − 8 Force “X” 0 − 21464 −
15 Rotation rad 0.0004 − −
16 Displacement “X” cm 6.0494 5.1725 1.1695 Table 7. Shear forces in the bars for a central frame (N)
17 6 Displacement “Y” cm 0.1337 − −
Bar Node Concept Model 1 Model 2 M1/M2
18 Rotation rad 0.0007 − −
0 Force “X” − 114463 − 91527 1.2506
19 Displacement “X” cm 4.3305 3.6668 1.1810 1
1 Force “X” +114463 +91527 1.2506
20 7 Displacement “Y” cm 0.0848 − −
1 Force “X” − 137144 − 69916 1.9616
21 Rotation rad 0.0012 − − 2
2 Force “X” +137144 +69916 1.9616
22 Displacement “X” cm 2.8763 2.6577 1.0823
2 Force “X” − 63500 − 46617 1.3622
23 8 Displacement “Y” cm 0.0504 − − 3
3 Force “X” +63500 +46617 1.3622
24 Rotation rad 0.0018 − −
3 Force “X” − 30038 − 19296 1.5567
4
4 Force “X” +30038 +19296 1.5567
Results 5
5 Force “X” +30038 − 19296 1.5567
The values for the building’s vibration mode frequencies for both 6 Force “X” − 30038 +19296 1.5567
models are shown in Table 2. It was observed that values for M1 6 Force “X” +63500 − 46598 1.3627
were lower regarding M2 in terms of frequency and logically the 6
7 Force “X” − 63500 +46598 1.3627
periods were inverse. The first four modes of M1 are presented
7 Force “X” +137144 − 69935 1.9610
in this Table, even though the work involved sixteen modes 7
resulting from the dynamic analysis. 8 Force “X” − 137144 +69935 1.9610
8 Force “X” +114463 − 91400 1.2523
Table 3 shows spectral acceleration. These values were obtained 8
9 Force “X” − 114463 +91400 1.2523
from the frequency of each of the structure’s vibration modes
4 Force “Y” +7956 − 6975 1.1406
and these results were found by means of the horizontal re- 9
sponse spectrum of the soil where the building was constructed; 5 Force “Y” − 7956 +6975 1.1406
this spectrum is presented in Figure 2. 3 Force “Y” +24054 − 23201 1.0368
10
6 Force “Y” − 24054 +23201 1.0368
The participation factors and maximum normal coordinates of
the system for each mode are given in Table 4, all values in M1 2 Force “Y” +41222 − 41536 0.9924
11
being higher than in M2. The participation factors increased by 7 Force “Y” − 41222 +41536 0.9924
34.14% in M1 and maximum normal coordinates increased by 1 Force “Y” +61842 − 66100 0.9356
51.74% (both percentages are presented in the third mode). 12
8 Force “Y” − 61842 − 66100 0.9356

40 INGENIERÍA E INVESTIGACIÓN VOL. 32 No. 3, DECEMBER 2012 (37-41)

 LUÉVANOS

Table 8. Moments in the bars for a central frame (N-m)

Conclusions
Bar Node Model 1 Model 2 M1/M2 Horizontal displacement, vertical displacement and rotation at
each joint was not restricted given the results obtained in M2
0 +272414 +243877 1.1170
1 taking into account four degrees of freedom, one for each floor
1 +299931 +213750 1.4032 (i.e. horizontal displacement at each level) and M1 considered
1 +244750 +116857 2.0944 twenty-four degrees of freedom. According to the above, it was
2
2 +235263 +127844 1.8402 noted that several degrees of freedom were ignored in M2, being
2 +112815 +79834 1.4172
so reflected in the system’s response.
3
3 +109421 +83326 1.3132 Frequency analysis revealed that M2 involved neglecting certain
3 +53592 +32687 1.6396 modal shapes (symmetric modes and/or anti-symmetrical) system
4
4 +51552 +34865 1.4786 which, in the case of soil excitation, are present and should be
5 +51552 +34845 1.4794
considered, since they often reflect relatively low frequencies.
5
6 +53592 +32677 1.6400 General practice considering a lumped mass for each level is thus
6 +109421 +83297 1.3136 not recommendable, whereas the model proposed in this paper
6 seems to be the most appropriate one for seismic analysis of
7 +112815 +79804 1.4136
buildings’ structural systems.
7 +235263 +127854 1.8401
7
8 +244750 +116925 2.0932
8 +299931 +213466 1.4051
8
9 +272414 +243543 1.1185
References
4 +39799 +34865 1.1415
Aguilar Falconi, R., Acciones para el diseño sísmico de estructu-
9 ras, Limusa-Wiley, 1998, pp. 119-135.
5 +39799 +34845 1.1422
Appendix. Formulario de Teoría de Estructuras. Matrices de Rigi-
3 +120251 +116013 1.0365 dez Elementales, de Masa Congruentes, y de Rigidez Geomé-
10
6 +120251 +115964 1.0370 trica. Disponible en:
http://www.esiold.us.es/php/infgen/aulav/teorestructurasind/M
2 +206138 +207678 0.9926
11 atrices_de_rigidez_elementales.pdf
7 +206138 +207648 0.9927
Bazan, E., Meli, R., Diseño sísmico de edificios, Limusa-Wiley, 1998,
1 +309211 +330607 0.9353 pp. 225-239.
12
8 +309211 +330391 0.9359 García Reyes, L. E., Dinámica estructural aplicada a diseño sísmi-
co, Universidad de los Andes, Facultad de Ingeniería, Departa-
Table 5 gives the structural system’s relative deformations; all mento de Ingeniería Civil, Bogotá, Colombia, 1998, pp. 321-548.
values were greater in M1 (18.10% increase). Only horizontal Clough, R. W., Penzien, J., Dynamics of Structures, McGraw-Hill,
displacements were compared as M2 did not consider the other 1975, pp. 145-284.
two deformations which would be present in any given structure. Luévanos Rojas, A., Seismic analysis of a building of four levels;
making a comparison, despising and considering the defor-
Table 6 shows the axial forces in the structure. There was a mations by Sharp, International Review of Civil Engineering
5.0731 times greater increase in M1 than M2; this only occured (I.RE.C.E.), Vol. 1, No. 4, Sep., 2010, pp. 275-279.
in the columns and axial load was not presented in beams for Luévanos Rojas, A., Betancourt Silva, F., Martinez Garcia, I.,
M1. Luévanos Rojas R. and Luévanos Soto, I., Vibrations in systems of
pipes with different excitation in its ends, International Journal of
Table 7 gives the shear forces; there were differences of up to Innovative Computing, Information and Control, Vol. 6, No. 12,
96.10% in all top columns in M1 compared to M2 and a 14.06% Dec., 2010, pp. 5333-5350.
increase in the upper members of the beams in M1. Such differ- Mc Cormac, J. C., Structural analysis: using classical and matrix
ence decreased when dealing with each lower floor and became methods, John Wiley & Sons, 2007, pp. 550-580.
greater in M2 when arriving at level 1. Przemieniecki, J. S., Theory of Matrix Structural Analysis, McGraw-
Hill, 1985, pp 150-163 y 278-287.
The moments acting on the bars of the structure are shown in
Tena Colunga, A., Análisis de estructuras con métodos matricia-
Table 8. M1 was greater for all columns by up to 2.0944 times les, Limusa-Wiley, 2007, pp. 93-98.
than M2, while shear forces behaved similarly in beams, having
Zárate, G., Ayala, A. G., García, O., Método sísmico estático para
14.22% increase in M1 in the top bar whilst the bottom bar in
edificios asimétricos: revisión de enfoques. Revista de Ingeniería
M2 was greater. Sísmica No. 69, 2003, pp. 25-44.

INGENIERÍA E INVESTIGACIÓN VOL. 32 No. 3, DECEMBER 2012 (37-41) 41