CE-5190: Application of Structural Dynamics to

Earthquake Engineering

MODULE 8

Modal Analysis & Free vibration Response of

undamped MDOF systems
1
 Free vibration in MDOF systems

Vibration of elastic MDOF systems is governed by equation

  ku  0
With p(t) =0 for free vibration and c= 0 for systems without damping: m u

The equation represents N homogenous differential equations* that are coupled through mass
matrix, the stiffness matrix or both matrices; N is the number of DOFs.

The solution u(t) can be determined using initial conditions

u  u(0) u  u (0)

* A homogeneous differential equation is an equation containing a differentiation and a function, with a set of variables

2
 Free vibration in MDOF systems

Contrary to SDOF system, the motion of each mass (or floor) is not a Simple Harmonic
Motion (SHM) during free vibration and thus the frequency of motion cannot be defined
video1.

Furthermore, the deflected shape (i.e., u1 /u2) varies with time , as is evident from
differing deflected shapes at time corresponding to a, b and c of Figure c (next slide),
which are different from each other.
An undamped structure would undergo SHM if free vibration is initiated by appropriate
disturbances of displacements video1 in the various DOFs. These sources of disturbance may
be earthquake, wind, blast, etc

3
 Free vibration in MDOF systems

𝑢 (0)

𝑢 (0)

c
Free vibration of an undamped system due to arbitrary initial displacement u1(0) and u2(0): (a) 2 story
frame; (b) deflected shape at time instants a, b and c; (c) displacement response history of floors.

4
 Free vibration in MDOF systems

As shown in figures given on slides 6 and 7, two characteristic deflected shapes exist
for 2 DOF system shown on slide 4. If the structure is displaced in one of these shapes
and released, it will vibrate in SHM, maintaining the initial deflected shape.
Each characteristic deflected shape is called a Natural mode or Mode shape. video2

5
 Free vibration in MDOF systems: Mode shapes
Mode 1
u2(0)=1

u1(0)= 0.5

a b c
Free vibration of an undamped system in its first natural mode of vibration: (a) two-story frame; (b) deflected
shapes at time instants a, b, c, d, and e; (c) displacement history of floors.

Please note that now u1/u2 remains constant for deflected shapes at various instants a, b, c, d, and e (Fig. c)

6
 Free vibration in MDOF systems: Mode shapes
Mode 2 u2(0)= 1

u1(0)= -1

a b
c

Free vibration of an undamped system in its second natural mode of vibration:

(a) two-story frame; (b) deflected shapes at the time instants a, b, c, d, and e; (c) displacement history.

Again note again that now u1/u2 remains constant for deflected shapes at various instants a, b, c, d, and e
7
 Undamped MDOF system: Natural time periods
A Natural period of vibration Tn of an MDOF system is the time required to complete
one cycle of the simple harmonic motion in one of its natural modes (n-th natural mode).
The corresponding natural circular frequency of vibration is ωn and the natural cyclic
frequency of vibration is fn, where & = 1/

It is to be noted that unlike SDOF (which has single vibration mode) where subscript n in Tn is
used to describe natural time period, subscript n in MODF system is used for any mode number n.
Figures on previous 2 slides show the two natural periods Tn and natural frequencies ωn (n = 1, 2)

of the two-story building vibrating in its natural modes . Where First subscript
indicate DOF and subscript indicate mode number . Thus =Mode shape corresponding to
DOF 2 while system is vibrating in n-th mode

8
 Undamped MDOF system: Natural Frequencies
and associated mode shape
The smaller of the two natural vibration frequencies is denoted by ω1, and the larger by ω2.
Correspondingly, the longer of the two natural vibration periods is denoted by T1 and the
shorter one by T2.
ω1 and T1 are known as Fundamental natural frequency and Fundamental natural time

period, respectively. Mode shape vector is called fundamental mode shape.

Thus, a MDOF system with N-DOF have N natural frequencies such that:
………..

9
Modal Analysis
 Modal Analysis

Modal analysis is a computational or experimental process for predicting or

measuring modal parameters. Resonant frequency, mode shape, and damping are all
of the modal parameters.

Since effect of damping ratio (which occasionally exceeds 5% in civil engineering structures) on the
natural frequencies is negligible, Civil engineers usually compute the frequencies and mode shapes of
an undamped structure.

11
 Evaluation of Natural frequencies & associated mode shapes
EOM of an undamped MDOF system subjected to free vibration is rewritten as

If it is assumed that the resulting motion of MDOF system is simple harmonic, the displacement,
just like in case of SDOF, can be expressed in matrix form as ϕ
where ϕ is a mode shape vector of free vibration and = Phase angle.
ϕ

ϕ ϕ

12
 Evaluation of Natural frequencies & associated mode shapes
ϕ ϕ

ϕ ϕ ϕ ϕ
Equation 2 is a generalized eigenvalue problem that can be written in the form of a standard
eigenvalue problem. For n-th mode, equation 2 can be written as ϕ ϕ

Pre-multiplying both terms of Eqn 2 by m-1, we obtain the following equation:

𝟏 ϕ ϕ 𝟏 ϕ ϕ
dynamic stiffness matrix. 𝟏

The dynamic stiffness matrix E contains the system characteristics, represented by mass matrix m
and stiffness matrix k.
COMMENT.– A physical interpretation of equation 2 is that a mode shape is a displacement configuration of the
structure for which the elastic forces 𝐤ϕ are in exact equilibrium with the inertia forces 𝜔 𝐦ϕ. This property is
the basis of some numerical methods used to evaluate frequencies and mode shapes. 13
 Evaluation of Natural frequencies & associated mode shapes
Eqn II can also be written as ϕ ϕ where λ ≡ ω2 is a real positive
scalar.
The trivial solution ϕi = 0 of the equation does not correspond to any motion of the structure and is
of no interest.
A non-trivial solution is only possible if the determinant of the coefficient matrix is
equal to zero. In other words, vibration is only possible if:

… 4

From determinant, we obtain a polynomial equation of degree N in λn = ωn2 for a system with
N DOFs. The N roots of this equation (λ1, λ2, . . . , λn, . . . , λN) are the angular frequencies, or
natural frequencies, of the system and are associated with a vibration modal vector or mode
shape. In linear algebra, the pairs (λn, ϕn) are the eigenvalues and eigenvectors of the
generalized eigenvalue problem

14
 Modal Analysis of a 2 Story Shear Frame

Problem 8.1
Calculate the vibration frequencies and associated mode shapes of the
structure illustrated in Figure. m = 20,000 kg and lateral story stiffness
k = 18×106 N/m.

Solution

15
Problem 8.1:Natural frequencies

Note: In a uniform moment resisting shear frame undergoing a shear type sway deformation the
natural frequencies of free vibration increase roughly in the ratios of 1:3:5:7:9 etc 16
 Problem 8.1:Mode shapes
The natural frequencies are expressed in the increasing order. The one having the least
value correspond to the first mode and known as fundamental natural frequency. Rest are
the frequencies corresponding to second and third mode shape, respectively.
Since time period is related to the reciprocal of frequencies, The natural time period are
expressed in the decfreasing order. The one having the hihest value correspond to the first
mode and known as fundamental natural time period.

To complete the solution of the eigenvalue problem, we calculate the N vibration mode shapes
corresponding to the N eigenvalues. If we substitute the values of ω for n-th mode in equation 2, we
get ϕ

where ϕ is the mode shape corresponding to ωn 17

Problem 8.1:Mode shapes (contd).
Fundamental mode shape ϕ
(2 − 1.111 ∗ 10 ∗ 343.83) −1 𝜙
18 ∗ 10 =0
−1 (1 − 1.111 ∗ 10 ∗ 343.83) 𝜙
1.618 −1 1
18 ∗ 10 =0
−1 0.618 𝜙
Note: In , j = DOF & i = mode number. e.g., ϕ21 represent mode shape coordinate corresponding
to DOF 2 while structure is vibrating in mode 1
Second mode shape ϕ
(2 − 1.111 ∗ 10 ∗ 2356.44 −1 𝜙
18 ∗ 10 =0
−1 1 − 1.111 ∗ 10 ∗ 2356.44 𝜙
−0.618 −1 1
Taking 𝜙 = 1; 18 ∗ 10 =0
−1 −1.618 𝜙
To increase numerical precision, the unit element in the vectors is associated with the largest amplitude. Modal
vectors are therefore represented in a non-dimensional format by dividing each element by the maximum value
in the vector. The mode shape is said to be normalized. There exist other forms of normalization,
18
Problem 8.1:Normalization of Modes

Scale factors are sometimes applied to natural modes to standardize their elements
associated with various DOFs. This process is called normalization. There are various
methods of normalization

1. Largest term in the vector is 1.0 (Easy to visualize, all coefficients between –1 and
+1, used in these lectures)

2.The first term in the vector is 1.0

3.The Euclidean norm of the vector ϕ is 1.0. i.e. 𝜙 𝜙 = 1 (Most eigenvalue

subroutines, used by mathematicians)

4.The generalized mass is 1.0. i. e. 𝜙 𝑚 𝜙 = 1. Easy to visualize, all coefficients

between –1 and +1. Used in computer programming.
19
Problem 8.1:Normalization of Mode shapes
Mode shapes being normalized by taking greatest floor term taken as 1 (or -1) for all the modes
1.000

0.618

-0.618

1.00

20
Problem 8.1: Modal and Spectral Matrices
All mode shapes information are assembled in a mode shapes matrix where each column n
represents a vibration mode shape . We have

where j indicates the DOFs.

Where is called modal matrix for the eigenvalue problem of Eq 3

The N eigenvalues ωn2 can be assembled into a diagonal matrix Ω2, which is known as the spectral
matrix of the eigenvalue problem, Eq 3

Each eigenvalue and eigenvector satisfies Eq.3, ϕ ϕ .

By using the modal and spectral matrices, it is possible to assemble all of such relations (n = 1, 2, . . . , N)
into a single matrix equation:
21
Problem 8.2 Calculate the natural frequencies and associated mode shapes of the two-
story rigid frame shown in Figure, accounting for rotation DOFs. Neglect the rotational mass
moment of inertia of lumped masses. The condensed mass and stiffness matrices were
determined in Problem 7.4.

𝑢 𝑢

𝑢
𝑢

22
Problem 8.2:Solution

The condensed stiffness matrix from problem 7.4 is

23
Problem 8.2:Solution (contd).

We can compare the natural frequencies for the building with flexible beams to those obtained for the
same building, but assuming beams that are rigid in flexure and calculated in previous problem (8.1):

The beam flexibility reduces lateral stiffness of the building and therefore reduces the natural frequencies.

24
Problem 8.2:Fundamental mode shape.

Translational DOF.
ϕ

1.000

0.503

25
Problem 8.2:Fundamental mode shape.
Rotational DOF. The relationship between the translation and
rotational DOFs is given by equation VIII, Module 7
𝐮

𝜙 = 1.00 𝜙
𝜙
𝐤 =

Replacing , we get
𝜙
= 0.503 𝜙 𝜙

Fundamental mode shape

Problem 8.2: Second mode shape.
Translational DOF. ϕ

-0.503

1.00

27
Problem 8.2: Second mode shape.
Rotational DOF.
Replacing , we get

𝜙
𝜙

𝜙 = −0.503

𝜙 𝜙
𝜙 = 1.00

Second mode shape

28
Problem 8.3:
Consider a rigid square slab supported by 3 rectangular columns. Assume that all columns are
the same size and have the same orientation. b = d = 20 ft. weight of slab = 40 kip. Height of
each column = 12 ft. Ix = 1728 in.4 Ix = 4Iy . E=4000 ksi. Calculate natural frequencies and
associated mode shapes
Assume that: (a) the roof slab is rigid in its plane; (b) the torsional stiffness of the columns is
negligible; (c) the columns are rigid axially and (d) the beams are rigid in flexure
Neglect the mass of the columns.
Solution 𝑢 𝑦

𝑥
𝑢 𝑢

29
Problem 8.3:Stiffness matrix
k
240" 𝑘 𝐹 = 0,⇒ 𝑘 = 3𝑘 = 20.82 ;
𝑘 in.
𝑘
ux=1
𝑘 𝑘 = 0;
240"
𝑘 𝑘

𝑀 = 0 ⇒𝑘 = 𝑘 ∗ 120 − 2 ∗ 𝑘 ∗ 120 = −832.8

uy=1 𝑘 = 0;
𝑘 𝑘 𝑘
k
𝑘 𝐹 = 0,⇒ 𝑘 = 3𝑘 = 83.28 ;
in.

𝑘 𝑘 𝑀 = 0, 𝑘 = 𝑘 ∗ 120 − 2 ∗ 𝑘 ∗ 120 = −120𝑘 = −3331.2

30
Problem 8.3:Stiffness matrix
𝑘 ∗ 120

𝑘 ∗ 120
240"

240" 𝑘 ∗ 120

𝑘 ∗ 120 𝑢
𝑘 ∗ 120 =1
𝑘 ∗ 120

𝐹 = 0,𝑘 = −120𝑘 = −832.8 ; 𝐹 = 0, 𝑘 = −120𝑘 = −3331.2

𝑀=0 ⇒𝑘 = 3 ∗ 120𝑘 ∗ 120 + 3 ∗ 120𝑘 ∗ 120 = 43,200 𝑘 + 𝑘 = 1.5 ∗ 10

31
Mass matrix
k. sec
𝑚 = 0.1035 ;
in.
𝑚 0 0 0.1035 0 0
𝐦= 0 𝑚 0 = 0 0.1035 0
0 0 𝐼 0 0 993.6

EOM for undamped free vibration to determine natural frequencies

It can be observed that multiplying any row of k and m result in coupled equations (each equation
involves displacement at least along 2 DOFs (for instance multiplication of first row result in an eqn
that involves ux and uθ ) and therefore natural frequencies and associated mode shapes shall be
determined by eigen value problem as used in previous 2 problems
32
Problem 8.4:Natural frequencies
20.82 0 −832.8 0.1035 0 0
𝐤= 0 83.82 −3331.2 𝐦= 0 0.1035 0
−832.8 −3331.2 1.5 ∗ 10 0 0 993.6

20.82 − 0.1035𝜔 0 −832.8

𝐤−𝜔 𝐦= 0 83.82 − 0.1035𝜔 −3331.2
−832.8 −3331.2 1.5 ∗ 10 − 993.6𝜔

20.82 − 0.1035𝜔 0 −832.8

det 𝐤 − 𝜔 𝐦 = 0 83.82 − 0.1035𝜔 −3331.2 =𝟎
−832.8 −3331.2 1.5 ∗ 10 − 993.6𝜔

⇒ 10.64 𝑥 − 26,829.35 𝑥 + 16,721,751.9 𝑥 − 2,341,794,088 = 0

33
 Problem 8.4:Mode shapes
Fundamental Mode shape: ϕ

20.82 − 0.1035 ∗ 198.02 0 −832.8 𝜙

0 83.82 − 0.1035 ∗ 198.02 −3331.2 𝜙 =0
−832.8 −3331.2 1.5 ∗ 10 − 993.6 ∗ 198.02 𝜙
0.325 0 −832.8 1
0 63.33 𝜙 First row: 0.325 − 832.8𝜙 = 0 ⇒ 𝜙 = 0.00039
⇒ −3331.2 =0
−832.8 −3331.2 1,303,247 𝜙 Second row: 63.33𝜙 − 3331.2 ∗ 𝜙 = 0 ⇒ 𝜙 = 0.0205

Second Mode shape: ϕ

−48.91 0 −832.8 1 First row: −48.91 − 832.8𝜙 =0⇒𝜙 = −0.0587
0 14.09 −3331.2 𝜙 =0
−832.8 −3331.2 Second row: 14.09𝜙 − 3331.2 ∗ 𝜙 =0⇒𝜙 = −13.87
830,612 𝜙
ϕ
Third Mode shape:
−149.93 0 −832.8 1 First row: −149.93 − 832.8𝜙 =0⇒𝜙 = −5.555
0 −86.90 −3331.2 𝜙 =0
Second row: −86.9𝜙 − 3331.2 ∗ 𝜙 =0⇒𝜙 = 212.94
−832.8 −3331.2 −139,241 𝜙 34
 Problem 8.4:Mode shapes
𝜙 = 0.0205 𝜙 = 0.00039
Fundamental mode
𝜙 =1

𝜙 =1
𝜙 = 0.072 𝜙 = 0.00423
Second mode

𝜙 = 0.0261
𝜙 =1 𝜙 = 0.0047

Third mode

35
Problem 8.4: Figure shows the plan view of a one-story building of Problem 7.6. The structure
consists of a roof, idealized as a rigid diaphragm, supported on three frames, A, B, and C, as shown.
The roof weight is uniformly distributed and has a magnitude of 100 lb/ft.2 The lateral stiffnesses of
the frames are ky = 75 kips/ft for frame A, and kx = 40 kips/ft for frames B and C.
The plan dimensions are b = 30 ft and d = 20 ft, the
eccentricity is e = 1.5 ft, and the height of the building 30 ft
is 12 ft. Determine the natural periods and modes of
vibration of the structure. 𝑘 = 40 𝑘/ft

𝑘
= 75 𝑘/ft
Solution: Selecting DOFs 20 ft
𝑘 = 40 𝑘/ft
𝑢 𝑦

𝑢 𝑢𝑥 𝑒 = 1.5 ft

36
Problem 8.4:Stiffness and Mass matrix
𝑑 20
𝑘 +𝑘 0 (𝑘 −𝑘 ) 40+ 40 0 (40 − 40)
2 2
𝐤= 0 𝑘 −𝑘 𝑒 = 0 75 −75 ∗ 1.5
𝑑 𝑑 𝑑 20
𝑘 −𝑘 −𝑘 ∗𝑒 (𝑘 +𝑘 ) +𝑘 𝑒 40 − 40 −75 ∗ 1.5 40 + 40 + 75 ∗ 1.5
2 4 2 4
𝑊 30 ∗ 20 ∗ 100/1000 kip k − sec
𝑚= = = 1.863 ;
𝑔 32.2 ft/sec ft
𝑚 𝑏 +𝑑 1.863 30 + 20
𝐼 = = = 201.83 k. ft. sec
12 12

37
Problem 8.4:Natural Frequencies & associated mode shapes

It can be observed that multiplying first row of k and m result in an uncoupled equation (i.e. it
involves response corresponding to only 1 DOF) as follows: .

However, multiplying second and third row results in

equations which are coupled (i.e each equation involves both
and and therefore natural frequencies and associated 𝜙 =0 𝜙 =0
mode shapes shall be determined by eigen value problem
𝑢 𝑦

𝑢𝑥
𝑢
𝜙 =1
In the other words given system will not displace at all in
direction ux in other modes 38
 Problem 8.4:Natural Frequencies & associated mode shapes
k and m for remaining 2 DOFs are:

39
Problem 8.4:Natural Frequencies & associated mode shapes
Mode shape associated with ϕ

ϕ
ϕ ϕ ⸪ ϕ
ϕ
Mode shape associated with ϕ

ϕ
ϕ ϕ ⸪ ϕ
ϕ 40
Problem 8.4:Natural Frequencies & associated mode shapes
As discussed at slide 31, system will not displace at all in 𝑥
direction in vibration modes other c 𝑥 𝜙 = 0.0945
Fundamental mode,
𝜙 =1

Second mode, 𝑢 𝑦
𝜙 =1
𝑢 𝑢𝑥

𝜙 = 0.0977
Third mode, 𝜙 =1

41
 Orthogonality of mode shapes
Applying eqn 3 to (any two) k and r mode such that , we have
ϕ ϕ ϕ ϕ
Transposing both sides of Eqn 5a and postmultiplying by ϕ
ϕ ϕ ϕ ϕ
Recalling that and due to symmetry =k and , the equation becomes
ϕ ϕ ϕ ϕ
Premultiplying both sides of Eqn 5b by ϕ gives ϕ ϕ ϕ ϕ
Comparing 6 and 7 results in ϕ ϕ ϕ ϕ ϕ ϕ
Since ϕ ϕ
This is the first orthogonality of eigenvectors, i.e. the different mode shapes are orthogonal with respect
to the mass matrix. This means that the off-diagonal elements of the product 𝐓 are zero. 42
 Orthogonality of mode shapes

The second orthogonality can be given from this relation. Pre-multiplying Eqn 5b with ϕ ,we have
ϕ ϕ ϕ ϕ
Substituting ϕ ϕ (i.e.eqn 8) in above eqn results in ϕ ϕ
This equation shows that the different mode shapes are orthogonal with respect to stiffness matrix.

COMMENT.– The mode shape orthogonality properties imply that the work carried out by the
inertia and elastic forces for mode k on the displacements of mode r is equal to zero.

Using orthogonality of natural modes, it can be proved that 𝐓 and 𝐓 results in diagonal
matrices*, M and K, respectively.
* A diagonal matrix is one in which all off diagonal terms are zero. Thus [A]3x3 will be diagonal matrix if
𝐷 0 0
[A]= 0 𝐷 0
0 0 𝐷 43
Modal Decoupling of EOMs
 Coupled EOMs of MDOF systems
Consider free vibration in an undamped MDOF shear frame. The equations of motion for the
system in matrix form can be written as:

For a shear frame

1 2 2
2 2 3 3
3 3

1
2
3

45
 MDOF systems: Coupled EOMs
By substituting k, and m in Eqn 10, we get

1 1 1 2 2 1
2 2 2 2 3 3 2
3 3
3 3 3

By multiplying first row, we get EOM for DOF 1

1 1+ 1 1 2 1 2 2

By multiplying second row, we get EOM for DOF 2

2 2 − 2 1 2 2 3 2 3 3

Similarly EOM for DOF3 3 3 3 2 3 3

It can be observed from the above given EOMs that each equation contains terms for displacements at
two or more DOFs. Hence, these equations are “coupled” and cannot be solved independently. 46
 Modal decoupling of the EOMs: Transformation from Geometric
coordinate system,u, to Modal coordinate system, q
Process of Uncoupling EOM (Eqn 11) is carried by a procedure known as “Modal Decoupling” of
EOMs. This is done by transformation of displacement coordinates from Geometric coordinates
(displacements corresponding to DOFs), u(t), to Modal coordinates (Amplitudes of natural modes)
q(t) using relation ,
A displacement vector for a MDOF system, u , can be obtained by superposition of the mode shapes, each
scaled by a specific modal amplitude. The procedure is illustrated in Figure for a 3 DOF shear frame.
u32
u31 u33
Displ along DOF 3 when system is u3
vibrating in mode no. 2

u21 u22 u23 u2

u11 u12 u13 u1

Deformed shaped obtained by the summation of individual modal components . 47

 Modal decoupling of the EOMs: Transformation from Geometric
coordinate system ,u, to Modal coordinate system, q
The displacement vector corresponding to DOFs of a MDOF vibrating in nth vibration mode, 𝑛,are given by
the product of the modal displacement, qn and the mode shape vector, , that is
to DOF 1, +
(see figure on previous slide)
is the amplitude of displacement (modal amplitude) when system vibrates in fundamental mode
+
+
𝑢 𝜙 𝜙 𝜙 𝑞 𝑞
These equations in matrix form can be written as: 𝑢 = 𝜙 𝜙 𝜙 𝑞 or {𝑢} = 𝜙 𝜙 𝜙 𝑞
𝑢 𝜙 𝜙 𝜙 𝑞 𝑞

For a MDOF with N−DOFs, the relation can be written as . 𝟏𝟑

The equation in vector form is written as 14
where q is the modal (generalized)coordinates vector and is the mode shapes matrix. Matrix is
therefore used to transform the geometric displacement coordinates, u, into modal displ. coordinates q.
48
 Modal decoupling of the EOMs: How to find q for each mode
Any modal coordinate can be easily calculated using the orthogonality properties of mode shapes as follows.
Premultiply both side of eqn 14 ( ) by , we get
However,

Since mode shape are orthogonal i.e. ϕ 𝑇 ϕ

⸫ ,from which Where 𝜙 𝐦𝜙 = 𝑀

is known as generalized mass for nth mode

Once initial conditions for nth mode are determined using

Eqns 16 a and 16 b using given values of (0) & (0),
EOM for nth mode is solved.
Geometric displacements in MDOF system for nth vibration mode are determined using . It is
𝐦𝐮
to be noted that effect of normalization cancels at this stage as
𝐦 49
 Modal EOMs
By substituting values of in Eqn 10, we get
Since from E

Pre-multiplying both sides by , we get

As discussed on slides 42 and 43 different modes are orthogonal w.r.t m and k due to which
𝐓
and 𝐓 results in diagonal matrices (i.e all off diagonal elements are zero).

Where [M],[K] and are known as generalized mass and stiffness matrices
* An orthogonal matrix is s a matrix whose transpose is equal to the inverse of the matrix . Product of an orthogonal matrix
50
and its transpose gives an identity matrix (i.e. matrix with diagonal elements as 1 and all off diagonal elements are zero)
 Modal EOMs
Since [M],[K] are diagonal matrices, therefore ,For given 3 story shear frame
1 1
2 2
3 3

By multiplying first row, we get EOM for mode 1 1 1(t)+ 1 1

By multiplying second row, we get EOM for mode 2 2 2(t)+ 2 2

By multiplying third row, we get EOM for mode 3 3 3(t)+ 3 3

For Mode n 𝑛 𝑛(t)+ 𝑛 𝑛

Where 𝑛 𝑛
Where 𝑛, & 𝑛 are referred as Modal mass and modal stiffness for nth mode, respectively
51
 Modal Superposition for free vibration response
It is a procedure of superposing modal coordinates to determine the solution of coupled EOMs. In
order to uncouple EOMs for free vibration in a MDOF system, following steps are followed
Step 1: Determine ω & ϕ using Eigen value problem ϕ=0
Step 2:Transform u to q using relation,
𝐦𝐮
Step 3: By employing Modal orthogonality properties of mass matrix, determine

First 3 steps are discussed in preceding slides. Whereas steps 4 and 5 are discussed in slides to follow

Step 4: After converting initial conditions by using the above eqn, solve
̇
EOMs for various modes. For nth mode, the solution is

Step 5: After solving EOMs for all modes, we can develop EOM for various DOFs using the same
relation which was used to uncouple EOMs 52
 𝑢 0 = 0.02 m
Problem 8.5 Calculate, by modal superposition, the free-vibration
response of the two DOFs for the two-story building of Problem 8.1.
The initial conditions are as follows: u1 = 0.02 m and u2 = 0.02 m.
Assume that the system is undamped.
𝑢 0 = 0.02 m
Solution 𝐮 0 =
𝑢 (0)
=
0.02
& 𝐮̇ (0) =
𝑢̇ 0
=
0
𝑢 (0) 0.02 𝑢̇ 0 0

From Problem 8.1, we have

Step 1: Determine Generalized mass for each mode

53
Step 2: Determine contribution of each mode in given initial conditions.
𝜙𝐦𝐮 𝜙 𝐦𝐮̇ 𝜙 𝐦𝐮 𝟎 𝜙 𝐦𝐮̇ (𝟎)
Eqn 16: 𝑞 (𝑡) = & 𝑞̇ (𝑡) = or 𝑞 (0) = & 𝑞̇ (0) =
𝑀 𝑀 𝑀 𝑀
𝜙 𝐦𝐮 0 𝜙 𝐦𝐮̇ (0)
⇒ 𝑞 (0) = & 𝑞̇ (0) =
𝑀 𝑀
20,000 0 0.02
0.618 1
0 20,000 0.02
⇒𝑞 0 = = 0.02342 m = 23.42 mm
27,629
20,000 0 0
0.618 1
0 20,000 0
& 𝑞̇ (0) = =0
27,639 23.42 mm −3.42 mm
0.618 0.01447 20 mm
⇒𝐮 0 =𝑞 0 𝜙 = 0.02342 = m
1 0.02342

20,000 0 0.02
1 −0.618 5.53 mm
0 20,000 0.02 20 mm
and 𝑞 0 =
27,639 14.47 mm
= 0.00553m = 5.53 mm & 𝑞̇ (0) = 0
1 0.00553
⇒ 𝐮 0 = 𝑞 (0)𝜙 = 0.00553 = =
−0.618 −0.00342
Step 3: Solve EOM for each mode
Solution to EOM for undamped free vibration in n-th mode

We can observe from these results that the motion of each floor includes the contribution of each
vibration frequency of the structure. 55
Problem (contd).

Displacements of DOFs 1 and 2 generated by displacements of 20 mm applied at t = 0 to both DOFs

56
Problem 8.6 Determine the free vibration response of the two-story shear frame with initial
displacement u(0) = [1 -1]T . The natural frequencies and mode shape vectors as calculated using eigen
value problem are as follows

Solution

Step 1: Determine Generalized mass for each mode

57
 Problem 8.6:Solution 𝑢 0 =1
Step 2: Determine contribution of each mode in given initial conditions.
2𝑚 0 1
𝜙 𝐦𝐮 0 0.5 1
0 𝑚 −1
𝑞 0 = = = 0 & 𝑞̇ (0) = 0
𝑀 1.5𝑚
𝑢 0 =1

2𝑚 0 1
𝜙 𝐦𝐮 0 1 −1 3𝑚
0 𝑚 −1
𝑞 0 = = = = 1 & 𝑞̇ (0) = 0 𝐮 0 = 𝐮𝟏 0 + 𝐮𝟐 (0)
𝑀 3𝑚 3𝑚
1 1

1 1
0.50
𝐮𝟏 0 = 0 =0
1
Step 3: Solve EOM for each mode

1 1
𝐮𝟐 0 = 1 =
−1 −1
 Problem 8.6:Solution

(a)

These solutions for 2 1 2 had

been plotted in Fig. a and b. Notethat 1
implies that the first mode has no contribution to
the response and the response is due entirely to
the second mode. Such is the case because the
initial displacement is proportional to the second
mode and hence orthogonal to the first mode. (b)
Free vibration of 2 story shear frame for given initial conditions:
(a) modal coordinate q2(t); (b) displacement history. 59
 Equivalent static forces & Base shear

Equivalent static force vector for n-th mode can be determined by using the basic relation

Inserting both equations in Eqn 17 results in =

Similary base shear in

60
Problem 8.7: Calculate, by modal superposition, the equivalent static forces in the two-story building
of Problem 8.5.
𝑢 0 = 0.02 m

Given data

𝑢 0 = 0.02 m
Problem 8.7: Solution
Eqn 1
1 0 0.618
𝐟 = 𝜔 𝐦 𝜙 𝑞 𝑡 = 18.54 ∗ 20,000 ∗ 0.02342 Cos 18.54𝑡
0 1 1

𝑓 = 161Cos 18.54𝑡

𝑓 = 99.5Cos 18.54𝑡

+
 1 0 1
𝐟 = 𝜔 𝐦 𝜙 𝑞 𝑡 = 48.54 ∗ 20 ∗ 10 ∗ 0.00553Cos 48.54𝑡
0 1 −0.618

𝑓 = −161Cos 48.54𝑡

𝑓 = 260.5Cos 48.54𝑡

+
 𝑓 = 𝑓 + 𝑓 = 99.5Cos 18.54𝑡 + 260.5Cos 48.54𝑡

𝑓 =𝑓 +𝑓 = 161Cos 18.54𝑡 − 161Cos 48.54𝑡

+
Once equivalent static forces are known , static analysis is carried out for the design purpose.
 Exercise 8
1. Determine the free vibration response of the structures of Exercise Problem 7.1 if they are
displaced as shown in Fig. a, b, and c and released. Plot floor displacements versus t/T1 and
comment on the relative contributions of the three vibration modes to the response that was
produced by each of the three initial displacements. Neglect damping.

Answer

Figure a

Figure c
Figure b
 Exercise 8

2. For the system defined in Exercise Problem 7.3, m = 90 kips/g, k = 1.5 kips/in., and b = 25 ft.
Determine the natural vibration frequencies
. and modes.

Answer

0.317 1 0.862
𝚽 = −0.317 1 −0.862
1 0 −1