DYNAMICS OF STRUCTURES

(INS3562)
- LESSON 6 -

Assist Prof. Dr. Muzaffer BÖREKÇİ

Email: dynamicsofstructuresmb@[Link]
 Natural Vibration Periods and Modes
of
MDOF Systems
For MDOF systems EOM is written in matrix form:
𝒎 𝒖 + 𝒌 𝒖 = 𝟎

Time variation of displacement can be given as:

For SDOf systems:
𝒖 = 𝜱 𝐬𝐢𝐧(𝝎𝒕 + 𝜽)
u(t) = Xsin(ωt+Φ)
• {Φ}: constant mode amplitudes
• θ: phase angle

𝒖 = 𝝎 𝜱 𝒄𝒐𝒔(𝝎𝒕 + 𝜽)
𝒖 = −𝝎𝟐 𝜱 𝒔𝒊𝒏(𝝎𝒕 + 𝜽)

Assist. Prof. Dr. Muzaffer BÖREKÇİ Dynamics of Structures

 𝒖 = 𝜱 𝐬𝐢𝐧(𝝎𝒕 + 𝜽)

𝒖 = −𝝎𝟐 𝜱 𝒔𝒊𝒏(𝝎𝒕 + 𝜽)

−𝝎𝟐 𝒎 𝜱 𝒔𝒊 𝒏 𝝎𝒕 + 𝜽 + 𝒌 𝜱 𝒔𝒊𝒏 𝝎𝒕 + 𝜽 = 𝟎

−𝝎𝟐 𝒎 𝜱 + 𝒌 𝜱 = 𝟎

𝒌 − 𝝎𝟐 𝒎 𝜱 = 𝟎

This algebric equation is called

«matrix eigenvalue problem».
[k] and [m] are known, the problem is to determine the scalar ω2
and vector {Φ}.
Eigenvalue equation can be interpreted as a set of N
homogenous algebraic equations for the N elements Φj (j = 1, 2,
….., N).
Assist. Prof. Dr. Muzaffer BÖREKÇİ Dynamics of Structures
One of the solution of the equation is {Φ} = 0; however this
means there is no motion. Thus, nontrivial solution is:
«characteristic equation» in mathematics
𝒌 − 𝝎𝟐 𝒎 = 𝟎
«frequency equation» in structural dynamics

Frequency equation has N real and positive roots for ω2.

The N roots ω2 give N natural frequencies ω of vibration.
These roots of the frequency equation are also known as
«eigenvalues» or «characteristic values».
When a natural frequency ω is known, «eigenvalue
equation» can be solved for the corresponding vector {Φ}
to within a multiplicative constant.

Assist. Prof. Dr. Muzaffer BÖREKÇİ Dynamics of Structures

It is very important, the eigenvalue problem does not fix the
absolute amplitude of the vectors {Φ}, only the shape of the
vectors is given by the relative values of the N displacements Φj
(j = 1, 2, ….., N).
Corresponding to the N natural vibration frequencies ω of an N-
DOF system, there are N independent vectors {Φ} which are
known as «natural modes of vibration» or «natural mode shapes
of vibration». These vectors are also known as «eigenvectors».

Summary:
 Number of ω of a system with N-DOF is N

(ω1 < ω2 < ω3 < …. < ωN)

 Number of T of a system with N-DOF is N

(T1 > T2 > T3 > …. > TN)

Assist. Prof. Dr. Muzaffer BÖREKÇİ Dynamics of Structures

EXAMPLE 1:
m
a) Estimate mass matrix ([m]).
k k b) Estimate stiffness matrix ([k]).
2m c) Estimate frequency equation.
d) Estimate freuqencies and mode amlitudes.
k k (m = 10 kNs2/m ; k = 1000 kN/m)

2-D Shear Frame

Assist. Prof. Dr. Muzaffer BÖREKÇİ Dynamics of Structures

Solution a: (Mass Matrix)
m

k k
2m

k k

𝒎𝟏 𝟎 𝟐𝒎 𝟎 𝟐𝟎 𝟎
a) 𝒎 = = =
𝟎 𝒎𝟐 𝟎 𝒎 𝟎 𝟏𝟎

Assist. Prof. Dr. Muzaffer BÖREKÇİ Dynamics of Structures

Solution b: (Siffness Matrix)
m

k k 𝐤 𝟐 = 𝟐𝐤
2m

k k 𝐤 𝟏 = 𝟐𝐤

𝒌𝟏 + 𝒌𝟐 −𝒌𝟐 𝟒𝒌 −𝟐𝒌 𝟒𝟎𝟎𝟎 −𝟐𝟎𝟎𝟎

b) 𝒌 = = =
−𝒌𝟐 𝒌𝟐 −𝟐𝒌 𝟐𝒌 −𝟐𝟎𝟎𝟎 𝟐𝟎𝟎𝟎

Assist. Prof. Dr. Muzaffer BÖREKÇİ Dynamics of Structures

Solution c: (Frequency Equation)
m

k k 𝐤 𝟐 = 𝟐𝐤
2m

k k 𝐤 𝟏 = 𝟐𝐤

𝟒𝒌 −𝟐𝒌 𝟐𝒎 𝟎
c) 𝒌 − 𝝎𝟐 𝒎 = − 𝝎𝟐
−𝟐𝒌 𝟐𝒌 𝟎 𝒎

Assist. Prof. Dr. Muzaffer BÖREKÇİ Dynamics of Structures

Solution d: (Frequencies)

𝟒𝒌 − 𝝎𝟐 𝟐𝒎 −𝟐𝒌 =𝟎
d) 𝟐
−𝟐𝒌 𝟐𝒌 − 𝝎 𝒎

𝟒𝒌 − 𝝎𝟐 𝟐𝒎 𝟐𝒌 − 𝝎𝟐 𝒎 − −𝟐𝒌 −𝟐𝒌 = 𝟎

𝟐
𝟐𝒎𝟐 𝝎𝟐 − 𝟖𝒌𝒎𝝎𝟐 + 𝟒𝒌𝟐 = 𝟎

𝟐
𝟐 𝒌 𝟐 𝒌 𝒌 𝟏𝟎𝟎𝟎
𝝎𝟐 −𝟒 𝝎 +𝟐 =𝟎 𝝎𝟐 = 𝝀 = = 𝟏𝟎𝟎
𝒎 𝒎 𝒎 𝟏𝟎

𝟐
𝒌 𝒌
𝝀 𝟐−𝟒 𝝀+𝟐 =𝟎
𝒎 𝒎
𝝀𝟏 = 𝝎𝟐𝟏 = 𝟓𝟖. 𝟓𝟖 𝝎𝟏 = 𝟕. 𝟔𝟓𝟑 𝒓𝒂𝒅/𝒔𝒆𝒄
𝝀 𝟐 − 𝟒𝟎𝟎𝝀 + 𝟐𝟎𝟎𝟎𝟎 = 𝟎
𝝀𝟐 = 𝝎𝟐𝟐 = 𝟑𝟒𝟏. 𝟒𝟐 𝝎𝟐 = 𝟏𝟖. 𝟒𝟖 𝒓𝒂𝒅/𝒔𝒆𝒄

Assist. Prof. Dr. Muzaffer BÖREKÇİ Dynamics of Structures

Solution d: (Mode amplitudes)
d) 𝝎𝟏 = 𝟕. 𝟔𝟓𝟑 𝒓𝒂𝒅/𝒔𝒆𝒄 𝑻𝟏 = 𝟎. 𝟖𝟐 𝒔𝒆𝒄

𝝎𝟐 = 𝟏𝟖. 𝟒𝟖 𝒓𝒂𝒅/𝒔𝒆𝒄 𝑻𝟐 = 𝟎. 𝟑𝟒 𝒔𝒆𝒄

First Mode:
𝟒𝒌 − 𝝎𝟐 𝟐𝒎 −𝟐𝒌 Φ𝟏𝟏 𝟎
=
−𝟐𝒌 𝟐𝒌 − 𝝎𝟐 𝒎 Φ𝟐𝟏 𝟎

𝟐𝟖𝟐𝟖. 𝟒 −𝟐𝟎𝟎𝟎 Φ𝟏𝟏 𝟎

=
−𝟐𝟎𝟎𝟎 𝟏𝟒𝟏𝟒. 𝟐 Φ𝟐𝟏 𝟎
𝜱𝟐𝟏 = 𝟏. 𝟒𝟏𝟒𝟐

𝟐𝟖𝟐𝟖. 𝟒𝜱𝟏𝟏 − 𝟐𝟎𝟎𝟎𝜱𝟐𝟏 = 𝟎

−𝟐𝟎𝟎𝟎𝜱𝟏𝟏 + 𝟏𝟒𝟏𝟒. 𝟐𝜱𝟐𝟏 = 𝟎 𝜱𝟏𝟏 = 𝟏

𝜱𝟏𝟏 = 𝟏 (𝑨𝒔𝒔𝒖𝒎𝒑𝒕𝒊𝒐𝒏)

𝜱𝟐𝟏 = 𝟏. 𝟒𝟏𝟒𝟐

Assist. Prof. Dr. Muzaffer BÖREKÇİ Dynamics of Structures

Solution d: (Mode amplitudes)
d) 𝝎𝟏 = 𝟕. 𝟔𝟓𝟑 𝒓𝒂𝒅/𝒔𝒆𝒄 𝑻𝟏 = 𝟎. 𝟖𝟐 𝒔𝒆𝒄

𝝎𝟐 = 𝟏𝟖. 𝟒𝟖 𝒓𝒂𝒅/𝒔𝒆𝒄 𝑻𝟐 = 𝟎. 𝟑𝟒 𝒔𝒆𝒄

Second Mode:
𝟒𝒌 − 𝝎𝟐 𝟐𝒎 −𝟐𝒌 Φ𝟏𝟐 𝟎
=
−𝟐𝒌 𝟐𝒌 − 𝝎𝟐 𝒎 Φ𝟐𝟐 𝟎

−𝟐𝟖𝟐𝟖. 𝟒 −𝟐𝟎𝟎𝟎 Φ𝟏𝟐 𝟎

=
−𝟐𝟎𝟎𝟎 −𝟏𝟒𝟏𝟒. 𝟐 Φ𝟐𝟐 𝟎
𝜱𝟐𝟐 = −𝟏. 𝟒𝟏𝟒𝟐
−𝟐𝟖𝟐𝟖. 𝟒𝜱𝟏𝟐 − 𝟐𝟎𝟎𝟎𝜱𝟐𝟐 = 𝟎
−𝟐𝟎𝟎𝟎𝜱𝟏𝟐 − 𝟏𝟒𝟏𝟒. 𝟐𝜱𝟐𝟐 = 𝟎 𝜱𝟏𝟐 = 𝟏

𝜱𝟏𝟐 = 𝟏 (𝑨𝒔𝒔𝒖𝒎𝒑𝒕𝒊𝒐𝒏)

𝜱𝟐𝟏 = −𝟏. 𝟒𝟏𝟒𝟐

Assist. Prof. Dr. Muzaffer BÖREKÇİ Dynamics of Structures

 ORTHOGONALITY of MODES
𝒌 𝜱 𝒏 = 𝝎𝟐𝒏 𝒎 𝜱 𝒏 𝒌 𝜱 𝒓 = 𝝎𝟐𝒓 𝒎 𝜱 𝒓
nth mode rth mode

𝜱 𝑻 𝒌 𝜱 = 𝝎𝟐𝒓 𝜱 𝑻 𝒎 𝜱 Eq. (1)

𝒏 𝒓 𝒏 𝒓

𝜱 𝑻 𝒌 𝜱 = 𝝎𝟐𝒏 𝜱 𝑻 𝒎 𝜱 Eq. (2)

𝒓 𝒏 𝒓 𝒏

Since [k] and [m] are symmetric matrices, then [k] = [k]T and
[m] = [m]T. Therefore, Eq. (2) can be transposed and rewritten
as: 𝑻 𝟐 𝑻
𝜱 𝒏 𝒌 𝜱 𝒓 = 𝝎𝒏 𝜱 𝒏 𝒎 𝜱 𝒓 Eq. (3)

Substract Eq.(1) from Eq.(3)

𝝎𝟐𝒏 − 𝝎𝟐𝒓 𝜱 𝑻
𝒏 𝒎 𝜱 𝒓 =𝟎
𝜱 𝑻 𝒎 𝜱 =𝟎
𝝎𝟐𝒏 ≠ 𝝎𝟐𝒓 𝒏 𝒓 should be zero
𝑻
Substitute 𝜱 𝒏 𝒎 𝜱 𝒓 into Eq.(1) Orthogonality
𝜱 𝑻 𝒌 𝜱 =𝟎 of modes
𝒏 𝒓

Assist. Prof. Dr. Muzaffer BÖREKÇİ Dynamics of Structures

GENERALIZED MASS and STIFFNESS MATRICES
 For r = n, orthogonality condition for the mass can be written
as: 𝑻 𝜱 𝒓 𝒎 𝜱 𝒓 = 𝑴𝒓

Mr: «Generalized mass» or «modal mass» for the rth mode

 For r = n, orthogonality condition for the mass can be written

as: 𝑻 𝜱 𝒓 𝒌 𝜱 𝒓 = 𝑲𝒓

Kr: «Generalized stiffness» or «modal stiffness» for the rth mode

𝜱 𝑻 𝒌 𝜱 = 𝝎𝟐𝒓 𝜱 𝑻 𝒎 𝜱
𝒓 𝒓 𝒓 𝒓

𝑲𝒓 = 𝝎𝟐𝒓 𝑴𝒓

𝑲𝒓
𝝎𝟐𝒓 =
𝑴𝒓

Assist. Prof. Dr. Muzaffer BÖREKÇİ Dynamics of Structures

 𝑲𝒓 = 𝝎𝟐𝒓 𝑴𝒓
This equation represents the natural frequency for the rth mode
in terms of the generalized mass and generalized stiffness of the
rth mode. It should be noted that Kr and Mr are non-negative
scalar quantities.
If the relationships of generalized mass and generalized
stiffness are applied along with the conditions of mass and
stiffness orthogonality, following matrices are estimated for the
entire system.
𝑴 = 𝜱𝑻𝒎 𝜱 𝑲 = 𝜱𝑻𝒌 𝜱
𝑴𝟏 𝟎 𝟎 ⋯ 𝟎 𝟎 𝑲𝟏 𝟎 𝟎 ⋯ 𝟎 𝟎
𝟎 𝑴𝟐 𝟎 ⋯ 𝟎 𝟎 𝟎 𝑲𝟐 𝟎 ⋯ 𝟎 𝟎
⋮ ⋮
𝑴 = 𝑲 =
⋮ ⋯⋯⋯⋯⋯⋯ ⋮ ⋯⋯⋯⋯⋯⋯
⋮ ⋮
𝟎 𝟎 𝟎 ⋯ ⋯ 𝑴𝒏 𝒏𝒙𝒏 𝟎 𝟎 𝟎 ⋯ ⋯ 𝑲𝒏 𝒏𝒙𝒏

Generalized mass matrix Generalized stiffness matrix

Assist. Prof. Dr. Muzaffer BÖREKÇİ Dynamics of Structures

 ANALYSIS of DYNAMIC RESPONSE
by
MODE SUPERPOSITION
The general matrix formulation of the equations of motion for an
undamped n-DOF system subjected to external excitation is
given below:

𝒎 𝒖 + 𝒌 𝒖 = 𝑭(𝒕)

These equations are «coupled equations»; that is, they cannot

be solved indepently, only simultaneously.

The mode superposition method cannot be applied to the

coupled system of equations. Thus, these equations should be
transformed to uncoupled form.

Assist. Prof. Dr. Muzaffer BÖREKÇİ Dynamics of Structures

Finding a coordinate system that exhibits neither form of coupling is the
essence of the mode superposition procedure. When these equations
are uncoupled, they may be decoupling of the equations of motion are
called «principle coordinates» or «normal coordinates».
If the principle modes (normal modes) of vibration for a MDOF system
are used as generalized coordinates for defining uncoupled, in these
coordinates each uncoupled equation may be solved independently as if
each equation pertained to an independent system having only one
degree of freedom.
Therefore, the individual responses of the n uncoupled equations of the
MDOF systems (one for each mode of vibration), to any form of
excitation may be determined by applying the techniques of analysis for
SDOF systems. The response of the MDOF system by mode
superposition is then defined as the summation of the responses of the
individual modes.
This procedure of dynamic analysis is referred to as the mode
superposition method, or simply modal analysis.

Assist. Prof. Dr. Muzaffer BÖREKÇİ Dynamics of Structures

To uncouple the differential equations, an alternative set of coordinates
can be used:
uN (t) ΦN1 ΦN2 ΦNN
N N
. .
. .
. .
. .
. .
u2 (t) = Φ21 x q1 (t) + Φ22 x q2 (t) + Φ2N
x qN (t)
2 2

u1 (t) Φ11 Φ1N

1 1 Φ12

𝒖(𝒕) = 𝜱 𝒒(𝒕) = 𝜱 𝟏 𝒒𝟏 𝒕 + 𝜱 𝟐 𝒒𝟐 𝒕 + 𝜱 𝟑 𝒒𝟑 𝒕 + ⋯ + 𝜱 𝑵 𝒒𝑵 𝒕

Assist. Prof. Dr. Muzaffer BÖREKÇİ Dynamics of Structures

 Mode Superposition Solution for Damped Systems
due to Support Motion
𝒎 𝒖(𝒕) + 𝒄 𝒖(𝒕) + 𝒌 𝒖(𝒕) = − 𝒎 𝑰𝒙 𝒖𝒈 (𝒕) Eq. (1)

• 𝒖𝒈 𝒕 : ground accelerarion
• 𝑰𝒙 : unit vector of dimension N

 Effective load acting on the mass of any floor «i»:

𝑭𝒆𝒇𝒇 (𝒕)𝒊 = 𝒎𝒊 𝒖𝒈 (𝒕)

 Effective load acting on the mass of each floor:

𝑭𝒆𝒇𝒇 (𝒕) = 𝒎 𝑰𝒙 𝒖𝒈 (𝒕)

Eq.(1) is coupled equation. Uncoupled version of the equation is given

below:
𝒎 𝜱 𝒒(𝒕) + 𝒄 𝜱 𝒒(𝒕) + 𝒌 𝜱 𝒒(𝒕) = − 𝒎 𝑰𝒙 𝒖𝒈 (𝒕)

Assist. Prof. Dr. Muzaffer BÖREKÇİ Dynamics of Structures

 𝜱 𝑻 𝒎 𝜱 𝒒(𝒕) + 𝜱 𝑻 𝒄 𝜱 𝒒(𝒕) + 𝜱 𝑻 𝒌 𝜱 𝒒(𝒕) = − 𝜱 𝑻 𝒎 𝑰𝒙 𝒖𝒈 (𝒕)

𝑴 𝒒(𝒕) + 𝑪 𝒒(𝒕) + 𝑲 𝒒(𝒕) = − 𝜱 𝑻 𝒎 𝑰𝒙 𝒖𝒈 (𝒕)

𝑪 = 𝜱 𝑻 𝒄 𝜱 is an nxn symmetric matrix that is diagonal for

only a special case of [c]. This special case of damping is
referred to as «classical damping» or «proportional damping»,
for which [c] is proportional to [m] and/or [k]. Thus, if it is
assumed that the damping matrix [c] is so constructed that
𝜱 𝑻 𝒄 𝜱 =𝟎 • 𝒓 ≠𝒔
𝒓 𝒔

𝑴 𝒒(𝒕) + 𝑪 𝒒(𝒕) + 𝑲 𝒒(𝒕) = − 𝜱 𝑻 𝒎 𝑰𝒙 𝒖𝒈 (𝒕)

𝑴𝒏 𝒒𝒏 𝒕 + 𝑪𝒏 𝒒𝒏 𝒕 + 𝑲𝒏 𝒒𝒏 (𝒕) = − 𝜱 𝑻 𝒎 𝑰𝒙 𝒖𝒈 (𝒕)
𝒏

𝑳𝒙𝒏
𝑴𝒏 𝒒𝒏 𝒕 + 𝑪𝒏 𝒒𝒏 𝒕 + 𝑲𝒏 𝒒𝒏 (𝒕) = −𝑳𝒙𝒏 𝒖𝒈 (𝒕)

Assist. Prof. Dr. Muzaffer BÖREKÇİ Dynamics of Structures

 𝑴𝒏 𝒒𝒏 𝒕 + 𝑪𝒏 𝒒𝒏 𝒕 + 𝑲𝒏 𝒒𝒏 (𝒕) = −𝑳𝒙𝒏 𝒖𝒈 (𝒕)

𝑪𝒏 𝑲𝒏 𝑳𝒙𝒏
𝒒𝒏 𝒕 + 𝒒𝒏 𝒕 + 𝒒𝒏 𝒕 = − 𝒖 (𝒕)
𝑴𝒏 𝑴𝒏 𝑴𝒏 𝒈

𝑳𝒙𝒏
Γ𝒙𝒏 = Mode participation factor
𝑴𝒏
𝒒𝒏 𝒕 + 𝟐𝜻𝒏 𝝎𝒏 𝒒𝒏 𝒕 + 𝝎𝟐𝒏 𝒒𝒏 𝒕 = −𝜞𝒙𝒏 𝒖𝒈 (𝒕)

𝒒𝒏 (𝒕) 𝒒𝒏 (𝒕) 𝒒𝒏 (𝒕)

+ 𝟐𝜻𝒏 𝝎𝒏 + 𝝎𝟐𝒏 = −𝒖𝒈 (𝒕)
𝜞𝒙𝒏 𝜞𝒙𝒏 𝜞𝒙𝒏

𝒅𝒏 𝒕 + 𝟐𝜻𝒏 𝝎𝒏 𝒅𝒏 𝒕 + 𝝎𝟐𝒏 𝒅𝒏 𝒕 = −𝒖𝒈 (𝒕)

𝒒𝒏 𝒕 = Γ𝒙𝒏 𝒅𝒏 (𝒕)
𝒅𝒏 𝒕 : modal displacement

𝒖𝒏 (𝒕) = 𝜱 𝒏 𝒒𝒏 (𝒕) 𝒖𝒏 (𝒕) = 𝜱 𝒏 𝜞𝒙𝒏 𝒅𝒏 (𝒕)

Assist. Prof. Dr. Muzaffer BÖREKÇİ Dynamics of Structures

For nth mode: 𝑵

𝑴𝒏 = 𝜱 𝑻
𝒏 𝒎 𝜱 𝒏 = 𝒎𝒊 𝜱𝟐𝒊𝒏 Modal mass
𝒊=𝟏
𝑵

𝑳𝒙𝒏 = 𝜱 𝑻 𝒎 𝑰𝒙 = 𝒎𝒊 𝜱𝒊𝒏
𝒏
𝒊=𝟏

𝑵
𝑳𝒙𝒏 𝒊=𝟏 𝒎𝒊 𝜱𝒊𝒏
𝜞𝒙𝒏 = = 𝑵 𝟐
𝑴𝒏 𝒊=𝟏 𝒎𝒊 𝜱𝒊𝒏

𝒇𝒔𝒏 (𝒕) = 𝒌 𝒖𝒏 (𝒕) → 𝒖𝒏 (𝒕) = 𝒌 −𝟏 𝒇𝒔𝒏 (𝒕)

{𝑓𝑠𝑛 𝑡 }: static equivalent seismic load (structural resistance)

𝒇𝒔𝒏 (𝒕) = 𝒌 𝜱 𝒏 Γ𝒙𝒏 𝒅𝒏 (𝒕)
𝒌 𝜱 𝒏 = 𝝎𝟐𝒏 𝒎 𝜱 𝒏

𝒇𝒔𝒏 (𝒕) = 𝝎𝟐𝒏 𝒎 𝜱 𝒏 Γ𝒙𝒏 𝒅𝒏 (𝒕)

𝒇𝒔𝒏 (𝒕) = 𝒎 𝜱 𝒏 Γ𝒙𝒏 𝒂𝒏 (𝒕)

𝒂𝒏 𝒕 = 𝝎𝟐𝒏 𝒅𝒏 (𝒕) Modal pseudo acceleration

Assist. Prof. Dr. Muzaffer BÖREKÇİ Dynamics of Structures

 𝒅𝒏 𝒕 + 𝟐𝜻𝒏 𝝎𝒏 𝒅𝒏 𝒕 + 𝝎𝟐𝒏 𝒅𝒏 𝒕 = −𝒖𝒈 (𝒕)
𝒅(𝒕) 𝒎𝒂𝒙 = 𝑺𝒅
𝒂(𝒕) 𝒎𝒂𝒙 = 𝑺𝒑𝒂 = 𝝎𝟐 𝑺𝒅
𝒖𝒏 (𝒕) 𝒎𝒂𝒙 = 𝜱 𝒏 Γ𝒙𝒏 𝑺𝒅𝒏
𝑵

𝑽𝒙𝒏 (𝒕) = 𝒇𝒙𝒊𝒏 → 𝑽𝒙𝒏 (𝒕) = 𝑰𝒙 𝑻 𝒇𝒔𝒏 (𝒕)

N 𝒊=𝟏

fxin 𝑽𝒙𝒏 𝒕 = 𝝎𝟐𝒏 𝑰𝒙 𝑻 𝒎 𝜱 𝒏 Γ𝒙𝒏 𝒅𝒏

i

𝑳𝒙𝒏 𝑳𝒙𝒏
𝑴𝒏
𝑳𝟐 𝑳𝟐
𝒙𝒏 𝒙𝒏
𝑽𝒙𝒏 𝒕 = 𝝎𝟐𝒏 𝒅𝒏 𝒕 = 𝒂 𝒕
𝑴𝒏 𝑴𝒏 𝒏

𝑳𝟐𝒙𝒏
𝑽𝒙𝒏 = 𝑺
𝑴𝒏 𝒑𝒂𝒏

Assist. Prof. Dr. Muzaffer BÖREKÇİ Dynamics of Structures

 𝑳𝟐𝒙𝒏
𝑽𝒙𝒏 = 𝑺
𝑴𝒏 𝒑𝒂𝒏

𝑽𝒙𝒏 = 𝑴𝒙𝒏 𝑺𝒑𝒂𝒏

𝑴𝒙𝒏 : Effective participating mass

𝑵

𝑳𝒙𝒏 = 𝒎𝒊 𝜱𝒊𝒏
𝒊=𝟏
𝑵
𝑳𝒙𝒏 𝒊=𝟏 𝒎𝒊 𝜱𝒊𝒏
𝜞𝒙𝒏 = = 𝑵 𝟐 Mode participation factor
𝑴𝒏 𝒊=𝟏 𝒎𝒊 𝜱𝒊𝒏

𝑵 𝟐
𝑳𝟐𝒙𝒏 𝒎 𝜱
𝒊=𝟏 𝒊 𝒊𝒏
𝑴𝒙𝒏 = = 𝑵 𝟐
Effective participating mass
𝑴𝒏 𝒊=𝟏 𝒎𝒊 𝜱𝒊𝒏

𝑴𝒙𝒏
𝜸𝒙𝒏 = Effective participating mass ratio
𝑴𝒕
𝑴𝒕 = 𝒎𝒊 Total mass 𝜸𝒙𝒏 = 𝟏

Assist. Prof. Dr. Muzaffer BÖREKÇİ Dynamics of Structures

 For a 3D Structure
Summary:
1. Modal Properties:
1.1 Modal Mass of Any Mode:
𝑴𝒓 = 𝜱 𝑻 𝒎 𝜱
𝒓 𝒓

𝑴𝒓 = 𝒎𝒊 𝜱𝟐𝒙𝒊𝒓 + 𝒎𝒊 𝜱𝟐𝒚𝒊𝒓 + 𝒎𝒊𝜽 𝜱𝟐𝜽𝒊𝒓

1.2 Modal Participation Factors:

𝑵
𝑳𝒙𝒓 𝒊=𝟏 𝒎𝒊 𝜱𝒙𝒊𝒓
𝜞𝒙𝒓 = = 𝑵 𝟐 𝟐
𝑴𝒓 𝒊=𝟏(𝒎𝒊 𝜱𝒙𝒊𝒓 + 𝒎𝒊 𝜱𝒚𝒊𝒓 + 𝒎𝒊𝜽 𝜱𝟐𝜽𝒊𝒓 )
𝑵
𝑳𝒚𝒓 𝒊=𝟏 𝒎𝒊 𝜱𝒚𝒊𝒓
𝜞𝒚𝒓 = = 𝑵 𝟐 𝟐
𝑴𝒓 𝒊=𝟏(𝒎𝒊 𝜱𝒙𝒊𝒓 + 𝒎𝒊 𝜱𝒚𝒊𝒓 + 𝒎𝒊𝜽 𝜱𝟐𝜽𝒊𝒓 )
𝑵
𝑳𝜽𝒓 𝒊=𝟏 𝒎𝒊𝜽 𝜱𝜽𝒊𝒓
𝜞𝜽𝒓 = = 𝑵 𝟐 𝟐
𝑴𝒓 𝒊=𝟏(𝒎𝒊 𝜱𝒙𝒊𝒓 + 𝒎𝒊 𝜱𝒚𝒊𝒓 + 𝒎𝒊𝜽 𝜱𝟐𝜽𝒊𝒓 )

Assist. Prof. Dr. Muzaffer BÖREKÇİ Dynamics of Structures

 For a 3D Structure
Summary:
1.3 Effective Participating Masses:
𝑵 𝟐
𝑳𝟐𝒙𝒓 𝒎 𝜱
𝒊=𝟏 𝒊 𝒙𝒊𝒓
𝑴𝒙𝒓 = = 𝑵 𝟐 𝟐 𝟐
𝑴𝒓 𝒊=𝟏(𝒎𝒊 𝜱𝒙𝒊𝒓 + 𝒎𝒊 𝜱𝒚𝒊𝒓 + 𝒎𝒊𝜽 𝜱𝜽𝒊𝒓 )

𝑵 𝟐
𝑳𝟐𝒚𝒓 𝒎 𝜱
𝒊=𝟏 𝒊 𝒚𝒊𝒓
𝑴𝒚𝒓 = = 𝑵 𝟐 𝟐 𝟐
𝑴𝒓 𝒊=𝟏(𝒎𝒊 𝜱𝒙𝒊𝒓 + 𝒎𝒊 𝜱𝒚𝒊𝒓 + 𝒎𝒊𝜽 𝜱𝜽𝒊𝒓 )

𝑵 𝟐
𝑳𝟐𝜽𝒓 𝒎 𝜱
𝒊=𝟏 𝜽𝒊 𝜽𝒊𝒓
𝑴𝜽𝒓 = = 𝑵 𝟐 𝟐 𝟐
𝑴𝒓 𝒊=𝟏(𝒎𝒊 𝜱𝒙𝒊𝒓 + 𝒎𝒊 𝜱𝒚𝒊𝒓 + 𝒎𝒊𝜽 𝜱𝜽𝒊𝒓 )

1.4 Participating Mass Ratios:

𝑴𝒙𝒓 𝑴𝒚𝒓 𝑴𝜽𝒓
𝜸𝒙𝒓 = 𝜸𝒚𝒓 = 𝜸𝜽𝒓 =
𝒎𝒊 𝒎𝒊 𝒎𝜽𝒊

Assist. Prof. Dr. Muzaffer BÖREKÇİ Dynamics of Structures

 MODAL CONBINATION RULES
1) Square Root of Sum of the Squares (SRSS):
𝑺𝑴

𝝉𝒋,𝒎𝒂𝒙 = 𝝉𝟐𝒋𝒏,𝒎𝒂𝒙 SM: Sufficient mode number

𝒏=𝟏

2)Complete Quadratic Combination (CQC):

It is seen that SRSS gives erroneous results when the
combination of results of modes of which have close frequncies.

𝑺𝑴 𝑺𝑴

𝝉𝒋,𝒎𝒂𝒙 = 𝝉𝒋𝒎,𝒎𝒂𝒙 𝝆𝒎𝒏 𝝉𝒋𝒏,𝒎𝒂𝒙

𝒎=𝟏 𝒏=𝟏

Assist. Prof. Dr. Muzaffer BÖREKÇİ Dynamics of Structures

2)Complete Quadratic Combination (CQC):
It is seen that SRSS gives erroneous results when the
combination of results of modes of which have close frequncies.
𝑺𝑴 𝑺𝑴

𝝉𝒋,𝒎𝒂𝒙 = 𝝉𝒋𝒎,𝒎𝒂𝒙 𝝆𝒎𝒏 𝝉𝒋𝒏,𝒎𝒂𝒙

𝒎=𝟏 𝒏=𝟏

𝟑/𝟐
𝟖𝜻𝟐 (𝟏 + 𝜷𝒎𝒏 )𝜷𝒎𝒏 𝝎𝒎
𝝆𝒎𝒏 = 𝟐
𝜷𝒎𝒏 =
𝟏− 𝜷𝟐𝒎𝒏 + 𝟒𝜻𝟐 𝜷 )𝟐 𝝎𝒏
𝒎𝒏 (𝟏 + 𝜷𝒎𝒏
τjm,max :displacement, velocity, acceleration, shear force, moment etc.
m and n: mode number
ρmn :cross correlation coefficient
𝟑/𝟐
𝟖(𝜻𝒎 𝜻𝒏 )𝟏/𝟐 (𝜻𝒎 + 𝜷𝒎𝒏 𝜻𝒏 )𝜷𝒎𝒏
𝝆𝒎𝒏 = 𝟐
𝟏 − 𝜷𝟐𝒎𝒏 + 𝟒𝜻𝒎𝜻𝒏 𝜷𝒎𝒏 𝟏 + 𝜷𝟐𝒎𝒏 + 𝟒(ζ𝟐𝒎 + ζ𝟐𝒏 )𝜷𝟐𝒎𝒏

If different damping ratios are used for each mode

Assist. Prof. Dr. Muzaffer BÖREKÇİ Dynamics of Structures

It is seen that SRSS gives erroneous results when the
combination of results of modes of which have close frequncies.
𝝎𝒎
𝜷𝒎𝒏 =
𝝎𝒏

NOTE:
According to Turkish Seismic Code 2018:

In case 𝜷𝒎𝒏 < 𝟎. 𝟖 for all considered modes, it is allowed to use SRSS.

Assist. Prof. Dr. Muzaffer BÖREKÇİ Dynamics of Structures

 EXAMPLE 2:  Column Dimension: 30x30 cmxcm
 Ec =2.8*107 kN/m2 .
 Design Spectrum:
SDS = 0.879
SD1 = 0.271
TA = 0.049 sec
TB = 0.271 sec
I =1 (Importance factor)
R=8
EI = ∞ (Rigid) m3 = 50 kNs2/m ζ = 0.05

3m
EI = ∞ (Rigid) m2 = 82.5 kNs2/m

3m
EI = ∞ (Rigid) m1 = 82.5 kNs2/m

3m

Assist. Prof. Dr. Muzaffer BÖREKÇİ Dynamics of Structures

 EXAMPLE 2:
Questions:
a) Determine the natural circular
frequencies and mode amplitudes of
the first 3 modes.
b) Prove that the multiplication of
orthogonal modes is zero.
c) Obtain the modal mass and modal
stiffness matrices.
d) Estimate total base shear of the
EI = ∞ (Rigid) m3 = 50 kNs2/m building due to design earthquake
given in TSCB 2018 using mode
3m
EI = ∞ (Rigid) m2 = 82.5 kNs2/m
superposition method.
e) Estimate the sufficient mode number.
3m
EI = ∞ (Rigid) m1 = 82.5 kNs2/m

3m

Assist. Prof. Dr. Muzaffer BÖREKÇİ Dynamics of Structures

 EXAMPLE 2:
Solution a):

0.3 ∗ 0. 33
12 ∗ 2.8 ∗ 107 ∗
12𝐸𝐼 12 (Stiffness of one column)
𝑘= 3 = = 8400 𝑘𝑁/𝑚
ℎ 33
𝑘1 = 𝑘2 = 𝑘3 = 9𝑘 = 9 ∗ 8400 = 75600 𝑘𝑁/𝑚 (Stiffness of one storey)

𝑘1 + 𝑘2 −𝑘2 0 151200 −75600 0

𝑘 = −𝑘2 𝑘2 + 𝑘3 −𝑘3 = −75600 151200 −75600
0 −𝑘3 𝑘3 0 −75600 75600

𝑚1 0 0 82.5 0 0
𝑚 = 0 𝑚2 0 = 0 82.5 0
0 0 𝑚3 0 0 50

Assist. Prof. Dr. Muzaffer BÖREKÇİ Dynamics of Structures

EXAMPLE 2:
Solution a):
𝑘 − 𝜔2 𝑚 = 0 (frequency equation)

151200 − 82.5𝜔2 −75600 0

−75600 151200 − 82.5𝜔2 −75600 =0
0 −75600 75600 − 50𝜔2

𝜔2 = 𝜆

151200 − 82.5𝜆 151200 − 82.5𝜆 75600 − 50𝜆 − 75600 75600 …

… + 75600 −75600 75600 − 50𝜆 = 0

340312.5 ∗ 𝜆3 − 1761952500 ∗ 𝜆2 + 22718556 ∗ 105 ∗ 𝜆 − 432081216 ∗ 106 = 0

 𝜆1 = (ω1)2 = 0.2291*103 (rad/sec)2 → ω1 = 15.1357 rad/sec

 𝜆2 = (ω2)2 = 1.7130*103 (rad/sec)2 → ω2 = 41.3883 rad/sec
 𝜆3 = (ω3)2 = 3.2354*103 (rad/sec)2 → ω3 = 56.8803 rad/sec

Assist. Prof. Dr. Muzaffer BÖREKÇİ Dynamics of Structures

EXAMPLE 2:
Solution a):
Mode amplitudes of the first mode
Φ11 0
𝑘 − 𝜔12 𝑚 Φ21 = 0  (ω1)2 = 0.2291*103
Φ31 0

151200 − 82.5 ∗ 0.2291 ∗ 103 Φ11 − 75600Φ21 = 0

• Φ11 = 1 (𝐴𝑠𝑠𝑢𝑚𝑝𝑡𝑖𝑜𝑛) • Φ21 = 1.75

𝜱𝟑𝟏 = 𝟐. 𝟎𝟔
−75600Φ21 + (75600 − 50 ∗ 0.2291 ∗ 103 )Φ31 = 0
• Φ21 = 1.75 • Φ31 = 2.06 𝜱𝟐𝟏 = 𝟏. 𝟕𝟓

𝜱𝟏𝟏 = 𝟏

Assist. Prof. Dr. Muzaffer BÖREKÇİ Dynamics of Structures

EXAMPLE 2:
Solution a):
Mode amplitudes of the second mode
Φ12 0
𝑘 − 𝜔22 𝑚 Φ22 = 0  (ω2)2 = 1.7130*103
Φ32 0

151200 − 82.5 ∗ 1.7130 ∗ 103 Φ12 − 75600Φ22 = 0

• Φ12 = 1 (𝐴𝑠𝑠𝑢𝑚𝑝𝑡𝑖𝑜𝑛) • Φ22 = 0.13
𝜱𝟑𝟐 = 𝟎. 𝟗𝟖
−75600Φ22 + (75600 − 50 ∗ 1.7130 ∗ 103 )Φ32 = 0
• Φ22 = 0.13 • Φ32 = −0.98 𝜱𝟐𝟐 = 𝟎. 𝟏𝟑

𝜱𝟏𝟐 = 𝟏

Assist. Prof. Dr. Muzaffer BÖREKÇİ Dynamics of Structures

EXAMPLE 2:
Solution a):
Mode amplitudes of the third mode
Φ13 0
𝑘 − 𝜔32 𝑚 Φ23 = 0  (ω3)2 = 3.2354*103
Φ33 0

151200 − 82.5 ∗ 3.2354 ∗ 103 Φ11 − 75600Φ21 = 0

• Φ13 = 1 (𝐴𝑠𝑠𝑢𝑚𝑝𝑡𝑖𝑜𝑛) • Φ23 = −1.53
𝜱𝟑𝟑 = 𝟏. 𝟑𝟒
−75600Φ23 + (75600 − 50 ∗ 3.2354 ∗ 103 )Φ33 = 0
• Φ23 = −1.53 • Φ33 = 1.34 𝜱𝟐𝟑 = 𝟏. 𝟓𝟑

𝜱𝟏𝟑 = 𝟏

Assist. Prof. Dr. Muzaffer BÖREKÇİ Dynamics of Structures

EXAMPLE 2:
Solution b):
First and second modes orthogonality:

𝟖𝟐. 𝟓 𝟎 𝟎 𝟏 𝟖𝟐. 𝟓
𝑻
𝜱 𝟏 𝒎 𝜱 𝟐 = 𝟏 𝟏. 𝟕𝟓 𝟐. 𝟎𝟔 𝟎 𝟖𝟐. 𝟓 𝟎 𝟎. 𝟏𝟑 = 𝟏 𝟏. 𝟕𝟓 𝟐. 𝟎𝟔 𝟏𝟎. 𝟕𝟐𝟓 ≅ 𝟎
𝟎 𝟎 𝟓𝟎 −𝟎. 𝟗𝟖 −𝟒𝟗

Assist. Prof. Dr. Muzaffer BÖREKÇİ Dynamics of Structures

EXAMPLE 2:
Solution c):
𝟖𝟐. 𝟓 𝟎 𝟎 𝟏
𝑻
𝑴𝟏 = 𝜱 𝟏 𝒎 𝜱 𝟏 = 𝟏 𝟏. 𝟕𝟓 𝟐. 𝟎𝟔 𝟎 𝟖𝟐. 𝟓 𝟎 𝟏. 𝟕𝟓 = 𝟓𝟕𝟒. 𝟑𝟒
𝟎 𝟎 𝟓𝟎 𝟐. 𝟎𝟔

𝑴𝟏 = 𝟖𝟐. 𝟓(𝟏)𝟐 +𝟖𝟐. 𝟓(𝟏. 𝟕𝟓)𝟐 +𝟓𝟎(𝟐. 𝟎𝟔)𝟐 = 𝟓𝟕𝟒. 𝟑𝟒

𝟖𝟐. 𝟓 𝟎 𝟎 𝟏
𝑻
𝑴𝟐 = 𝜱 𝟐 𝒎 𝜱 𝟐 = 𝟏 𝟎. 𝟏𝟑 −𝟎. 𝟗𝟖 𝟎 𝟖𝟐. 𝟓 𝟎 𝟎. 𝟏𝟑 = 𝟏𝟑𝟏. 𝟗𝟏
𝟎 𝟎 𝟓𝟎 −𝟎. 𝟗𝟖
𝟖𝟐. 𝟓 𝟎 𝟎 𝟏
𝑻
𝑴𝟑 = 𝜱 𝟑 𝒎 𝜱 𝟑 = 𝟏 −𝟏. 𝟓𝟑 𝟏. 𝟑𝟒 𝟎 𝟖𝟐. 𝟓 𝟎 −𝟏. 𝟓𝟑 = 𝟑𝟔𝟓. 𝟒𝟎
𝟎 𝟎 𝟓𝟎 𝟏. 𝟑𝟒

𝑴𝟏 𝟎 𝟎 𝟓𝟕𝟒. 𝟑𝟒 𝟎 𝟎
𝑴 = 𝟎 𝑴𝟐 𝟎 = 𝟎 𝟏𝟑𝟏. 𝟗𝟏 𝟎
𝟎 𝟎 𝑴𝟑 𝟎 𝟎 𝟑𝟔𝟓. 𝟒𝟎

Assist. Prof. Dr. Muzaffer BÖREKÇİ Dynamics of Structures

EXAMPLE 2:
Solution c):
𝑲𝟏 = 𝝎𝟐𝟏 𝑴𝟏 = 𝟐𝟐𝟗. 𝟏 ∗ 𝟓𝟒𝟕. 𝟑𝟒 = 𝟏𝟐𝟓𝟑𝟗𝟓. 𝟓𝟗 𝒌𝑵/𝒎
𝑲𝟐 = 𝝎𝟐𝟐 𝑴𝟐 = 𝟏𝟕𝟏𝟑 ∗ 𝟏𝟑𝟏. 𝟗𝟏 = 𝟐𝟐𝟓𝟗𝟔𝟏. 𝟖𝟑 𝒌𝑵/𝒎
𝑲𝟑 = 𝝎𝟐𝟑 𝑴𝟑 = 𝟑𝟐𝟑𝟓. 𝟒 ∗ 𝟑𝟔𝟓. 𝟒𝟎 = 𝟏𝟏𝟖𝟐𝟐𝟏𝟓. 𝟏𝟔 𝒌𝑵/𝒎

𝑲𝟏 𝟎 𝟎 𝟏𝟐𝟓𝟑𝟗𝟓. 𝟓𝟗 𝟎 𝟎
𝑲 = 𝟎 𝑲𝟐 𝟎 = 𝟎 𝟐𝟐𝟓𝟗𝟔𝟏. 𝟖𝟑 𝟎
𝟎 𝟎 𝑲𝟑 𝟎 𝟎 𝟏𝟏𝟖𝟐𝟐𝟏𝟓. 𝟏𝟔

Assist. Prof. Dr. Muzaffer BÖREKÇİ Dynamics of Structures

EXAMPLE 2:
Solution d):
First mode:
𝑺𝑫𝟏 𝟎. 𝟐𝟕𝟏
𝑻𝟏 = 𝟎. 𝟒𝟐 𝒔𝒆𝒄 𝑺𝒂𝒆 (𝑻𝟏 ) = = = 𝟎. 𝟔𝟒𝟓 (𝒈)
𝑻𝟏 𝟎. 𝟒𝟐
𝑳𝟐𝒙𝟏
𝑽𝒙𝟏 = 𝑺 𝑻 = 𝑴𝒙𝟏 𝑺𝒂𝒆 (𝑻𝟏 )
𝑴𝟏 𝒂𝒆 𝟏
𝟑
𝑳𝒙𝟏 𝒊=𝟏 𝒎𝒊 𝜱𝒊𝟏 𝟖𝟐. 𝟓 ∗ 𝟏 + 𝟖𝟐. 𝟓 ∗ 𝟏. 𝟕𝟓 + 𝟓𝟎 ∗ 𝟐. 𝟎𝟔 𝟑𝟐𝟗. 𝟖𝟕𝟓
𝜞𝒙𝟏 = = = = = 𝟎. 𝟔
𝑴𝟏 𝟑 𝟐
𝒊=𝟏 𝒎𝒊 𝜱𝒊𝟏 𝟖𝟐. 𝟓 ∗ 𝟏𝟐 + 𝟖𝟐. 𝟓 ∗ 𝟏. 𝟕𝟓𝟐 + 𝟓𝟎 ∗ 𝟐. 𝟎𝟔𝟐 𝟓𝟒𝟕. 𝟑𝟒

𝑳𝒙𝟏 𝟐 𝟑𝟐𝟗. 𝟖𝟕𝟓𝟐 𝒌𝑵𝒔𝟐

𝑴𝒙𝟏 = = = 𝟏𝟗𝟖. 𝟖𝟏 Effective participating mass of first mode
𝑴𝟏 𝟓𝟒𝟕. 𝟑𝟒 𝒎

𝑳𝟐𝒙𝟏
𝑽𝒙𝟏 = 𝑺 𝑻 = 𝑴𝒙𝟏 𝑺𝒂𝒆 𝑻𝟏 = 𝟏𝟗𝟖. 𝟖𝟏 ∗ 𝟎. 𝟔𝟒𝟓 ∗ 𝟗. 𝟖𝟏 = 𝟏𝟐𝟓𝟕. 𝟗𝟔 𝒌𝑵
𝑴𝟏 𝒂𝒆 𝟏

Assist. Prof. Dr. Muzaffer BÖREKÇİ Dynamics of Structures

EXAMPLE 2:
Solution d):
Second mode:
𝑺𝒂𝒆 (𝑻𝟐 ) = 𝑺𝑫𝑺 = 𝟎. 𝟖𝟕𝟗 (𝒈)
𝑻𝟐 = 𝟎. 𝟏𝟓 𝒔𝒆𝒄

𝑳𝟐𝒙𝟐
𝑽𝒙𝟐 = 𝑺 𝑻 = 𝑴𝒙𝟐 𝑺𝒂𝒆 (𝑻𝟐 )
𝑴𝟐 𝒂𝒆 𝟐
𝟑
𝑳𝒙𝟐 𝒊=𝟏 𝒎𝒊 𝜱𝒊𝟐 𝟖𝟐. 𝟓 ∗ 𝟏 + 𝟖𝟐. 𝟓 ∗ 𝟎. 𝟏𝟑 + 𝟓𝟎 ∗ (−𝟎. 𝟗𝟖) 𝟒𝟒. 𝟐𝟑
𝜞𝒙𝟐 = = = = = 𝟎. 𝟑𝟒
𝑴𝟐 𝟑 𝟐
𝒊=𝟏 𝒎𝒊 𝜱𝒊𝟐 𝟖𝟐. 𝟓 ∗ 𝟏𝟐 + 𝟖𝟐. 𝟓 ∗ 𝟎. 𝟏𝟑𝟐 + 𝟓𝟎 ∗ 𝟎. 𝟗𝟖𝟐 𝟏𝟑𝟏. 𝟗𝟏

𝑳𝒙𝟐 𝟐 𝟒𝟒. 𝟐𝟑𝟐 𝒌𝑵𝒔𝒆𝒄𝟐

𝑴𝒙𝟐 = = = 𝟏𝟒. 𝟖𝟑 Effective participating mass of second mode
𝑴𝟐 𝟏𝟑𝟏. 𝟗𝟏 𝒎

𝑳𝟐𝒙𝟐
𝑽𝒙𝟐 = 𝑺 𝑻 = 𝑴𝒙𝟐 𝑺𝒂𝒆 𝑻𝟐 = 𝟏𝟒. 𝟖𝟑 ∗ 𝟎. 𝟖𝟕𝟗 ∗ 𝟗. 𝟖𝟏 = 𝟏𝟐𝟕. 𝟖𝟖 𝒌𝑵
𝑴𝟐 𝒂𝒆 𝟐

Assist. Prof. Dr. Muzaffer BÖREKÇİ Dynamics of Structures

 EXAMPLE 2:
Solution d):
Third mode:
𝑺𝒂𝒆 (𝑻𝟑 ) = 𝑺𝑫𝑺 = 𝟎. 𝟖𝟕𝟗 (𝒈)
𝑻𝟑 = 𝟎. 𝟏𝟏 𝒔𝒆𝒄

𝑳𝟐𝒙𝟑
𝑽𝒙𝟑 = 𝑺 𝑻 = 𝑴𝒙𝟑 𝑺𝒂𝒆 (𝑻𝟑 )
𝑴𝟑 𝒂𝒆 𝟑
𝟑
𝑳𝒙𝟑 𝒊=𝟏 𝒎𝒊 𝜱𝒊𝟑 𝟖𝟐. 𝟓 ∗ 𝟏 + 𝟖𝟐. 𝟓 ∗ −𝟏. 𝟓𝟑 + 𝟓𝟎 ∗ 𝟏. 𝟑𝟒 𝟐𝟑. 𝟐𝟖
𝜞𝒙𝟑 = = = = = 𝟎. 𝟎𝟔𝟒
𝑴𝟑 𝟑 𝟐
𝒊=𝟏 𝒎𝒊 𝜱𝒊𝟑 𝟖𝟐. 𝟓 ∗ 𝟏𝟐 + 𝟖𝟐. 𝟓 ∗ 𝟏. 𝟓𝟑𝟐 + 𝟓𝟎 ∗ 𝟏. 𝟑𝟒𝟐 𝟑𝟔𝟓. 𝟒𝟎

𝑳𝒙𝟑 𝟐 𝟐𝟑. 𝟐𝟖𝟐 𝒌𝑵𝒔𝒆𝒄𝟐

𝑴𝒙𝟑 = = = 𝟏. 𝟒𝟖 Effective participating mass of third mode
𝑴𝟑 𝟑𝟔𝟓. 𝟒𝟎 𝒎

𝑳𝟐𝒙𝟑
𝑽𝒙𝟑 = 𝑺 𝑻 = 𝑴𝒙𝟑 𝑺𝒂𝒆 𝑻𝟑 = 𝟏. 𝟒𝟖 ∗ 𝟎. 𝟖𝟕𝟗 ∗ 𝟗. 𝟖𝟏 = 𝟏𝟐. 𝟕𝟗 𝒌𝑵
𝑴𝟑 𝒂𝒆 𝟑

Assist. Prof. Dr. Muzaffer BÖREKÇİ Dynamics of Structures

EXAMPLE 2:
Solution d):
Combination of modes:

SRSS 𝑽𝒙 = 𝑽𝟐𝒙𝟏 + 𝑽𝟐𝒙𝟐 + 𝑽𝟐𝒙𝟑 = 𝟏𝟐𝟓𝟕. 𝟗𝟔𝟐 + 𝟏𝟐𝟕. 𝟖𝟖𝟐 + 𝟏𝟐. 𝟕𝟗𝟐

𝑽𝒙,𝑺𝑹𝑺𝑺 = 𝟏𝟐𝟔𝟒. 𝟓𝟏 𝒌𝑵

Assist. Prof. Dr. Muzaffer BÖREKÇİ Dynamics of Structures

EXAMPLE 2:
Solution d):
Combination of modes:
𝟑/𝟐
𝝎𝒎 𝟖𝜻𝟐 (𝟏 + 𝜷𝒎𝒏 )𝜷𝒎𝒏
CQC 𝜷𝒎𝒏 = 𝝆𝒎𝒏 = 𝟐
𝝎𝒏 𝟏 − 𝜷𝟐𝒎𝒏 + 𝟒𝜻𝟐 𝜷𝒎𝒏 (𝟏 + 𝜷𝒎𝒏 )𝟐

βmn n=1 n=2 n=3 ρmn n=1 n=2 n=3

m=1 1 0.3657 0.2661 m=1 1 0.00798 0.0.00401
m=2 2.7345 1 0.7276 m=2 0.00798 1 0.008818
m=3 3.7580 1.3743 1 m=3 0.00401 0.08821 1

𝟑 𝟑 𝑽𝒙𝟏 𝑽𝒙𝟐 = 𝟏𝟔𝟎𝟖𝟔𝟕. 𝟗𝟐𝟒𝟖 𝑽𝒙𝟏 𝑽𝒙𝟏 = 𝟏𝟐𝟓𝟕. 𝟗𝟔𝟐

𝑽𝒙 = 𝑽𝒙𝒎 𝝆𝒎𝒏 𝑽𝒙𝒏 𝑽𝒙𝟏 𝑽𝒙𝟑 = 𝟏𝟔𝟎𝟖𝟗. 𝟑𝟎𝟖𝟒 𝑽𝒙𝟐 𝑽𝒙𝟐 = 𝟏𝟐𝟕. 𝟖𝟖𝟐
𝒎=𝟏 𝒏=𝟏 𝑽𝒙𝟐 𝑽𝒙𝟑 = 𝟏𝟔𝟑𝟓. 𝟓𝟖𝟓𝟐 𝑽𝒙𝟑 𝑽𝒙𝟑 = 𝟏𝟐. 𝟕𝟗𝟐

𝑽𝒙,𝑪𝑸𝑪 = 𝟏𝟐𝟔𝟓. 𝟔𝟗 𝒌𝑵

Assist. Prof. Dr. Muzaffer BÖREKÇİ Dynamics of Structures

EXAMPLE 2:
Solution e):
Participating mass ratio:

𝑴𝒙𝟏 𝟏𝟗𝟖. 𝟖𝟏
𝜸𝒙𝟏 = = = 𝟎. 𝟗𝟐𝟓 • 𝜸𝒙𝟏 = %𝟗𝟐. 𝟓
𝑴𝒕 𝟐𝟏𝟓

𝑴𝒙𝟐 𝟏𝟒. 𝟖𝟑
𝜸𝒙𝟐 = = = 𝟎. 𝟎𝟕 • 𝜸𝒙𝟐 = %𝟕
𝑴𝒕 𝟐𝟏𝟓

𝑴𝒙𝟑 𝟏. 𝟒𝟗
𝜸𝒙𝟑 = = = 𝟎. 𝟎𝟎𝟕 • 𝜸𝒙𝟑 = %𝟎. 𝟕
𝑴𝒕 𝟐𝟏𝟓

𝑻=𝟑

𝜸𝒙𝒏 = 𝟗𝟐. 𝟓 + 𝟕 + 𝟎. 𝟎𝟕 ≅ 𝟏𝟎𝟎

𝒏=𝟏

Assist. Prof. Dr. Muzaffer BÖREKÇİ Dynamics of Structures

MODE SUPERPOSITION SOLUTION for SYSTEMS
with
CLASSICAL DAMPING
Equations of motion for a «general viscously damped» MDOF
system
𝒎 𝒖(𝒕) + 𝒄 𝒖(𝒕) + 𝒌 𝒖(𝒕) = 𝑭(𝒕) Eq. (1)

• [c] is nxn viscous damping matrix

For the mode superposition method to be valid for viscously

damped MDOF systems, Eq(1) must be expressed in uncoupled
form (in principal coordinates)
𝑴 𝒒(𝒕) + 𝑪 𝒒(𝒕) + 𝑲 𝒒(𝒕) = 𝒑(𝒕) Eq. (2)

𝑪 = 𝜱 𝑻 𝒄 𝜱 Eq. (3)

Assist. Prof. Dr. Muzaffer BÖREKÇİ Dynamics of Structures

 𝑴 𝒒(𝒕) + 𝑪 𝒒(𝒕) + 𝑲 𝒒(𝒕) = 𝒑(𝒕) Eq. (2)

𝑪 = 𝜱 𝑻 𝒄 𝜱 Eq. (3)

[C] is an nxn symmetric matrix that is diagonal for only a special

case of [c]. This special case of damping is referred to as
«classical damping» or «proportional damping» for which [c] is
proportional to [m] and/or [k]. If it is assumed the damping matrix
[c] is constructed that
𝜱 𝑻 𝒄 𝜱 =𝟎 ; 𝒓≠𝒔 Eq. (4)
𝒓 𝒔

then [C] is diagonal and Eq(2) are uncoupled and may be

written in the form
𝟏
𝒒𝒓 𝒕 + 𝟐𝜻𝒓 𝝎𝒓 𝒒𝒓 𝒕 + 𝝎𝟐𝒓 𝒒𝒓 𝒕 = 𝒑 (𝒕) Eq. (5)
𝑴𝒓 𝒓
ζr: is the modal damping factor for the rth mode
𝑪𝒓 𝟏
𝜻𝒓 = = 𝜱 𝑻𝒓 [𝒄] 𝜱 𝒓
𝟐𝑴𝒓 𝝎𝒓 𝟐𝑴𝒓 𝝎𝒓

Assist. Prof. Dr. Muzaffer BÖREKÇİ Dynamics of Structures

Eq. (2) is given in principal ccordinates and damping forces
become uncoupled in the same manner as the inertia and
elastic forces are uncoupled. Therefore, the vibration mode
shapes for the damped system are the same as the mode
shapes for the undamped system. However, the principal
coordinate transformation is valid only for classical or
proportional damping. A very popular form of proportional
dampig is known as «Rahleigh Damping».

Rayleigh proposed that a damping matrix having the form:

𝒄 =𝜶 𝒎 +𝜷 𝒌

α and β are arbitrary proportionality factors will satisfy the

orthogonality condition.

Assist. Prof. Dr. Muzaffer BÖREKÇİ Dynamics of Structures

Damping matrix proportional to the mass and/or stiffness
matrices will permit uncoupling of the equation of motion.
𝒌 − 𝝎𝟐𝒓 𝒎 𝜱 𝒓 = 𝟎

The orthogonality conditions specify

𝜱 𝑻 𝒎 𝜱 = 𝑴𝒓
𝒓 𝒓

𝜱 𝑻 𝒌 𝜱 = 𝑲𝒓 = 𝝎𝟐𝒓 𝑴𝒓
𝒓 𝒓

Then for the Rayleigh damping

𝑪𝒓 = 𝜱 𝑻 𝒄 𝜱 = 𝜶 + 𝜷𝝎𝟐𝒓 𝑴𝒓
𝒓 𝒓

𝑪𝒓 = 𝜱 𝑻 𝒄 𝜱 = 𝟐𝑴𝒓 𝝎𝒓 𝜻𝒓
𝒓 𝒓

𝟏 𝜶
𝜻𝒓 = + 𝜷𝝎𝒓 modal damping factor Eq. (6)
𝟐 𝝎𝒓

Assist. Prof. Dr. Muzaffer BÖREKÇİ Dynamics of Structures

 𝑪𝒓 = 𝜱 𝑻 𝒄 𝜱 = 𝟐𝑴𝒓 𝝎𝒓 𝜻𝒓
𝒓 𝒓

𝟏 𝜶
𝜻𝒓 = + 𝜷𝝎𝒓 modal damping factor
𝟐 𝝎𝒓

Modal damping matrix [C] is determined as

𝑪 = 𝜱 𝑻 𝒄 𝜱

𝜻𝟏 𝝎𝟏 𝑴𝟏 𝟎 𝟎 …
𝑪 =𝟐 𝟎 𝜻𝟐 𝝎𝟐 𝑴𝟐 𝟎 …
… = 𝒅𝒊𝒂𝒈(𝟐𝜻𝒓 𝝎𝒓 𝑴𝒓 )
𝟎 𝟎 𝜻 𝝎 𝑴
… … … 𝟑 𝟑 𝟑…

Assist. Prof. Dr. Muzaffer BÖREKÇİ Dynamics of Structures

Rayleigh damping is defined for a MDOF system by specifying ζr for
two different and unequal frequencies of vibration, and then solving for
α and β by simultaneous solution of Eq. (6).
α[m] (mass proportional) renders a contribution to ζr that is inversely
proportional to ωr. However β[k] (stiffness proportional) term yields a
contribution to ζr that increases linearly with ωr. Therefore, for large
values of ωr, the stiffness proportional term dominates the system
damping. This trend generally leads to unrealistically high damping
ratios in the higher vibration modes for large MDOF system.
Fortunately, in most superposition analyses of large MDOF systems,
since the considered number of vibration modes «p» is significantly
less than the total «n» mode of vibration (p<<n), damping in the higher
modes is generally not a critical issue. Moreover, to incorporate
damping in a mode superposition analysis, it is not necessary to
formulate the damping matrix [c] explicitly, since the system response
can be determined from the solution of the set of independent
equations. However, explicit formulation of the system damping matrix
is required to evaluate the dynamic response of a MDOF system by
direct numerical integration.
Assist. Prof. Dr. Muzaffer BÖREKÇİ Dynamics of Structures
 𝟏 𝜶
𝜻𝒓 = + 𝜷𝝎𝒓 modal damping factor
𝟐 𝝎𝒓

If both modes are assumed to have the same damping ratio ζ, which is
reasonale based on experimental data, then;
𝟐𝝎𝒓 𝝎𝒔 𝟐
𝜶=𝜻 𝜷=𝜻
𝝎𝒓 + 𝝎𝒔 𝝎𝒓 + 𝝎𝒔

𝜻𝒓 𝒄 = 𝜶𝒎 𝜻𝒓 𝟏 𝜶
𝜻𝒓 = + 𝜷𝝎𝒓
𝜶 𝟐 𝝎𝒓
𝜻𝒓 = 𝒄 = 𝜷𝒌
𝟐𝝎𝒓 𝑹𝒂𝒚𝒍𝒆𝒊𝒈𝒉 𝑫𝒂𝒎𝒑𝒊𝒏𝒈
𝜷𝝎𝒓
𝜻𝒓 = 𝜻
𝟐

𝑴𝒂𝒔𝒔 𝒑𝒓𝒐𝒑𝒐𝒓𝒕𝒊𝒐𝒏𝒂𝒍

𝝎𝒊 𝝎𝒊
𝝎𝒓 𝝎𝒔
Mass proportional damping and
stiffness proportional damping Rayleigh damping

Assist. Prof. Dr. Muzaffer BÖREKÇİ Dynamics of Structures

In applying this procedure to a practical problem, the modes «r»
and «s» with specified damping ratios should be chosen to
ensure reasonable values for the damping ratios in all the
modes contributing significantly to the response.
For example, consider the five modes in the response analysis
and in the estimation of α and β, reasonable modes should be
chosen. These reasonable two modes will be chosen of which ζ
is same. This ζ should be specified for the first mode and
possibly for the fourth mode. Damping ratio for the second and
third modes will be smaller than ζ and for the fifth mode it will be
somewhat larger than ζ.
The damping ratio for modes higher than the fifth will increase
monotonically with frequency and the corresponding modal
responses will be essential eliminated because of their high
damping.

Assist. Prof. Dr. Muzaffer BÖREKÇİ Dynamics of Structures