Ch.

4
Multi-Degree of Freedom
Systems

Dr. Ahmad Alshyyab

 We have learned vibration of linear systems of single degree of
freedom and methods of solution for almost all possible external
excitations forms. In this chapter we will introduce linear systems of
multi-degree of freedom (MDOF) and its methods of solution.

Definition:
Degree of Freedom (DOF) :
• it is the number of independent coordinates needed to describe the
motion of the system components.
• it is also the number of equations required to describe the motion
of the system components.
Example 1:
Example 2:
Example 3:
Moving carriage and pendulum Example 4:
Double pendulum
Example 5:
Equations of Motion: Double pendulum
Steps of writing the equations of motion are same as those for single degree of
freedom systems
1) Free Body Diagram 𝑂1𝑦 𝑂2 𝑦
𝑂1 𝑥 𝑂2 𝑥
𝐽𝑜 𝜃ሷ1 𝐽𝑜 𝜃ሷ 2
𝜃1 𝜃2

𝑚𝑟 𝑔
𝑚𝑟 𝑔
2) Apply Newton’s law:
𝜃1 𝑎𝑛𝑑 𝜃2 are small angles
∑𝑀𝑜1 = 0:
𝜃1
−𝒌𝑳𝟐 (𝜽𝟐 − 𝜽𝟏 ) + 𝑚𝑟 𝑔𝐿 𝑠𝑖𝑛𝜃1 + 𝐽𝑜 𝜃ሷ1 = 0, Linearize,
−𝒌𝑳𝟐 (𝜽𝟐 − 𝜽𝟏 ) + 𝑚𝑟 𝑔𝐿𝜃1 + 𝐽𝑜 𝜃ሷ1 = 0, (1a)
∑𝑀𝑜2 = 0:
𝜃2
𝒌𝑳𝟐 (𝜽𝟐 − 𝜽𝟏 ) + 𝑚𝑟 𝑔𝐿 𝑠𝑖𝑛𝜃2 + 𝐽𝑜 𝜃ሷ2 = 0, Linearize,
𝒌𝑳𝟐 (𝜽𝟐 − 𝜽𝟏 ) + 𝑚𝑟 𝑔𝐿 𝜃2 + 𝐽𝑜 𝜃ሷ2 = 0, (1b)
Rearrange
𝐽𝑜 𝜃ሷ1 + 𝑚𝑟 𝑔𝐿 + 𝒌𝑳𝟐 𝜃1 − 𝒌𝑳𝟐 𝜽𝟐 = 0, (2a)

𝐽𝑜 𝜃ሷ2 − 𝒌𝑳𝟐 𝜽𝟏 + 𝑚𝑟 𝑔𝐿 + 𝒌𝑳𝟐 𝜃2 = 0, (2b)

𝐽𝑜 0 𝜃ሷ1 𝑚𝑟 𝑔𝐿 + 𝒌𝑳𝟐 −𝒌𝑳𝟐 𝜽𝟏

+ = 0, (2c)
0 𝐽𝑜 ሷ𝜃2 −𝒌𝑳𝟐 𝑚𝑟 𝑔𝐿 + 𝒌𝑳𝟐 𝜽𝟐
ถ
𝑚𝑎𝑠𝑠 𝑚𝑎𝑡𝑟𝑖𝑥 [𝐌], 𝑺𝒕𝒊𝒇𝒇𝒏𝒆𝒔𝒔 𝒎𝒂𝒕𝒓𝒊𝒙 [𝐊], 𝐷𝑖𝑠𝑝𝑙𝑎𝑐𝑒𝑚𝑒𝑛𝑡
𝑠𝑦𝑚𝑚𝑒𝑡𝑟𝑖𝑐 𝒔𝒚𝒎𝒎𝒆𝒕𝒓𝒊𝒄 𝑣𝑒𝑐𝑡𝑜𝑟

In the simplified matrix form

[𝐌] 𝚯ሷ + 𝐊 𝜣 = 𝟎
 3) Find the transient solution for 𝚯
.

We will follow the same steps we have learned for SDOF. Assume the solution
𝒖𝟏
𝚯 = 𝐮 𝒆𝒋𝝀𝒕 , ሶ 𝒋𝝀𝒕 ሷ 𝟐 𝒋𝝀𝒕
𝚯 = 𝐮 𝒋𝝀𝒆 , 𝚯 = − 𝐮 𝝀 𝒆 , 𝐮 = 𝒖 , (𝟒𝒂)
𝟐

And substitute in the equations of motion (Eq. 3)

− 𝐌 𝐮 𝝀𝟐 𝒆𝒋𝝀𝒕 + 𝐊 𝐮 𝒆𝒋𝝀𝒕 = 𝟎, (𝟒𝒃)

Rearrange,
[ 𝐊 − 𝐌 𝝀𝟐 ] 𝐮 𝒆𝒋𝝀𝒕 = 𝟎, (𝟒𝒄)
𝑒 𝑗𝜆𝑡 is never zero. Therefore

[ 𝐊 − 𝐌 𝝀𝟐 ]] 𝐮 = 𝟎, (𝟒𝒅)
 −1
If the inverse 𝐊 − 𝜆2 𝐌 exist, then , 𝐮 = 𝐾 − 𝑀 𝜆2 −1 0 = 0,
which is a trivial solution. One last possibility to get a solution for 𝐮 is to
−1
assume the inverse 𝐊 − 𝐌 𝜆2does not exist, which means that the
determinant of 𝐾 − 𝑀 𝜆2 must be zero

2
𝐊 −𝜆 𝐌 =0, (4c)

Substitute for 𝐊 and 𝐌 and find the determinant,

 𝑚𝑟 𝑔𝐿 + 𝒌𝑳𝟐 −𝒌𝑳𝟐 2 𝐽𝑜 0
+ −𝜆 = 0, (4f)
−𝒌𝑳𝟐 𝑚𝑟 𝑔𝐿 + 𝒌𝑳𝟐 0 𝐽𝑜

𝑚𝑟 𝑔𝐿 + 𝒌𝑳𝟐 − 𝜆2 𝑱𝒐 −𝒌𝑳𝟐
= 0, (4g)
−𝒌𝑳𝟐 𝟐 2
𝑚𝑟 𝑔𝐿 + 𝒌𝑳 − 𝜆 𝑱𝒐

𝑚𝑟 𝑔𝐿 + 𝒌𝑳𝟐 − 𝜆2 𝑱𝒐 −𝒌𝑳𝟐
=0
−𝒌𝑳𝟐 𝑚𝑟 𝑔𝐿 + 𝒌𝑳𝟐 − 𝜆2 𝑱𝒐

2
(𝑚𝑟 𝑔𝐿 + 𝒌𝑳𝟐 − 𝜆2 𝑱𝒐 )2 − 𝒌𝑳 𝟐 = 0, (4h)

The last equation is called the characteristic equation, its roots are called
the eigenvalues. Rearrange Eq.(4h) and find the eigenvalues (the roots).
 (𝑚𝑟 𝑔𝐿 + 𝒌𝑳𝟐 − 𝜆2 𝑱𝒐 ) = ±𝒌𝑳𝟐 , (4i)
The eigenvalues are:
𝒌𝑳𝟐 𝑚𝑟 𝑔𝐿+𝒌𝑳𝟐 𝑚 𝑔𝐿 𝟐𝒌𝑳 𝟐 𝑚𝑟 𝑔𝐿
2 𝑟
𝜆 = + ± (5) 2,
𝜆1 = 2
, 𝜆2 = + , (6)
𝑱𝒐 𝑱𝒐 𝑱𝒐 𝑱𝒐 𝑱𝒐
The square roots of the eigenvalues are
𝑚𝑟 𝑔𝐿 𝟐𝒌𝑳𝟐 𝑚𝑟 𝑔𝐿 (7a,b)
𝜆1 = ± 𝜆2 = ± +
𝑱𝒐 𝑱𝒐 𝑱𝒐
Arrange the eigenvalues such that +𝝀𝟏 <+𝝀𝟐 ,
𝝎𝒏𝟏 = 𝝀𝟏 :is called the first natural frequency, it is the smallest one
𝝎𝒏𝟐 = 𝝀𝟐 :is called the 2ed natural frequency, it is the next smallest one, and so one if you have more 𝝀’s
−1
We have found the values of 𝜆 which render the inverse 𝐊 − 𝜆 𝐌 not exist, 2

so what is 𝐮 , to find the answer substitute the values of 𝜆12 into eq. (4d),
distinguish the solution 𝐮 corresponds 𝜆12 by subscript 1 to be 𝐮 𝟏

𝟐 𝒖𝟏𝟏
𝐊 − 𝐌𝝀 ]𝐮 𝟏 = 𝟎, 𝐮 𝟏= 𝒖 , (𝟒𝒅)
𝟐𝟏
 𝒎𝒓 𝒈𝑳
a) 𝝀𝟐𝟏 =
𝑱𝒐
𝒎𝒓 𝒈𝑳
𝒎𝒓 𝒈𝑳 + 𝒌𝑳𝟐 −( )𝑱𝒐 −𝒌𝑳𝟐 𝒖𝟏𝟏 𝒖𝟏𝟏
𝑱𝒐
𝒎 𝒈𝑳 𝒖𝟐𝟏 = 𝟎, 𝐮 𝟏 = 𝒖
𝟐
−𝒌𝑳 𝒎𝒓 𝒈𝑳 + 𝒌𝑳𝟐 − ( 𝒓 )𝑱𝒐 𝟐𝟏
𝑱𝒐

(𝒎𝒓 𝒈𝑳 + 𝒌𝑳𝟐 − 𝒎𝒓 𝒈𝑳) 𝒖𝟏𝟏 − 𝒌𝑳𝟐 𝒖𝟐𝟏 =0 𝒖𝟏𝟏 = 𝒖𝟐𝟏 (8a)

−𝒌𝑳𝟐 𝒖𝟏𝟏 + (𝒎𝒓 𝒈𝑳 + 𝒌𝑳𝟐 − 𝒎𝒓 𝒈𝑳)𝒖𝟐𝟏 =0 𝒖𝟏𝟏 = 𝒖𝟐𝟏 (8b)

The last two equations are just one equation, we can know the value of
one element of the vector 𝐮 . if we assume the another on, thus
A mode shape is a vector 𝒖𝟏𝟏 𝟏
that describes the relative 𝐮 𝟏 = 𝒖𝟏𝟏 = 𝒖𝟏𝟏
motion between the two ด
𝟏 (9)
masses or between two
degrees of freedom
𝟏𝒔𝒕 𝒎𝒐𝒅𝒆 𝒏𝒐𝒓𝒎𝒂𝒍𝒊𝒛𝒆𝒅
𝒔𝒉𝒂𝒑𝒆 𝟏𝒔𝒕 𝒎𝒐𝒅𝒆
𝒔𝒉𝒂𝒑𝒆
 𝟐𝒌𝑳𝟐 𝒎𝒓 𝒈𝑳
b) 𝝀𝟐𝟐 = +
𝑱𝒐 𝑱𝒐
to find the other values of 𝐮 corresponds the values of 𝝀𝟐𝟐 , substitute
𝝀𝟐𝟐 into eq. (4d), distinguish the solution 𝐟𝐨𝐫 𝐮 by subscript 2 to be 𝐮 𝟐
𝟐𝒌𝑳𝟐 𝒎𝒓 𝒈𝑳
𝒎𝒓 𝒈𝑳 + 𝒌𝑳𝟐 − ( + )𝑱𝒐 −𝒌𝑳𝟐 𝒖𝟏𝟐 𝒖𝟏𝟐
𝑱𝒐 𝑱𝒐
𝟐 𝟐 𝟐𝒌𝑳𝟐 𝒎𝒓 𝒈𝑳 𝒖𝟐𝟐 = 𝟎, 𝐮 𝟐 = 𝒖
𝟐𝟐
−𝒌𝑳 𝒎𝒓 𝒈𝑳 + 𝒌𝑳 − ( + )𝑱𝒐
𝑱𝒐 𝑱𝒐

−𝒌𝑳𝟐 𝒖𝟏𝟐 − 𝒌𝑳𝟐 𝒖𝟐𝟐 =0 𝒖𝟏𝟐 = −𝒖𝟏𝟐 (10a)

−𝒌𝑳𝟐 𝒖𝟏𝟐 − 𝒌𝑳𝟐 𝒖𝟐𝟐 =0 𝒖𝟏𝟐 = −𝒖𝟐𝟐 (10b)

𝒖𝟏𝟐 𝟏
𝐮 𝟐 = −𝒖 = 𝒖𝟏𝟐 (10c)
𝟏𝟐 −𝟏
𝟐𝒔𝒕 𝒎𝒐𝒅𝒆 𝒏𝒐𝒓𝒎𝒂𝒍𝒊𝒛𝒆𝒅
𝒔𝒉𝒂𝒑𝒆 𝟐𝒔𝒕 𝒎𝒐𝒅𝒆
𝒔𝒉𝒂𝒑𝒆
Hence, the solution we have assumed in eq.(4a) is
𝚯 = 𝒂𝟏 𝐮 𝟏 𝒆𝒋𝝎𝒏𝟏 𝒕 + 𝒃𝟏 𝐮 𝟏 𝒆−𝒋𝝎𝒏𝟏 𝒕 + 𝒂𝟐 𝐮 𝟐 𝒆𝒋𝝎𝒏 𝟐 𝒕 + 𝒃𝟐 𝐮 𝟐 𝒆−𝒋𝝎𝒏 𝟐 𝒕

𝟏 𝒋𝝎𝒏𝟏 𝒕 𝟏 −𝒋𝝎𝒏𝟏 𝒕 𝟏 𝒋𝝎𝒏 𝟐 𝒕 𝟏

𝚯 = 𝒂𝟏 𝒖𝟏𝟏 𝒆 + 𝒃𝟏 𝒖𝟏𝟏 𝒆 + 𝒂𝟐 𝒖𝟏𝟐 𝒆 + 𝒃𝟐 𝒖𝟏𝟐 𝒆−𝒋𝝎𝒏 𝟐 𝒕
𝟏 𝟏 −𝟏 −𝟏
𝑨𝟏 𝑩𝟏 𝑨𝟐 𝑩𝟐

𝟏 𝒋𝝎𝒏𝟏 𝒕 𝟏 −𝒋𝝎𝒏𝟏 𝒕 𝟏 𝒋𝝎𝒏 𝟐 𝒕 𝟏

𝚯 = 𝑨𝟏 𝒆 + 𝑩𝟏 𝒆 + 𝑨𝟐 𝒆 + 𝑩𝟐 𝒆−𝒋𝝎𝒏 𝟐 𝒕
𝟏 𝟏 −𝟏 −𝟏
𝟏 𝟏
𝚯 = 𝑨𝟏 +𝑩𝟏 𝒄𝒐𝒔𝝎𝒏𝟏 𝒕 + 𝒋 𝑨𝟏 −𝑩𝟏 𝒔𝒊𝒏𝝎𝒏𝟏 𝒕
𝟏 𝟏
𝟏 𝟏
+ 𝑨𝟐 +𝑩𝟐 𝒄𝒐𝒔𝝎𝒏𝟐 𝒕 + +𝐣 𝑨𝟐 −𝑩𝟐 𝒔𝒊𝒏𝝎𝒏𝟐 𝒕
−𝟏 −𝟏
The solution is, 𝟏 𝟏
𝚯 = 𝑨 𝒄𝒐𝒔𝝎𝒏𝟏 𝒕 + 𝑩 𝒔𝒊𝒏𝝎𝒏𝟏 𝒕
𝟏 𝟏
𝟏 𝟏
+𝑪 𝒄𝒐𝒔𝝎𝒏𝟐 𝒕 + +𝑫 𝒔𝒊𝒏𝝎𝒏𝟐 𝒕
−𝟏 −𝟏
The solution is,
1 1
𝚯 =𝑨 𝒄𝒐𝒔𝜔𝑛1 𝒕 + 𝑩 𝒔𝒊𝒏𝜔𝑛1 𝒕
1 1
1 1
+𝑪 𝒄𝒐𝒔𝜔𝑛2 𝒕 + 𝐷 𝒔𝒊𝒏𝜔𝑛2 𝒕
−1 −1

Where, the constants 𝑨, 𝑩,𝑪, 𝐚𝐧𝐝 𝐷 are to be determined

from the initial conditions. Hence, the number of needed
initial conditions are four. Furthermore, there are two
natural frequencies for the system 𝜔𝑛1 , and 𝜔𝑛2
Given the following pendulum parameters, plot 𝜃1 𝑡 and 𝜃2 (𝑡)
𝑙 1 2 N
𝑙 = 1 m; 𝐿 = 2 ; 𝑚 = 1 kg; Jo = 𝑚𝑙 𝑘𝑔. 𝑚2 ; 𝑘 = 10 ;
3 m

9.81 𝑚 𝑙
ω1 = = 3.84 𝑟𝑎𝑑/𝑠
2Jo

9.81𝑚𝑙 𝑘𝐿2
ω2 = +2 = 5.45 rad/s
2Jo Jo
 Modes of Motion
a) First mode motion:

𝟏 𝟏
𝜣 =𝑨 𝒄𝒐𝒔𝝎𝒏𝟏 𝒕 + 𝑩 𝒔𝒊𝒏𝝎𝒏𝟏 𝒕
𝟏 𝟏

1 ሶ 0 0 1
{Θ(0)} = ; {Θ(0)} = {Θ(0)} = ሶ
; {Θ(0)} =
1 0 0 1
cos(3.836𝑡) 0.26sin(3.836 𝑡)
𝜣(𝒕) = 𝜣(𝒕) =
cos(3.836𝑡) 0.26sin(3.836 𝑡)
𝜃1 (𝑡)
𝜃1 (𝑡) 𝜃2 (𝑡)
𝜃2 (𝑡)
b) Second mode motion:

1 1 1(t )
2
𝚯 =𝑪 𝒄𝒐𝒔𝜔𝑛2 𝒕 + 𝐷 𝒔𝒊𝒏𝜔𝑛2 𝒕 1 = − 2
−1 −1
Fig. 4b, Second mode motion.

0.183 Sin[5.45 𝑡]
𝜃 𝑡
−0.183 4Sin[5.45 𝑡]
With linear combination of both modes you get the complete solution of eq.(11)

1 1 1 1
𝚯 =𝑨 𝒄𝒐𝒔𝜔𝑛1 𝒕 + 𝑩 𝒔𝒊𝒏𝜔𝑛1 𝒕 + 𝑪 𝒄𝒐𝒔𝜔𝑛2 𝒕 + D 𝒔𝒊𝒏𝜔𝑛2 𝒕
1 1 −1 −1

1 −.2
𝚯(0) = , 𝚯′ (0) = ,
0.5 0.7
0.75Cos[3.83𝑡] + 0.25Cos[5.45𝑡] + 0.0652Sin[3.83𝑡] − 0.082Sin[5.45𝑡]
𝛩𝑡 =
0.75Cos[3.83𝑡] − 0.25Cos[5.45𝑡] + 0.0652Sin[3.83𝑡] + 0.082Sin[5.45𝑡]
In Eqs. 9 and 11, we have normalized the mode shapes by dividing by its largest absolute
element value. However, this is not the only way to normalize a vector (mode shape), another
mean is to divide the vector by its norm (magnitude), and the resulting normalized vector is
a unit vector. The norm of any vector x is x , where ,

n
x = x .x = 
T
x i .,
2 (13a)
i =1

𝐿
Example:
Find the norm x of the vector x and its normalized vector

0.5 
x =  1  , (13b)
0.7 
 
0.5
 
x .x = 0.5 1 0.7  1  = 0.52 + 12 + 0.72 = 1.74
T
0.7  , (13c)
 
x = 1.74 = 1.32

0.5 0.37 
1 1    
xnormalized = x =   
1 = 0.75 , , (13d)
x 1.32    
  
0.7 0.53 
Example 3: Moving carriage and pendulum
𝑥𝐵 = 𝑥𝐴 + 𝑥𝐵/𝐴 = 𝑥 + 𝐿𝜃, 𝑥𝐴 = 𝑥
𝑚 𝑐 𝑥ሶ 𝑘𝑥
𝑣𝐵 = 𝑥ሶ + 𝐿𝜃ሶ 𝑚𝑥ሷ
𝜃
𝜃 +𝐿𝜃ሶ 2
𝑎𝐵 = 𝑥ሷ +𝐿 𝜃ሷ
𝐵 𝑚𝑟 𝑥ሷ
𝒎𝒓 𝑳𝜽ሶ 𝟐 𝜃
𝐿𝜃ሶ 2
𝑚𝑟
𝐿 𝜃ሷ 𝑚𝑟 𝐿𝜃ሶ 2
𝒎𝒓 𝑳 𝜽ሷ 𝑚𝑔 𝜃
𝑥ሷ
𝒎𝒓 𝒙ሷ
𝜃
𝑚𝑔

∑𝐹𝑥 = 0: 𝑚 𝑥ሷ 𝑡 + 𝑐𝑥ሶ 𝑡 + 𝑘𝑥 𝑡 + 𝑚𝑟 𝑥ሷ 𝑡 + 𝑚𝑟 𝐿 𝜃ሷ cos 𝜃 − 𝑚𝑟 𝐿𝜃ሶ 2 sin 𝜃 = 0, (1)

∑𝑀𝐴 = 0: 𝐽𝐵 𝜃ሷ + 𝑚𝑟 𝑥ሷ 𝑡 𝑐𝑜𝑠𝜃 𝐿 + 𝑚𝑟 𝐿 𝜃ሷ 𝐿 + 𝑚𝑔 𝑠𝑖𝑛𝜃 𝐿 = 0, (2)

 Example 3: Moving carriage and pendulum
2 2
𝑣𝐵 = 𝑣𝐴 + 𝐿𝜃,ሶ 𝑣𝐵2 = 𝑥ሶ + 𝐿𝜃ሶ cos𝜃 + 𝐿𝜃ሶ sin𝜃
1 1 1
𝑇 = 𝑚𝑟 𝑣𝐵2 + 𝐽𝐵 𝜃ሶ 2 + 𝑚𝑣𝐴2 𝐿𝜃ሶ
2 2 2
1 2 2 1 1 𝜃
= 𝑚𝑟 𝑥ሶ + 𝐿𝜃ሶ cos𝜃 + 𝐿𝜃ሶ sin𝜃 ሶ
+ 𝐽𝐵 𝜃 + 𝑚𝑥ሶ 2
2
𝑥ሶ
2 2 2
1 2
𝑈 = 𝑚𝑟 𝑔 1 − cos 𝜃 𝐿 + 𝑘𝑥
2
𝜃
1 2 2 1 1 1
𝐿 = 𝑇 − 𝑈 = 𝑚𝑟 [ 𝑥ሶ + 𝐿𝜃ሶ cos𝜃 + 𝐿𝜃ሶ sin𝜃 ] + 𝐽𝐵 𝜃ሶ 2 + 𝑚𝑥ሶ 2 − 𝑚𝑟 𝑔 1 − cos 𝜃 𝐿 − 𝑘𝑥 2
2 2 2 2
𝜕𝐿 𝑑 𝜕𝐿 𝜕𝐿
= 𝑚𝑟 𝑥ሶ + 𝐿𝜃ሶ cos𝜃 + 𝑚 𝑥;ሶ = 𝑚𝑟 𝑥ሷ + 𝐿𝜃ሷ cos𝜃 − 𝐿𝜃ሶ 2 sin𝜃 + 𝑚 𝑥ሷ = −𝑘𝑥
𝜕𝑥ሶ 𝑑𝑡 𝜕𝑥ሶ 𝜕𝑥

𝑑 𝜕𝐿 𝜕𝐿
− =0
𝑑𝑡 𝜕𝑥ሶ 𝜕𝑥

𝑚 𝑥ሷ + 𝑘𝑥 + 𝑚𝑟 𝑥ሷ + 𝑚𝑟 𝐿𝜃ሷ cos𝜃 − 𝑚𝑟 𝐿𝜃ሶ 2 sin𝜃 += 0, (1)

𝜕𝐿 𝑑 𝜕𝐿
= 𝑚𝑟 𝑥ሶ 𝐿 cos𝜃 + 𝐿2 𝜃ሶ 𝒄𝒐𝒔𝟐 𝜽 + 𝐿2 𝜃ሶ 𝒔𝒊𝒏𝟐 𝜽 + 𝐽𝐵 𝜃;ሶ = 𝑚𝑟 (𝑥ሷ 𝐿 cos𝜃 − 𝐿 𝑥ሶ 𝜃ሶ sin𝜃 + 𝑳𝟐 𝜽)
ሷ + 𝐽𝐵 𝜃;ሷ
𝜕𝜃ሶ 𝑑𝑡 𝜕𝜃ሶ
𝜕𝐿
ሶ ሶ ሶ ሶ
𝑑 𝜕𝐿
= 𝑚𝑟 (𝑥ሷ 𝐿 cos𝜃 − 𝐿 𝑥ሶ 𝜃ሶ sin𝜃 + 𝑳𝟐 𝜽)
ሷ + 𝐽𝐵 𝜃;ሷ
𝑑𝑡 𝜕𝜃ሶ
𝜕𝐿
= 𝑚𝑟 ( 𝑥ሶ + 𝐿 𝜃ሶ cos𝜃 )(−𝐿𝜃ሶ sin 𝜃 ) + 𝑚𝑟 𝐿𝜃ሶ sin𝜃 𝐿𝜃ሶ cos𝜃 − 𝑚𝑟 𝑔𝐿 sin𝜃
𝜕𝜃

𝑑 𝜕𝐿 𝜕𝐿
− =0
𝑑𝑡 𝜕𝜃ሶ 𝜕𝜃

⇒ 𝑚𝑟 𝑥ሷ 𝐿 cos𝜃 − 𝐿 𝑥ሶ 𝜃ሶ sin𝜃 + 𝑳𝟐 𝜽ሷ + 𝐽𝐵 𝜃ሷ − 𝒎𝒓 𝒙ሶ + 𝑳 𝜽ሶ 𝒄𝒐𝒔𝜽 −𝑳𝜽ሶ 𝐬𝐢𝐧 𝜽 − 𝒎𝒓 𝑳𝜽ሶ 𝒔𝒊𝒏𝜽 𝑳𝜽ሶ 𝐜𝐨𝐬𝜽 + 𝒎𝒓 𝒈𝑳 𝒔𝒊𝒏𝜽 = 𝟎

⇒ 𝑚𝑟 𝑥ሷ 𝐿 cos𝜃 − 𝐿 𝑥ሶ 𝜃ሶ sin𝜃 + 𝑳𝟐 𝜽ሷ + 𝐽𝐵 𝜃ሷ − 𝑚𝑟 𝑥ሶ + 𝐿 𝜃ሶ cos𝜃 −𝐿𝜃ሶ sin 𝜃 − 𝑚𝑟 𝐿𝜃ሶ sin𝜃 𝐿𝜃ሶ cos𝜃 + 𝑚𝑟 𝑔𝐿 sin𝜃 = 0

⇒ 𝑚𝑟 𝑥ሷ 𝐿 cos𝜃 + 𝑳𝟐 𝜽ሷ + 𝐽𝐵 𝜃ሷ + 𝑚𝑟 𝑔𝐿 sin𝜃 = 0

Linearize: sin𝜃 ≈ 𝜃, cos𝜃 ≈ 1

⇒ 𝐽𝐵 𝜃ሷ + 𝑚𝑟 𝑥ሷ 𝐿 + 𝒎𝒓 𝑳𝟐 𝜽ሷ + 𝑚𝑟 𝑔𝐿 𝜃 = 0, (2)
 𝒌𝟏 (𝒙 + 𝒍𝟏 𝜽)
Example 2:

𝒎𝒆𝜽𝒄𝒐𝒔𝜽
𝒙

ሷ
𝒍𝟏 𝒍𝟐

𝒌𝟐 (𝒙 − 𝒍𝟐 𝜽)
𝒆

𝒄𝟏 (𝒙ሶ + 𝒍𝟏 𝜽)
𝒎𝒆𝜽ሷ

ሶ
𝜽
ሷ
𝒎(𝒆𝜽)

𝒄𝟐 (𝒙ሶ − 𝒍𝟐 𝜽)
ሷ
𝒎𝒆𝜽𝒔𝒊𝒏𝜽

ሶ
𝑱𝒄𝒈 𝜽ሷ
𝑪𝑮
𝑶
𝑪𝑮 𝜽
𝒎𝒙ሷ

𝑪𝑮 𝒎𝒆𝜽ሶ 𝟐 𝒄𝒐𝒔𝜽

𝒎𝒙ሷ
𝜽
𝒙ሷ

𝒎𝒆𝜽ሶ 𝟐 𝒔𝒊𝒏𝜽
𝒂𝑪𝑮 = 𝒂𝒐 + 𝒂𝑪𝑮/𝑶 = 𝒙ሷ + 𝒆𝜽ሷ + 𝒆𝜽ሶ 𝟐 𝜽

−𝒎𝒂𝒄𝑮 = −𝒎𝒂𝒐 + −𝒎𝒂𝑪𝑮/𝑶 = 𝒎𝒙ሷ + 𝒎𝒆𝜽ሷ + 𝒎𝜽ሶ 𝟐 𝒂𝑪𝑮

ሷ ≈0
∑𝑭𝒙 = −𝒎𝒙ሷ + 𝒎𝒆 𝜽𝒄𝒐𝒔𝜽 − 𝒎𝒆𝜽ሶ 𝟐 𝒔𝒊𝒏𝜽 − 𝒌𝟏 𝒙 + 𝒍𝟏 𝜽 − 𝒌𝟐 𝒙 − 𝒍𝟐 𝜽 − 𝑐1 𝑥ሶ + 𝑙1 𝜃ሶ − 𝑐2 𝑥ሶ − 𝑙2 𝜃ሶ = 𝟎

𝒎 𝒙ሷ − 𝒆 𝜽ሷ + 𝒌𝟏 + 𝒌𝟐 𝒙 + 𝒌𝟏 𝒍𝟏 − 𝒌𝟐 𝒍𝟐 𝜽 + 𝑐1 + 𝑐2 𝑥ሶ + 𝑐1 𝑙1 − 𝑐2 𝑙2 𝜃ሶ = 𝟎, (𝟏)
 𝒌𝟏 (𝒙 + 𝒍𝟏 𝜽)
Example 2:
𝒙
𝒍𝟏 𝒍𝟐

𝒌𝟐 (𝒙 − 𝒍𝟐 𝜽)
𝒆

𝒎 𝒆𝜽ሷ 𝒄𝒐𝒔𝜽
𝒄𝟏 (𝒙ሶ + 𝒍𝟏 𝜽)
ሶ
ሷ
𝒎(𝒆𝜽)

𝒄𝟐 (𝒙ሶ − 𝒍𝟐 𝜽)
ሶ
𝑱𝒄𝒈 𝜽ሷ
𝑶
ሷ
𝒎(𝒆𝜽)
𝑪𝑮 𝜽 𝜽
𝒎 𝒆𝜽ሷ 𝒔𝒊𝒏𝜽 𝒙ሷ

𝒎(𝒙)
𝜽 𝒎𝒆𝜽ሶ 𝒄𝒐𝒔𝜽
𝟐
𝜽

𝒎𝒆𝜽ሶ 𝟐 𝒔𝒊𝒏𝜽
𝒂𝑪𝑮

𝒎𝒙ሷ
∑𝑴𝒐 = 𝑱𝑪𝑮 𝜽ሷ − 𝒎𝒆 𝒙𝒄𝒐𝒔𝜽
ሷ + 𝒎 𝒆𝟐 𝜽ሷ + 𝒌𝟏 𝒍𝟏 𝒙 + 𝒍𝟏 𝜽 − 𝒌𝟐 𝒍𝟐 𝒙 − 𝒍𝟐 𝜽 + 𝑙1 𝑐1 𝑥ሶ + 𝑙1 𝜃ሶ − 𝑙2 𝑐2 𝑥ሶ − 𝑙2 𝜃ሶ = 𝟎

(𝑱𝑪𝑮 + 𝒎 𝒆𝟐 )𝜽ሷ − 𝒎𝒆 𝒙ሷ + 𝒌𝟏 𝒍𝟏 − 𝒌𝟐 𝒍𝟐 𝒙 + 𝒌𝟏 𝒍𝟐𝟏 + 𝒌𝟐 𝒍𝟐𝟐 𝜽 + 𝑙1 𝑐1 − 𝑙2 𝑐2 𝑥ሶ + 𝑐1 𝑙12 + 𝑐2 𝑙22 𝜃ሶ = 𝟎, (𝟐)

 .. .
𝒎 −𝒎𝒆 𝒙 𝒄𝟏 + 𝒄 𝟐 𝒄𝟏 𝒍 𝟏 − 𝒄𝟐 𝒍 𝟐 𝒙 𝒌𝟏 + 𝒌𝟐 𝒌𝟏 𝒍𝟏 − 𝒌𝟐 𝒍𝟐 𝒙
𝟐 { .. } + {.}+ { }=𝟎
−𝒎𝒆 𝑱 + 𝒎𝒆 𝜽 𝒄𝟏 𝒍 𝟏 − 𝒄 𝟐 𝒍 𝟐 𝒄𝟏 𝒍 𝟏 𝟐 + 𝒄𝟐 𝒍 𝟐 𝟐 𝜽 𝒌𝟏 𝒍𝟏 − 𝒌𝟐 𝒍𝟐 𝒌 𝟏 + 𝒌𝟐 𝒍 𝟐 𝟐 𝜽
𝑫𝒂𝒎𝒑𝒊𝒏𝒈 𝒎𝒂𝒕𝒓𝒊𝒙 [𝑪]
𝑰𝒏𝒆𝒓𝒕𝒊𝒂 𝒎𝒂𝒕𝒓𝒊𝒙 [𝑴], 𝑺𝒕𝒊𝒇𝒇𝒏𝒆𝒔𝒔 𝑴𝒂𝒕𝒓𝒊𝒙 [𝑲],
𝑴 𝒊𝒔 𝒏𝒐𝒏 𝒅𝒊𝒂𝒈𝒐𝒏𝒂𝒍 𝒎𝒂𝒕𝒓𝒊𝒙 𝑲 𝒊𝒔 𝒏𝒐𝒏 𝒅𝒊𝒂𝒈𝒐𝒏𝒂𝒍 𝒎𝒂𝒕𝒓𝒊𝒙
𝒅𝒚𝒏𝒂𝒎𝒊𝒄 𝒄𝒐𝒖𝒑𝒍𝒊𝒏𝒈 𝑺𝒕𝒂𝒕𝒊𝒄 𝒄𝒐𝒖𝒑𝒍𝒊𝒏𝒈
Dynamic and static coupling:
•If the inertia matrix is non-diagonal, we said the system has dynamic coupling,
•If the stiffness matrix is non-diagonal, we said the system has static coupling,
•If the inertia matrix and the stiffness one are non-diagonal, we said the system
has static and dynamic coupling,
The last equations have both static and dynamic coupling. If we chose the
coordinate x , such that 𝒆 = 𝟎 , the equations of motions will be only couple
statically.
.. .
𝒎 𝟎 𝒙 𝒄 𝟏 + 𝒄𝟐 𝒄𝟏 𝒍 𝟏 − 𝒄𝟐 𝒍 𝟐 𝒙 𝒌𝟏 + 𝒌𝟐 𝒌𝟏 𝒍𝟏 − 𝒌𝟐 𝒍𝟐 𝒙
{ .. } + {.}+ { }=𝟎
𝟎 𝑱 𝜽 𝒄𝟏 𝒍 𝟏 − 𝒄𝟐 𝒍 𝟐 𝒄𝟏 𝒍 𝟏 𝟐 − 𝒄𝟐 𝒍 𝟐 𝟐 𝜽 𝒌𝟏 𝒍 𝟏 − 𝒌𝟐 𝒍 𝟐 𝒌𝟏 𝒍𝟏 𝟐 − 𝒌𝟐 𝒍𝟐 𝟐 𝜽
𝑰𝒏𝒆𝒓𝒕𝒊𝒂 𝒎𝒂𝒕𝒓𝒊𝒙 [𝑴], 𝑫𝒂𝒎𝒑𝒊𝒏𝒈 𝒎𝒂𝒕𝒓𝒊𝒙 [𝑪] 𝑺𝒕𝒊𝒇𝒇𝒏𝒆𝒔𝒔 𝑴𝒂𝒕𝒓𝒊𝒙 [𝑲],
𝑴 𝒊𝒔 𝒅𝒊𝒂𝒈𝒐𝒏𝒂𝒍 𝒎𝒂𝒕𝒓𝒊𝒙 𝑲 𝒊𝒔 𝒏𝒐𝒏 𝒅𝒊𝒂𝒈𝒐𝒏𝒂𝒍 𝒎𝒂𝒕𝒓𝒊𝒙
𝑵𝒐 𝒅𝒚𝒏𝒂𝒎𝒊𝒄 𝒄𝒐𝒖𝒑𝒍𝒊𝒏𝒈 𝑺𝒕𝒂𝒕𝒊𝒄 𝒄𝒐𝒖𝒑𝒍𝒊𝒏𝒈
,Instead, if we chose the coordinates such that, k1l1 = k 2l 2 , the equations of motions are only coupled
dynamically.

.. .
𝑚 −𝑚𝑒 𝑥 𝑐1 + 𝑐2 𝑐1 𝑙1 − 𝑐2 𝑙2 𝑥 𝑘1 + 𝑘2 0 𝑥
𝐽 + 𝑚𝑒 2 { .. } + {.}+ { }=0
−𝑚𝑒 𝜃 𝑐1 𝑙1 − 𝑐2 𝑙2 𝑐1 𝑙1 2 − 𝑐2 𝑙2 2 𝜃 0 𝑘1 𝑙1 2 − 𝑘2 𝑙2 2 𝜃
𝐼𝑛𝑒𝑟𝑡𝑖𝑎 𝑚𝑎𝑡𝑟𝑖𝑥 [𝑀],
𝐷𝑎𝑚𝑝𝑖𝑛𝑔 𝑚𝑎𝑡𝑟𝑖𝑥 [𝐶 ] 𝑆𝑡𝑖𝑓𝑓𝑛𝑒𝑠𝑠 𝑀𝑎𝑡𝑟𝑖𝑥 [𝐾],
𝑀 𝑖𝑠 𝑛𝑜𝑛 𝑑𝑖𝑎𝑔𝑜𝑛𝑎𝑙 𝑚𝑎𝑡𝑟𝑖𝑥 𝐾 𝑖𝑠 𝑑𝑖𝑎𝑔𝑜𝑛𝑎𝑙 𝑚𝑎𝑡𝑟𝑖𝑥
𝐷𝑦𝑛𝑎𝑚𝑖𝑐 𝑐𝑜𝑢𝑝𝑙𝑖𝑛𝑔 𝑁𝑜 𝑠𝑡𝑎𝑡𝑖𝑐 𝑐𝑜𝑢𝑝𝑙𝑖𝑛𝑔
Find the equations of motion for the last problem using Lagrange method, do not include the damping

1 2 1 1
𝑅 = 𝑐𝑣 ሶ 2 ሶ 2
𝑅 = 𝑐2 (𝑥ሶ − 𝑙2 𝜃) + 𝑐1 (𝑥ሶ + 𝑙1 𝜃)
2 2 2
Damped system:
1 1
ሶ 2 + 𝑐1 (𝑥ሶ + 𝑙1 𝜃)
𝑅 = 𝑐2 (𝑥ሶ − 𝑙2 𝜃) ሶ 2
2 2

𝜕𝑅 𝜕𝑅
= 𝑐2 𝑥ሶ − 𝑙2 𝜃ሶ + 𝑐1 𝑥ሶ + 𝑙1 𝜃ሶ = −𝑙2 𝑐2 𝑥ሶ − 𝑙2 𝜃ሶ + 𝑙1 𝑐1 𝑥ሶ + 𝑙1 𝜃ሶ
𝜕𝑥ሶ 𝜕𝜃ሶ

𝒅 𝝏𝑳 𝝏𝑳 𝝏𝑹
− + =𝟎
𝒅𝒕 𝝏𝒙 𝝏𝒙 𝝏𝒙ሶ

𝒎 𝒙ሷ − 𝒆 𝜽ሷ + 𝒌𝟏 𝒙 + 𝒍𝟏 𝜽 + 𝒌𝟐 𝒙 − 𝒍𝟐 𝜽 + 𝑐2 𝑥ሶ − 𝑒𝜃ሶ + 𝑐1 𝑥ሶ + 𝑒𝜃ሶ = 𝟎, (𝟏)

𝑱 𝜽ሷ − 𝒎𝒆 𝒙ሷ + 𝒎 𝒆𝟐 𝜽ሷ + 𝒌𝟏 𝒍𝟏 𝒙 + 𝒍𝟏 𝜽 − 𝒌𝟐 𝒍𝟐 𝒙 − 𝒍𝟐 𝜽 + 𝑙1 𝑐1 𝑥ሶ + 𝑙1 𝜃ሶ − 𝑙2 𝑐2 𝑥ሶ − 𝑙2 𝜃ሶ = 𝟎, (𝟐)
 Forced Excitations
The governing equations of motion with forced
excitations is
𝑀 𝜃ሷ 𝑡 + 𝐾 𝜃 𝑡 = 𝐹 𝑡 , (18𝑎)

𝐹 𝑡 could be any excitation force. For harmonic excitation 𝐹 𝑡 𝐵 cos(𝜔𝑡)

𝐹 𝑡 = 𝐹 𝑅𝑒(𝑒 𝑗𝜔𝑡 ), (18𝑏)

The solution pattern follows the excitation pattern. Assume the solution
is also harmonic of the form

𝜃 𝑡 = 𝑉 𝑒 𝑗𝜔𝑡 , 𝜃ሷ 𝑡 = −𝜔2 𝑉 𝑒 𝑗𝜔𝑡 , (18𝑐)

𝑉 : is the displacement magnitude vector. Substitute eq. (16b,c) into eq. (16a)
−𝜔2 𝑀 𝑉 𝑒 𝑗𝜔𝑡 + 𝐾 V 𝑒 𝑗𝜔𝑡 = F 𝑅𝑒(𝑒 𝑗𝜔𝑡 ),

[ 𝐾 − 𝜔2 𝑀 ] . V 𝑒 𝑗𝜔𝑡 = {𝐹}𝑒 𝑗𝜔𝑡 ⇒ [ 𝐾 − 𝜔2 𝑀 ] . V = F , ( 18d)

𝑉 =[ 𝐾 − 𝜔2 𝑀 ]−1 {F}, ( 18e)

Θ(𝑡) = [ 𝐾 − 𝜔2 𝑀 ]−1 {F}𝑒 𝑗𝜔𝑡 , ( 18f)

Now take the real part of the solution.

Θ(𝑡) = [ 𝐾 − 𝜔2 𝑀 ]−1 {F}cos(𝜔𝑡), ( 19)
Recall the double pendulum equations of motion, with forcing terms (Troques)
.. .
𝐽𝑜 0 𝜃1 𝑚𝑔𝐿 + kL2 −𝑘𝐿2 𝜃1 M1
.. + . = cos(𝜔 𝑡), (2c)
0 𝐽𝑜 𝜃2 −𝑘𝐿2 𝑚𝑔𝐿 + kL 2
𝜃2 𝑀2

Δ = −𝑘 2 𝐿4 + kL2 + 𝑔𝐿𝑚 − 𝜔2 𝐽𝑜 2
𝑀𝑚1 cos(𝜔𝑡) 𝑀2 cos(𝜔𝑡)

𝒌𝑳𝟐 + 𝒎𝒈𝑳 − 𝝎𝟐 𝑱𝒐 𝒌𝑳𝟐

−𝟏 𝚫 𝚫
𝑲 − 𝝎𝟐 𝑴 =
𝒌𝑳𝟐 𝒌𝑳 + 𝒎𝒈𝑳 − 𝝎𝟐 𝑱𝒐
𝟐

𝜟 𝜟)
𝒉𝒐𝒎𝒐𝒈𝒆𝒏𝒐𝒖𝒔 𝒔𝒐𝒍𝒖𝒕𝒊𝒐𝒏
𝟏 𝟏
𝚯(𝒕) = 𝑨 𝐂𝐨𝐬 𝛚𝐧𝟏 𝒕 + 𝑩 𝐒𝐢𝐧𝛚𝐧𝟏 𝒕 + 𝑪 𝐂𝐨𝐬 𝛚𝐧𝟐 𝒕 + 𝑫 𝐒𝐢𝐧 𝛚𝐧𝟐 𝒕
𝟏 −𝟏
𝜣 𝒕 𝒑 : 𝒑𝒂𝒓𝒕𝒊𝒄𝒖𝒍𝒂𝒓 𝒔𝒐𝒍𝒖𝒕𝒊𝒐𝒏

𝒌𝑳𝟐 + 𝒎𝒈𝑳 − 𝝎𝟐 𝑱𝒐 𝒌𝑳𝟐

𝚫 𝚫 𝑭𝟏
+ 𝒄𝒐𝒔 𝝎𝒕
𝒌𝑳𝟐 𝒌𝑳𝟐 + 𝒎𝒈𝑳 − 𝝎𝟐 𝑱𝒐 𝑭𝟐
𝜟 𝜟)
−𝟏
𝑲−𝝎𝟐 𝑴
Given, the system parameters are,
𝒍 𝟏 𝑵
𝒍 = 𝟏 𝒎 ; 𝒈 = 𝟗. 𝟖𝟏 𝒎/𝒔𝟐 ; 𝑳 = ; 𝒎 = 𝟏 𝒌𝒈; 𝐉𝐨 = 𝒎𝒍𝟐 𝒌𝒈. 𝒎𝟐 ; 𝒌 = 𝟏𝟎 ;
𝟐 𝟑 𝒎

𝒈𝒎𝒍 𝒓𝒂𝒅 𝒈𝒎𝒍 𝒌𝑳𝟐

𝛚𝒏𝟏 = = 𝟑. 𝟖𝟑𝟔 ; 𝛚𝒏𝟐 = +𝟐 = 𝟓. 𝟒𝟓 𝒓𝒂𝒅/𝒔
𝟐𝐉𝐨 𝒔 𝟐𝐉𝐨 𝐉𝐨

The forcing term is,

𝑭𝟏
𝑭 𝒕 = 𝐅 𝒄𝒐𝒔 𝝎 𝒕 = 𝒄𝒐𝒔 𝝎 𝒕 , 𝑭𝟏 = 𝟏; 𝑭𝟐 = 𝟎;
𝑭𝟐
Plot the solution for
a) The initial conditions are, 𝐱𝟎 = 𝟏, 𝟏 ; 𝐯𝟎 = 𝟎, 𝟎 ; 𝝎=2, 3.7, 3.836, 5.3 , 5.45 and 8.0 rad/s
b) The initial conditions are, 𝐱𝟎 = 𝟎, 𝟎 ; 𝐯𝟎 = 𝟏, 𝟏 ; 𝝎=8.0 rad/s
c) The initial conditions are, 𝐱𝟎 = −𝟏, 𝟏 ; 𝐯𝟎 = 𝟎, 𝟎 ; 𝝎=8.0 rad/s
d) The initial conditions are, 𝐱𝟎 = 𝟎, 𝟎 ; 𝐯𝟎 = −𝟏, 𝟏 ; 𝝎=8.0 rad/s

d) The initial conditions are, 𝐱𝟎 = 𝟎. 𝟕, 𝟎. 𝟏 ; 𝐯𝟎 = −𝟎. 𝟖, 𝟏. 𝟏 ; 𝝎=8.0 rad/s

a) The initial conditions are, 𝐱 𝟎 = 𝟏, 𝟎. 𝟓 ; 𝐯𝟎 = 𝟎, 𝟎 ;

𝟏
𝐌 = , 𝝎 =2 rad/s
𝟏. 𝟓
𝜃2 𝑡 = 0.38cos(2. 𝑡) + 0.4cos(3.836𝑡) + 0.279cos(5.45𝑡)

𝑥𝑝 (𝑡) 𝑥ℎ (𝑡)
𝜃1 𝑡 = 0.32cos(2. 𝑡) + 0.4cos(3.836𝑡) + 0.279cos(5.45𝑡)
𝑥𝑝 (𝑡) 𝑥ℎ (𝑡)
𝜃2 𝑡 = 0.38cos(2. 𝑡) + 0.4cos(3.836𝑡) + 0.279cos(5.45𝑡)
 a) The initial conditions are, 𝐱 𝟎 = 𝟏, 𝟎. 𝟓 ; 𝐯𝟎 = 𝟎, 𝟎 ;

𝟏 𝝎 → 𝝎n1 = 3.836; Beating

𝐅 = , 𝝎 =3.7 rad/s
𝟏. 𝟓

𝜃1 𝑡 = 3.611 cos 3.7𝑡 − 2.9cos(3.836𝑡) + 0.296cos(5.45𝑡) 𝜃2 𝑡 = 3.69cos 3.7𝑡 − 2.9cos 3.836𝑡 − 0.297cos(5.45𝑡)
 a) The initial conditions are, 𝐱 𝟎 = 𝟏, 𝟎. 𝟓 ; 𝐯𝟎 = 𝟎, 𝟎 ;
𝟏
𝐅 = , 𝝎 =3.836 rad/s 𝝎 = 𝝎n1 = 3.836 ⇒ Resonance: the solution grows without limit.
𝟏. 𝟓

𝜃1 𝑡 = 0.7cos[3. . 83𝑡] + 0.3 cos[5.45𝑡] + 0.48𝟗 𝒕 sin[3.836𝑡]

𝜃2 𝑡 = 0.8 𝑐𝑜𝑠[3.83𝑡] − 0.3 cos[5.45𝑡] + 0.489 𝑡 sin[3.836𝑡]
a) The initial conditions are, 𝐱 𝟎 = 𝟏, 𝟎. 𝟓 ; 𝐯𝟎 = 𝟎, 𝟎 ; 𝝎=2, 3.7, 3.836, 5.3 , 5.45 and 8.0 rad/s
𝟏
𝐅 = , 𝝎 =5.3 rad/s 𝝎 → 𝝎n2 = 5.45 rad/s ⇒ Beating
𝟏. 𝟓

𝜃1 𝑡 = 1.03Cos[3.836𝑡] − 0.742Cos[5.3𝑡] + 0.711Cos[5.451𝑡]

𝜃2 𝑡 = 1.03Cos[3.836𝑡] + 1.092Cos[5.4𝑡] − 1.6Cos[5.45𝑡]
a) The initial conditions are, 𝐱 𝟎 = 𝟏, 𝟎. 𝟓 ; 𝐯𝟎 = 𝟎, 𝟎 ; 𝝎=2, 3.7, 3.836, 5.3 , 5.45 and 8.0 rad/s
𝟏
𝐅 = , 𝝎 =5.45 rad/s 𝝎 = 𝝎n2 = 5.45 rad/s ⇒⇒ Resonance: the solution grows without limit.
𝟏. 𝟓

𝜃1 𝑡 𝜃2 𝑡
a) The initial conditions are, 𝐱 𝟎 = 𝟏, 𝟎. 𝟓 ; 𝐯𝟎 = 𝟎, 𝟎 ; 𝝎=2, 3.7, 3.836, 5.3 , 5.45 and 8.0 rad/s
𝟏
𝐅 = , 𝝎 =8 rad/s
𝟏. 𝟓

𝜃1 𝑡 = 0.826 Cos[3.836 t]+0.228 Cos[5.451 t]−0.0542 cos[8. t]

𝜃2 𝑡 = 0.826 Cos[3.836 t]−0.228 Cos[5.451 t]−0.098 cos[8. t]

b) Particular solution due to the forcing terms:

𝐱 𝟎 = 𝟎, 𝟎 ; 𝐯𝟎 = 𝟎, 𝟎 ;

𝒌𝑳𝟐 + 𝒎𝒈𝑳 − 𝝎𝟐 𝑱𝒐 𝒌𝑳𝟐

𝚫 𝚫 𝑴𝟏
𝚯 𝒕 𝒑𝒂𝒓𝒕𝒊𝒄𝒖𝒍𝒂𝒓 = 𝒄𝒐𝒔 𝝎𝒕
𝒌𝑳𝟐 𝒌𝑳𝟐 + 𝒎𝒈𝑳 − 𝝎𝟐 𝑱𝒐 𝑴𝟐
𝜟 𝜟)
−𝟏
𝑲−𝝎𝟐 𝑴

Therefore;
𝒌𝑳𝟐 + 𝒎𝒈𝑳 − 𝝎𝟐 𝑱𝒐 𝒌𝑳𝟐
𝜽𝟏 𝒕 = M1 + 𝐅𝟐 𝒄𝒐𝒔 𝝎𝒕
𝚫 𝚫

𝒌𝑳𝟐 𝒌𝑳𝟐 + 𝒎𝒈𝑳 − 𝝎𝟐 𝑱𝒐

𝜽𝟐 𝒕 = M1 + 𝐅𝟐 𝒄𝒐𝒔 𝝎𝒕
𝚫 𝚫
b) Particular solution due to the forcing terms:

𝒍
If 𝒍 = 𝟏; 𝒈 = 𝟗. 𝟖𝟏; 𝑳 = ; 𝑭 = F1 , F2 , ; 𝐱 𝟎 = 𝟎, 𝟎 ; 𝐯𝟎 = 𝟎, 𝟎 ;
𝟐
𝟏 𝟐
𝒎 = 𝟏; 𝐉𝐨 = 𝒎𝒍 ; 𝒌 = 𝟏𝟎;
𝟑
𝒈𝒎𝒍 𝒓𝒂𝒅 𝒈𝒎𝒍 𝒌𝑳𝟐
𝛚𝒏𝟏 = = 𝟑. 𝟖𝟑𝟔 ; 𝛚𝒏𝟐 = +𝟐 = 𝟓. 𝟒𝟓 𝒓𝒂𝒅/𝒔
𝟐𝐉𝐨 𝒔 𝟐𝐉𝐨 𝐉𝐨

2𝜔2
2 5M2 + M1 14.81 − 3

2𝜔 2 2𝜔 2
9.81 − 3 19.81 − 3
𝚯t 𝑝𝑎𝑟𝑡𝑖𝑐𝑢𝑙𝑎𝑟 = 𝑐𝑜𝑠 ω𝑡
2𝜔2
2 5M1 + M2 14.81 −
3
2𝜔 2 2𝜔 2
9.81 − 3 19.81 − 3
b) Particular solution due to the forcing terms:
𝒍 𝟏
If 𝒍 = 𝟏; 𝒈 = 𝟗. 𝟖𝟏; 𝑳 = ; 𝒎 = 𝟏; 𝐉𝐨 = 𝒎𝒍𝟐 ; 𝒌 = 𝟏𝟎; 𝐱 𝟎 = 𝟎, 𝟎 ; 𝐯𝟎 = 𝟎, 𝟎 ;
𝟐 𝟑

𝒈𝒎𝒍 𝒓𝒂𝒅 𝒈𝒎𝒍 𝒌𝑳𝟐

𝛚𝒏𝟏 = = 𝟑. 𝟖𝟑𝟔 ; 𝛚𝒏𝟐 = +𝟐 = 𝟓. 𝟒𝟓 𝒓𝒂𝒅/𝒔
𝟐𝐉𝐨 𝒔 𝟐𝐉𝐨 𝐉𝐨
2𝜔2
2 −9.81 + 3

2𝜔 2 2𝜔 2
𝛩t 𝑝𝑎𝑟𝑡𝑖𝑐𝑢𝑙𝑎𝑟 = 9.81 − 3 19.81 − 3 𝑐𝑜𝑠 ω𝑡
2
2𝜔 2
19.81 − 3

−1
−1 {𝐌} = ;
{𝐌} = ; 1
1
Θ2

𝜔
|Θ2 | Θ1
|Θ1 |

𝜔
b) Particular solution due to the forcing terms:
𝒍 𝟏
If 𝒍 = 𝟏; 𝒈 = 𝟗. 𝟖𝟏; 𝑳 = ; 𝒎 = 𝟏; 𝐉𝐨 = 𝒎𝒍𝟐 ; 𝒌 = 𝟏𝟎; 𝐱 𝟎 = 𝟎, 𝟎 ; 𝐯𝟎 = 𝟎, 𝟎 ;
𝟐 𝟑

𝒈𝒎𝒍 𝒓𝒂𝒅 𝒈𝒎𝒍 𝒌𝑳𝟐

𝛚𝒏𝟏 = = 𝟑. 𝟖𝟑𝟔 ; 𝛚𝒏𝟐 = +𝟐 = 𝟓. 𝟒𝟓 𝒓𝒂𝒅/𝒔
𝟐𝐉𝐨 𝒔 𝟐𝐉𝐨 𝐉𝐨
2
2𝜔 2
9.81 − 3
𝛩t 𝑝𝑎𝑟𝑡𝑖𝑐𝑢𝑙𝑎𝑟 = 𝑐𝑜𝑠 ω𝑡
2
1
2𝜔 2 {𝐌} = ;
9.81 − 3 1

1
{𝐌} = ; |Θ2 |
Θ1 Θ2 1
ω |Θ1 |
𝜔
b) Particular solution due to the forcing terms:
𝒍 𝟏
If 𝒍 = 𝟏; 𝒈 = 𝟗. 𝟖𝟏; 𝑳 = ; 𝒎 = 𝟏; 𝐉𝐨 = 𝒎𝒍𝟐 ; 𝒌 = 𝟏𝟎; 𝐱 𝟎 = 𝟎, 𝟎 ; 𝐯𝟎 = 𝟎, 𝟎 ;
𝟐 𝟑

𝒈𝒎𝒍 𝒓𝒂𝒅 𝒈𝒎𝒍 𝒌𝑳𝟐

𝛚𝒏𝟏 = = 𝟑. 𝟖𝟑𝟔 ; 𝛚𝒏𝟐 = +𝟐 = 𝟓. 𝟒𝟓 𝒓𝒂𝒅/𝒔
𝟐𝐉𝐨 𝒔 𝟐𝐉𝐨 𝐉𝐨

1
1 {𝐌} = ;
{𝐌} = ; 0
0

Θ1
Θ2 𝜔
𝜔𝑛1 𝜔𝑛2
|Θ1 |
|Θ2 | 𝜔

𝜔𝑛1 𝜔𝑛2
 Example: (important)
Vibration absorber is an addition of spring-mass to a single degree of freedom system as
shown to minimize the motion of the original mass (m)

F(t)=F0sin(ωt) F(t)=F0cos(ωt)
x x

m m

k a xa
k/2 k/2 k/2 k/2
ma
Absorber added
m m
Fig. (a) A single degree Fig. (b) A two degrees of freedom
of freedom
The new system is two degree of freedom system with two natural frequencies.
The values of the absorber mass and stiffness are chosen such that the motion of
the original mass (m) is minimum (zero). For the absorber shown
1) Find the natural frequencies
2) chose the values of absorber stiffness 𝑘𝑎 and mass 𝑚𝑎 and , respectively, such
that the motion of m is zero.
3) find the magnitude of the force acts on the absorber mass 𝑚𝑎

Eq. of motion
mx + kx + k a (x − x a ) = F0 cos(t ) ,, (21a)
ma x a − k a (x − x a ) = 0 ,, (21b)

m 0   x  k + k a −k a   x  F0 cos(t )  F0  j t )

   +     = =
   Re(e , (21c)
0 m a  x a   −k a k a  x a   0  0
Solution:
1) To find the natural frequencies you need only the un-damped free system

m 0   x  k + k a −k a   x 
   +     = 0 , (22a)
0 ma  x a   −k a k a  x a 

a) find the characteristic equation

k + k a −  2m −k a
| [K ] −  2[M ] |= 0; = 0, (22b)
−k a k a −  2ma

(k + k a −  2m )(k a −  2m a ) − k a2 = 0,
(kk a −  2km a ) + (k a −  2m )(k a −  2m a ) − k a2 = 0 , (22c)
kk a −  2km a + k a2 −  2k a m a −  2k a m +  4m a2 = 0
Rearrange,
ma  − (kma + k a m a + k a m )  + (k a + kk a ) = 0 ,
2 4 2 2

A B C
this is the characteristic equation. The eigenvalues are,
2
−B  B − 4AC
2
1,2 = ,
2A
The natural frequencies are
1 = +1; 2 = +2 , (22d)
Find the steady state solution
Since the excitation force is F0 cos(t ) = F0 .Re(e j t ) . Assume the solution with imaginary
exponent; once you get the solution take its real part only.
 x   X  j t  X  j t
{x} =   =   e 2
;[x ] = −   e (23a)
x a  X a  X a 

m 0   X  j t k + k a −k a   X  j t F0  j t
− 2
   e +   e =  e ,, (23b)
 0 ma  X a   −k a k a  X a  0

k + k a −  2m −k a 
  X  j t F0  j t
   e = e ,, (23c)

 −k a k a −  2ma 
  Xa  0
−1
X  k + k −  2m −k a  F0 
a
 =    ., (23d)
X a   −k a k a −  2ma  0
X  1 
−  2m
 a
k a ka  F0  2m )(k + k −  2m ) − k 2 (23e)
  =    ,  = ( k a −  a a a

 a  
X  2
k + ka − m  0
 ka

(k a −  2ma )F0 k a F0
X = ;X a = (23f)
(k a −  2ma )(k + k a −  2m ) − k a2 (k a −  2ma )(k + k a −  2m ) − k a2

The steady state motion is

 (k a −  2m a )F0 
 
   (k −  m a )(k + k a −  m ) − k a 
x 2 2 2
j t ) , (24a)
{x} =   =  a  Re(e
x a   k a F0 
 2 2 2
 (k a −  m a )(k + k a −  m ) − k a 
 (k a −  2ma )F0 
 
 x   (k −  ma )(k + k a −  m ) − k a 
2 2 2
{x} =   =  a  cos(t ) (24b)
 a 
x k a F0 
 2 2 2
 (k a −  ma )(k + k a −  m ) − k a 
To minimize the motion (x(t)) of the mass (m), you need to minimize
its magnitude X by selecting proper values of k a , and ma this is
accomplished by equating X to zero
(k a −  2ma )F0 k
X = = 0; →  2 = a ., (24c)
(k a −  2ma )(k + k a −  2m ) − k a2 ma

Hence, if the absorber parameters k a , and ma are chosen such

that  2 m a = k a , the motion of the mass (m) will be zero x(t)=0 .
Doing so, the magnitude of the motion of the absorber mass ( ma )
F
is Xa = 0 . (24d)
−k a
Why this choice of k a , and ma ? You can answer the last
question once you know”What is the magnitude of the force
acts on the mass m ?”.
=0
F0
Fa (t ) = k a (x (t ) − x a (t )) = −k a X a cos(t ) = − k a cos(t ) = F0 cos(t ) ,
− ka
Fa (t ) = F0 cos(t ) ,
A very nice conclusion is that the summation of
forces on mass m is zero. With this proper choice of
k a , and ma (  2 m a = k a ).
F (t ) = F0 cos(t )

Fa (t ) = F0 cos(t )

ma
Problem 4.1. Consider the system of the Figure. For 𝑐1 = 𝑐2 = 𝑐3 = 0, derive the equation of motion and
calculate the mass and stiffness matrices. Note that setting 𝑘3 = 0 in your solution should result in the stiffness
matrix given by equation (4.9).

For 𝑐1 = 𝑐2 = 𝑐3 = 0 and 𝑘3 = 0
Find The modeshapes,

81 − 32.45 81 + 32.45
2
1 = = 2.7 , 2
2 = = 6.3, (26e)
18 18
The system natural frequencies are 1 = 12 = 2.7 = 1.64 Rad / s and

2 = 22 = 6.3 = 2.51 Rad / s , (26g)

Problem 4.3. Calculate the mode shapes 𝐮 1 and 𝐮 2 for Problem 4.2.

First mode shape: 𝐊 − 2.7 𝐌 𝐮 1 = 0, (27a)

27 − 9 2.7 −3 𝑢11
{𝑢 } = 0, (27b)
−3 6 − 2.7 21
 2.7𝑢11 − 3𝑢21 = 0 u 1

−3𝑢11 + 3.3𝑢21 = 0, therefore 𝑢11 = 1.11𝑢21 𝑚2 𝑢21

𝑚1 𝑢11
𝑢11 1.11
u 1 = {𝑢 } = 𝑢21 { }, (27c) 1.0 1.1
21 1

Second mode shape: 𝐊 − 6.3 𝐌 𝐮 1 =0

27 − 9 6.3 −3 𝑢11
{ } = 0,
−3 6 − 6.3 𝑢21 u 1
−29.7𝑢11 − 3.0𝑢21 = 0 𝑚2 𝑢21
3
−3𝑢11 − 0.3𝑢21 = 0 , therefore 𝑢11 = − 29.7 𝑢21 (27d) 𝑚1 𝑢11

−0.101 1.0
𝑢12 −0.11
u 2 = {𝑢 } = 𝑢22 { }, (27e)
22 1
Problem 4.4. For initial conditions 𝐱 0 = 𝟏 𝟎 𝑻 and 𝐱ሶ 0 = 𝟎 𝟎 𝑻 , calculate the
free response of the system of Problem 4.2. Plot the response x1(t) and x2(t).
Note the solution contained both modes of motions, that is it has both natural frequencies. The
choice of the initial conditions, determine either both modes of motion are excited or only one of
them is excited.
  x 1(0) 
For all cases of initial conditions, if the velocity initial condition is zero,   = 0 , you can
x 2 (0) 
excite one mode of motion only if the displacement initial condition is equal that mode shape
or its multiples (scaled), for example to excite the last system with the one mode chose the
initial conditions as
Problem 4.6
For k1 = k 3 = 0 , the system is not restrained and is allowed to translate as a one unit as a
rigid body (as if k=infinity) may be without any oscillations or as rigid body but with
oscillations, so you should anticipate one of the natural frequencies to be zero (rigid body
motion), lets see x1(t ) x (t ) 2

m1 0   x 1   k 2 −k 2   x 1  k2
   +    = 0 (31a)
 0 m 2  x 2  −k 2 k 2  x 2  m m

Problem 4.7
Given k 2 = 10 N / m . Find the characteristic equation, natural frequencies and mode shapes
| [K ] −  2 [M ] |= 0
10 − 9 2 −10
= 0, (31b) x1(t )
−10 10 −  2 const.
x2 (t ) = x1 (t )
(10 − 9 2 )(10 −  2 ) − 100 = 0 , 100 − 100 2 + 9 4 − 100 = 0, k2
m
(9 2 − 100) 2 = 0 ,is the characteristic equation. m
The eigenvalues are
12 = 0 rad / s , 22 = 100 / 9 rad / s
Rigid body mode, ω=0
x (t)=x (t)
 𝑚 𝑢21

𝑚 𝑢11
1.0

𝑚 𝑢21

𝑢11 𝑚

−0.11 1.0
4.12. Determine the equation of motion in matrix form, then calculate the natural
frequencies and mode shapes of the torsional system of Figure P4.12. Assume that the
torsional stiffness values provided by the shafts are equal (𝒌𝟏 = 𝒌𝟐 , 𝑵. 𝒎/𝒓𝒂𝒅) and that
disk 1 has three times the inertia as that of disk 2 (𝐽1 = 3𝐽2 ).

3𝐽2 𝜃ሷ1 + k1 𝜃1 − k 2 𝜃2 − 𝜃1 = 0
𝐽2 𝜃ሷ2 + k 2 𝜃2 − 𝜃1 = 0

3𝐽2 0 𝜃ሷ1 𝑘1 + 𝑘2 0 𝜃1
+ =0
0 𝐽2 ሷ𝜃2 0 𝐽2 𝜃2

3𝐽2 0 𝜃ሷ1 𝑘1 + 𝑘2 0 𝜃1
+ =0
0 𝐽2 ሷ
𝜃2 0 𝐽2 𝜃2
4.2 Eigenvalues and Natural Frequencies
A matrix [M] is positive definite if the scalar formed from the product 𝐗 𝐓 . 𝐀. 𝐗 > 𝟎 for every
nonzero choice of the vector 𝐗
{𝐗}𝐓 . M . {𝐗} > 𝟎,
Cholesky decomposition:
The factorization L is called the Cholesky decomposition and is examined, along with the notion of
positive definite
A.symmetric and positive definite matrix M (inertia matrix) may be factored into two terms:
𝑀 = [𝐿]. [𝐿]T

𝐿11 0 0 ⋯ 0 𝐿11 𝐿21 𝐿31 ⋯ 𝐿𝑛1

𝐿21 𝐿22 0 0 0 𝐿22 𝐿32 0 𝐿𝑛2
𝐿 = 𝐿31 𝐿32 𝐿33 0 ⋮ , [𝐿]T = 𝐿33 0 ⋮
⋮ ⋱ ⋮ ⋮ ⋱ ⋮
𝐿𝑛1 𝐿𝑛2 ⋯ ⋯ 𝐿𝑛𝑛 ⋯ ⋯ 𝐿𝑛𝑛
Diagonal Matrix:
A diagonal matrix [M] is positive definite if its all elements >0

𝑚11 0 ⋯ 0 0
0 𝑚22 0 0 0
[M] = ⋮ 0 𝑚33 0 0
0 0 0 ⋱ 0
0 0 0 0 𝑚𝑛𝑛 1
0 ⋯ 0 0
𝑚11
𝑚11 0 ⋯ 0 0 1
0 0 0 0
1 0 𝑚22 0 0 0 −1 𝑚22
L = M = ⋮ 0 T M =
2
𝑚33 0 0 = L And 2
⋮ 0
1
0 0
(33b)
𝑚33
0 0 0 ⋱ 0
0 0 0 ⋱ 0
0 0 0 0 𝑚𝑛𝑛 1
0 0 0 0 𝑚𝑛𝑛
1 1 1
M : matrix square root and M
2 2. M 2 =M
1 −1 1 −1
The inverse of the diagonal matrix M , is M2 2 ⇒ M . M
2 2 = I
Eigenvectors:

We have determined the eigenvalues and modeshapes for [M] xሷ + 𝐊 {𝐱} = 0

Where the eigenvalues 𝜆′𝑖 𝑠 and the corresponding modeshapes 𝐮 ′𝑖 𝑠 are found from
|[M] + 𝐊 | = 0, (34b) gives the characteristic eq.
𝐊 − 𝜆𝑖 𝐌 𝐮 𝑖 = 0 (34b)

For two degrees of freedom the solution was,

𝑥 = 𝐮 1 [𝐴1 𝑐𝑜𝑠 𝜔𝑛1 𝑡 + 𝐵1 𝑠𝑖𝑛 𝜔𝑛1 𝑡 ] + 𝐮 2 [𝐴2 𝑐𝑜𝑠 𝜔𝑛2 𝑡 + 𝐵2 𝑠𝑖𝑛 𝜔𝑛2 𝑡 ]
𝜔𝑛1 = + 𝜆1 , 𝜔𝑛2 = + 𝜆2 (34c)

Or,
𝐱 = 𝐮 1 𝐴𝑐𝑜𝑠 𝜔𝑛1 𝑡 − 𝜙1 + 𝐮 𝟐 𝐵𝑐𝑜𝑠 𝜔𝑛2 𝑡 − 𝜙2 ,

𝐵1 𝐵
𝜙1 = 𝑡𝑎𝑛−1 , 𝜙2 = 𝑡𝑎𝑛−1 ( 2 ) (34d)
𝐴1 𝐴2
Coordinate Transformation or Change

Introduce the coordinate transformation

−1
𝐱 = M 2 𝐪, M : Mass matrix
Substitute into the equation of motion,
−1 −1
M 𝐱ሷ + K 𝐱 = 0 ⇒ M M −1
2 𝐪ሷ + K M 2 𝐪 = 𝟎, (35b)
Pre-multiply eq. (35b) by M 2 ,

−1 −1 −1 −1
M 2 M M 2 𝐪ሷ + M 2 K M 2 𝐪 = 𝟎, (35c)
𝑰 ෩
𝐊
ณ
Identity matrix:dynamic decoupling
Mass normalized matrix

෩ ] likes the matrix [𝐊], is a symmetric matrix. The matrix 𝐊

[𝐊 ෩ is called the mass-
normalized stiffness 𝐪ሷ + [𝑲෩ ] 𝐪 = 0, (35d)
Solution: Solve to get 𝐪 :
Assume 𝐪 = 𝐯 e𝑗 𝜆 𝑡

෩ 𝐪 = 0 ⇒ , −𝜆2 𝐯 e𝑗 𝜆 𝑡 + 𝐯 e𝑗 𝜆 𝑡 =0,
𝐪ሷ + [𝑲] (35d)

෩ − 𝜆2 [𝐼] 𝐯 e𝑗 𝜆 𝑡 =0,
[𝑲] (35d

෩ − 𝜆2 [𝐼]| = 0 ⇒ 𝜆2 = 𝜆12 , 𝜆22 ……𝜆2𝑛 . Where, 𝜆12 , < 𝜆22 ,<……< 𝜆2𝑛
|[𝑲]
Where 𝜆2𝑖 are the eigenvalues of the symmetric matrix [𝑲].
෩ The natural frequencies are, 𝜔𝑛 𝑖 = 𝜆2𝑖 , 𝑟𝑎𝑑𝑙𝑠
𝜔𝑛1 = 𝜆12 , 𝜔𝑛2 = 𝜆22 , … … , 𝜔𝑛𝑖 = 𝜆2𝑖

The corresponding eigenvectors are 𝐯 𝟏 , 𝐯 𝟐 , 𝐯 𝟑 , … . 𝐯 𝒏 ,

The eigenvectors of symmetric matrix [𝑲] ෩ are orthogonal and never zero just like the unit vectors
(𝐢, 𝐣, 𝐤, 𝐞ො 𝟏 , 𝐞ො 𝟐 , 𝐞ො 𝟑 … Orthogonal means 𝐯 𝑻 𝒊 . 𝐯 𝒋 = 𝟎, 𝒊 ≠ 𝒋

෩ is symmetric, this is called the symmetric eigenvalue problem. The eigenvector 𝐯

Since the matrix [𝑲]
generalizes the concept of a mode shape 𝐮 used in Section 4.1.
The norm (magnitude) of a vector 𝐯𝒊 is denoted by and defined by

𝐯𝒊 = 𝐯 𝑻
𝒊. 𝐯 𝒊

𝐯 𝒊
The normalized eigenvector {ത
𝐯}𝒊 is defined by {ത
𝐯}𝒊 =
𝐯 𝑻
𝒊. 𝐯 𝒊

𝐯}𝒊 is {ത
The norm (magnitude) of a vector{ത 𝐯}𝒊 = 𝟏, (𝒖𝒏𝒊𝒕 𝒗𝒆𝒄𝒕𝒐𝒓)

Define matrix [P] = [ 𝐯ത 𝟏 , {ത

𝐯}𝟐 , {ത
𝐯}𝟑 , …., {ത
𝐯 }𝒏 ]

Since the colons of [P] are orthogonal and has a normal , {ത

𝐯}𝒊 = 𝟏, matrix [P] is called orthonormal matrix . It is
also symmetric matrix. Therefore

Matrices that satisfy the equation P T . [P] = [I] are called orthogonal matrices.

Also note that P −1 = P 𝑇

 4.49. The vibration is the vertical direction of an airplane and its wings can be modeled as a three-degree-of-
freedom system, with one mass corresponding to the right wing, one mass for the left wing, and one mass for the
fuselage. The stiffness connecting the three masses corresponds to that of the wing and is a function of the
modulus E of the wing. The equation of motion is

The model is given in the Figure. Calculate the natural frequencies and mode
shapes. Plot the mode shapes and interpret them according to the airplane’s
deflection.

Solve for the free response of the system, where E = 6.9*109 N/m2, l = 2 m, m
= 3000 kg, and I = 5.2 * 10-6 m4. Let the initial displacement correspond to a
ሶ
gust of wind that causes an initial condition of 𝒙(0) = 0, x(0) = {0.2 0 0}T m.
Discuss your solution.
 −1 1 1 0 0 𝐸𝐼 3 −3 0
𝑀2 = 0 0.5 0 [𝐊] = 3 −3 6 −3
𝑚 𝐿
0 0 1 0 −3 3

෩
Mass normalized matrix 𝐾

−1 −1 1 0 0 3 −3 0 1 0 0
1 𝐸𝐼 1
෩ = 𝑀
𝐾 2 . 𝐊. 𝑀 2 = 0 0.5 0 −3 6 −3 0 0.5 0
𝑚 𝐿3 𝑚
0 0 1 0 −3 3 0 0 1

𝐸𝐼 3.0 −1.5 0.0

෩ =
𝐾 −1.5 1.5 −1.5
𝑚 𝐿3
0.0 −1.5 3.0
 𝐸𝐼 3.0 −1.5 0.0
෩ =
𝐾 −1.5 1.5 −1.5
𝑚 𝐿3
0.0 −1.5 3.0

෩ − 𝜆2 𝐼 = 0 13.5 𝐸 2 𝐼 2 𝜆2 + 7.5 𝐸 𝐼 𝐿3 𝑚 𝜆4 − 𝐿6 𝑚2 𝜆6
Eigenvalues: 𝐾 =0
𝐿6 𝑚2
13.5 𝐸 2 𝐼 2 𝜆2 + 7.5 𝐸 𝐼 𝐿3 𝑚 𝜆4 − 𝐿6 𝑚2 𝜆6 = 0
EI EI EI EI
𝜆2 = 0, 𝜆2 = 3 3 , 𝜆2 = 4.5 3 , 𝜔𝑛1 = 0, 𝜔𝑛2 = +1.73 , 𝜔 = +2.12
𝐿 𝑚 𝐿 𝑚 𝐿3 𝑚 𝑛3 𝐿3 𝑚
Eigenvectors (normalized by magnitude) :

EI EI
𝜔𝑛1 = 0, v 1 = 𝜔𝑛2 = 1.73 3 , v 2 = 𝜔𝑛3 = 2.12 3 , v 2 =
𝐿 𝑚 𝐿 𝑚

Mode shapes:
−1 −1 1 1
𝑥 = M 2 𝑞 ⇒ u 1 = M 2 v 1 = =
𝑚 𝑚
 −1 −1
𝑥 = M 2 𝑞 ⇒ ui= M 2 v i
−1
The relationship between the mode shapes and eigenvectors u i is just u i = M 2 v i

First mode: semidefinite mode: Rigid body translation:

u 1
0.41
−1 1 1
𝜔𝑛1 = 0, u 1 = M 2 v 1 = =
𝑚 𝑚
𝑚 4𝑚 𝑚

uത 1 =

ഥ : over head bar means normalized by its magnitude

i

uത i =1
Second mode:

EI −1 u 2
𝜔𝑛2 = 1.73 3 , u 2 = M 2 v 2 = 0.71
𝐿 𝑚
4𝑚 𝑚
𝑚
1 1 -0.71
u 2 = =
𝑚 𝑚

uത 2 =
 Third mode:

EI −1 u 3 0.58
𝜔𝑛3 = 2.12 3 , u = M 2 v 0.58
3 3
𝐿 𝑚
4𝑚
𝑚 𝑚
1 1 -0.58
u 3 = =
𝑚 𝑚

uത 3 =
Orthogonal 𝑚𝑒𝑎𝑛𝑠 𝑓𝑜𝑟 𝑒𝑣𝑒𝑟𝑦𝑗 ≠ 𝑖, 𝑡ℎ𝑒𝑛 u 𝑇𝑖 . u 𝑗 =0, 𝑗 ≠ 𝑖

Unlike the Eigenvectors v i , in general the mode-shape vectors u i are not orthogonal:

Example: u 1 and u 3 are not orthogonal:

0.41
𝑇
u 3. u 1 = 0.58, −0.58, 0.58 . 0.41 = (0.58)(0.41)+(−0.58)(0.41)+(0.58)(0.41)≠ 0
0.41
2) Introduce second coordinate transformation defined as, 𝐪 = 𝐏 {𝐫} (37a)

Substitute into Eq. 35d, then ෩ 𝐏 𝐫 =𝟎

𝐏 𝐫ሷ + 𝐊

Pre − multiply by 𝐏 𝐓 , 𝐏 𝐓
𝐏 𝐫ሷ + 𝐏 𝐓 ෩ 𝐏 𝐫 =𝟎
𝐊
𝑰 Λ

𝐫ሷ + Λ 𝐫 = 𝟎, (𝟑𝟕𝐛)

Λ : is diagonal matrix, whose elements are the eigenvalues or the square of the natural frequencies.

2
𝜆12 𝜔𝑛1
𝜆22 2
𝜔𝑛2
Λ = = Λ = , (𝟑𝟕𝐜)
⋱ ⋱
𝜆2𝑛 2
𝜔𝑛𝑛
Equation 37a are decoupled equations, they are

𝑟1ሷ + 𝝎𝟐𝒏𝟏 𝑟1 = 𝟎
𝑟2ሷ + 𝝎𝟐𝒏𝟐 𝑟2 = 𝟎,
(𝟑𝟕𝐜)
⋮
𝑟𝑛ሷ + 𝝎𝟐𝒏𝒏 𝑟𝑛 = 𝟎,

The coordinates 𝑟𝑖, 𝑖 = 1, 2, … . 𝑛 are called modal coordinates and equations 37d are called modal equations.
To get the solution you need also to transform the IC,s.
Note that:
𝟏 𝟏
− −
𝐱 = 𝐌 𝟐 𝐪 = 𝐌 𝟐 𝐏 𝐫,
𝐒
𝐂𝐚𝐥𝐥𝐞𝐝 𝐦𝐨𝐝𝐚𝐥 𝐌𝐚𝐭𝐫𝐢𝐱
𝟏
𝐫 = 𝐏 𝑻 𝐌 𝟐 𝐱 ,

The solutions of Eqs. 37d subject to is,

𝟏
𝑻
𝐫(𝟎) = 𝐏 𝐌 𝟐 𝐱𝟎 , (37f)

The solution of Eq’s 37d subject to IC’s of EQ. 37f is,

𝝎𝟐𝒏𝒊 (𝒓𝒊 𝟎 𝟐 + 𝒓ሶ 𝟎 𝟐
𝝎𝒏𝒊 𝒓𝒊 𝟎
𝒓𝒊 (𝒕) = 𝒔𝒊𝒏 𝝎𝒏𝒊 𝒕 + 𝒕𝒂𝒏−𝟏 , 𝐢 = 𝟏, 𝟐, … , 𝐧, (37g)
𝝎𝒏𝒊 𝒓ሶ 𝒊 𝟎
Note we have introduced two coordinate transformations; even we can combine this
transformation into one transformation as follows
−1  −1
x = M  2 q  x = M  2  P r ,

(38a)
q =  P r 
Or,
−1
x = Sr , S  =  M  2  P  . (38b)

Matrix S  is called modal matrix. Its columns are the mode shapes of the motion
Introduce Eq. 38b into Eq. 34a to get,

[M ]S r + [K ] S r = 0 , (38c)

Pre-multiply by S  ,
T

ST [M]S r + S  [K ]S r = 0

T
(38d)
[ Λ]
−1 −1 −1 −1
 P  M  2 [M]M  2  P r +  P  M  2 [K ]M  2  P r = 0
T T

I I
I [ Λ]
r + [ Λ]r = 0 (38e)
Example : multi degree of freedom system

x1(t ) x2 (t ) x3 (t ) 𝑥𝑛−1 (𝑡) 𝑥𝑛 (𝑡)

F1 (t ) F2 (t ) F3 (t ) 𝐹𝑛−1 (𝑡) 𝐹𝑛 (𝑡)
k1 k2 k3
𝑚𝑛−1 𝑚𝑛
𝑘𝑛−1 𝑘𝑛
…….

Equations of motion: 𝑚𝑖 𝑥𝑖ሷ + 𝑘𝑖 𝑥𝑖 − 𝑥𝑖−1 − 𝑘𝑖+1 𝑥𝑖+1 − 𝑥𝑖 = 0, i = 1,2,3 … , n

𝑘1 + 𝑘2 −𝑘2 0 0 ... 0
−𝑘2 𝑘2 + 𝑘3 −𝑘3
𝐌 = 𝑑𝑖𝑎𝑔𝑜𝑛𝑎𝑙 𝑚1 , 𝑚2 , … … 𝑚𝑛 , 0 −𝑘3 𝑘3 + 𝑘4 −𝑘4
K =
0 −𝑘4 ... −𝑘𝑛−2
−𝑘𝑛−2 𝑘𝑛−2 + 𝑘𝑛−1 −𝑘𝑛−1
−𝑘𝑛−1 𝑘𝑛−1 + 𝑘𝑛
 Example4.4.2 : Three degrees of freedom x1(t ) x2 (t ) x3 (t )

k1
𝐹1 = (𝑡)
𝒎𝟏 = 𝟒, 𝒎𝟐 = 𝟒 𝒌𝒈, 𝒎𝟑 = 𝟒 𝒌𝒈,
𝑵 x3 (t ) k3
𝒌𝟏 = 𝟒, 𝒌𝟐 = 𝟒 , 𝒌𝟑 = 𝟒 𝑵/𝒎
𝒎
Equations of motion: 𝑚𝑖 𝑥𝑖ሷ + 𝑘𝑖 𝑥𝑖 − 𝑥𝑖−1 − 𝑘𝑖+1 𝑥𝑖+1 − 𝑥𝑖 = 0, i = 1,2,3
𝑚1 𝑥1ሷ + 𝑘1 𝑥1 − 𝑘2 𝑥2 − 𝑥1 = 0
𝑚2 𝑥2ሷ + 𝑘2 𝑥2 − 𝑥1 − 𝑘3 𝑥3 − 𝑥2 = 0,
𝑚3 𝑥3ሷ + 𝑘3 𝑥3 − 𝑥2 = 0

Matrix form: ሷ
𝐌 𝐱(𝐭) + 𝐊 𝐱(𝐭) = {𝐅(𝐭)}
4 0 0 8 −4 0
𝐌 = 𝑑𝑖𝑎𝑔𝑜𝑛𝑎𝑙 4,4,4 , = 0 4 0 K = −4 8 −4
0 0 4 0 −4 4
1
Solution: 0 0
2 1 0 0
𝟏
−𝟐 1 1 1
1) 𝐌 = 0 2
0 = 0 1 0 = 2𝐈
2
1 0 0 1
0 0
2
𝟏 𝟏 2 −1 0
−𝟐 −𝟐 1
෩ = 𝐌
2) Mass normalized stiffness: 𝐊 K 𝐌 = M = −1 2 −1
4
0 −1 1
 𝜆 − 2 −1 0 x1 (t ) x2 (t ) x3 (t )
෩ = −1 𝜆 − 2 −1
3) Determinant: 𝜆 𝐈 − 𝐊 𝐹1 = (𝑡)
0 −1 𝜆 − 1 k1
= 𝜆−2 𝜆−2 𝜆−1 −1 − 𝜆−1 =0
x3 (t ) k3

Solve for 𝜆: 𝜆 − 2 𝜆−2 𝜆−1 −1 − 𝜆−1 =0

𝜆3 − 5𝜆2 + 6𝜆 − 1 = 0
𝜆1 = 0.19806, 𝜆2 = 1.555, 𝜆3 → 3.2477

𝜔𝑛1 = 𝜆1 = 0.4450, 𝜔𝑛2 = 𝜆2 = 1.247, 𝜔𝑛3 = 𝜆3 = 1.801938

𝜆𝒊 − 2 −1 0 𝑣1𝑖
෩ 𝐯 = −1
𝜆𝒊 𝐈 − 𝐊 𝜆𝒊 − 2 −1 𝑣2𝑖
0 −1 𝜆𝒊 − 1 𝑣3𝑖
0.4450𝑣13 0.4450
− 1.8019𝑣11 + 1. 𝑣11 = 0,
𝐯 1 = 0.8019𝑣13 =𝑣13 0.8019 ,
𝜆1 = 0.19806, 𝑣11 − 1.8019𝑣21 + 𝑣21 = 0,
𝑣13 1
𝑣21 − 0.8019𝑣31 = 0
,
 Example:

−0.4450𝑣12 + 𝑣22 = 0 −1.247

𝜆2 = 1.555, 𝑣12 − 0.445𝑣22 + 𝑣32 = 0 𝐯 2 = 𝑣32 −0.555 ,
𝑣22 + 0.555𝑣32 = 0 1

1.247𝑣13 + 1𝑣23 = 0 1.8019

𝜆3 = 3.247, 𝑣13 + 1.2467𝑣23 + 𝑣33 = 0 𝐯 3 = 𝑣33 −2.247 ,
𝑣23 + 2.2467𝑣33 = 0 1
 0.4450 x1(t ) x2 (t ) x3 (t )
4)𝑁𝑜𝑟𝑚𝑎𝑙𝑖𝑧𝑒 𝐯 1 : 𝐯 𝑻
1. 𝐯 1 = 𝑣13 0.8019 , k1
𝐹1 = (𝑡)
1 x3 (t ) k3

𝐯 = 𝐯 𝑻 𝐯 1 = 𝑣13 (0.44502 + 0.80192 + 1)=1.357 𝑣13 ,

1 1.

0.328 −0.7370 −0.591

.𝐯1
𝐯 1,nor𝑚 = = 0.591 . Similarly, . 𝐯 2,norm = −0.3280 , 𝐯 3,norm = 0.7370
𝐯1
0.737 0.5910 −0.3280

0.328 −0.7370 −0.5910 −1

1/2 0 0 1 0 0
1
The matrix [P] is then, [𝑷] = 0.591 −0.3280 0.7370 . [𝐌] = 0
2 1/2 0 = 0 1 0.
2
0.737 0.5910 −0.3280 0 0 1/2 0 0 1

−1 1 0 0 0.328 −0.7370 −0.5910 0.164 −0.3685 −0.2955

1
5) The matrix [S] is then, [𝐒] = 𝐌 2 [𝐏] = 0 1 0 0.591 −0.3280 0.7370 = 0.2955 −0.164 0.3685
2
0 0 1 0.737 0.5910 −0.3280 0.3685 0.2955 −0.164

1 0.328 0.591 0.737 1 0 0 0.656 −1.182 1.474

[𝐒]−1 = 𝐓
[𝐏] [𝐌]2 = 2 −0.7370 −0.3280 0.5910 0 1 0 = −1.474 −0.656 1.182
−0.5910 0.7370 −0.3280 0 0 1 −1.182 1.474 −0.656
6) The initial conditions in the modal coordinates 𝐫 are
𝒙𝟏 (𝟎) 𝟏 𝒓𝟏 (𝟎) 𝟎. 𝟔𝟓𝟔 −𝟏. 𝟏𝟖𝟐 𝟏. 𝟒𝟕𝟒 𝟏 𝟎. 𝟔𝟓𝟔
𝒙 𝟎 = 𝒙𝟐 (𝟎) = 𝟎 , 𝐫 𝟎 = 𝑺−𝟏 𝐱 𝟎 , 𝒓 𝟎 = 𝒓𝟐 (𝟎) = −𝟏. 𝟒𝟕𝟒 −𝟎. 𝟔𝟓𝟔 𝟏. 𝟏𝟖𝟐 𝟎 = −𝟏. 𝟒𝟕𝟒
𝒙𝟑 (𝟎) 𝟎 𝒓𝟑 (𝟎) −𝟏. 𝟏𝟖𝟐 𝟏. 𝟒𝟕𝟒 −𝟎. 𝟔𝟓𝟔 𝟎 −𝟏. 𝟏𝟖𝟐
𝒙ሶ 𝟏 (𝟎) 𝟎 𝐴
𝐴𝑖
𝒙ሶ 𝟎 ሶ
= 𝒙(𝟎) = 𝟎 , 𝒓ሶ 𝟎 = 𝑺−𝟏 𝒙ሶ 𝟎 = 𝟎, 𝑟𝑖 (0) 𝑟𝑖ሶ 0
𝒙ሶ 𝟑 (𝟎) 𝟎 𝜙 𝜔𝑛𝑖
𝜙
7) The solution 𝐫 are given as 𝑟𝑖ሶ 0
𝜔𝑖
𝝎𝟐𝒏𝒊 (𝒓𝒊 𝟎 )𝟐 + (𝒓ሶ 𝒊 𝟎 )𝟐 𝝎𝒏𝒊 (𝒓𝒊 (𝟎) 𝐴𝑖 𝑟𝑖 (0)
𝒓𝒊 𝒕 = 𝒔𝒊𝒏 𝝎𝒏𝒊 𝒕 + 𝒕𝒂𝒏−𝟏 , 𝒊 = 𝟏, 𝟐, … 𝒏
𝝎𝒏𝒊 (𝒓ሶ 𝒊 (𝟎)
𝟎. 𝟒𝟒 𝟐 𝟎. 𝟔𝟓𝟔 𝟐 𝟎. 𝟒𝟒 𝟎. 𝟔𝟓𝟔 𝝅
𝒓𝟏 𝒕 = 𝒔𝒊𝒏 𝟎. 𝟒𝟒 𝒕 + 𝒕𝒂𝒏−𝟏 = 𝟎. 𝟔𝟓𝟔𝒔𝒊𝒏 𝟎. 𝟒𝟒𝒕 + = 𝟎. 𝟔𝟓𝟔𝒄𝒐𝒔 𝟎. 𝟒𝟒𝒕
𝟎. 𝟒𝟒 𝟎 𝟐
𝟏.𝟐𝟒𝟕𝟐 −𝟏.𝟒𝟕𝟒 𝟐 𝟏.𝟐𝟒𝟕 −𝟏.𝟒𝟕𝟒𝒕 𝝅
𝒓𝟐 𝒕 = 𝒔𝒊𝒏 𝟏. 𝟐𝟒𝟕𝒕 + 𝒕𝒂𝒏−𝟏 = −𝟏. 𝟒𝟕𝟒𝒔𝒊𝒏(𝟏. 𝟐𝟒𝟕𝒕 + 𝟐 ) = −𝟏. 𝟒𝟕𝟒𝒄𝒐𝒔(𝟏. 𝟐𝟒𝟕𝒕)
𝟏.𝟐𝟒𝟕 𝟎

𝟏.𝟖𝟎𝟏𝟗𝟐 −𝟏.𝟏𝟖𝟐 𝟐 𝟏.𝟐𝟒𝟕 −𝟏.𝟏𝟖𝟐𝒕 𝝅

𝒓𝟑 𝒕 = 𝒔𝒊𝒏 𝟏. 𝟖𝟎𝟏𝟗𝒕 + 𝒕𝒂𝒏−𝟏 = −𝟏. 𝟏𝟖𝟐𝒔𝒊𝒏(𝟏. 𝟖𝟎𝟏𝟗𝒕 + 𝟐 )=−𝟏. 𝟏𝟖𝟐𝒄𝒐𝒔(𝟏. 𝟖𝟎𝟏𝟗𝒕)
𝟏.𝟖𝟎𝟏𝟗 𝟎
 𝑟1 𝑡 = 0.656𝑐𝑜𝑠 0.44𝑡 0.656𝑐𝑜𝑠 0.44𝑡
𝑟2 𝑡 = −1.474𝑐𝑜𝑠(1.247𝑡) {𝐫 𝑡 } = −1.474𝑐𝑜𝑠(1.247𝑡
𝑟3 𝑡 = −1.182𝑐𝑜𝑠(1.8019𝑡) −1.182𝑐𝑜𝑠(1.8019𝑡)

8) The solution in terms of the physical coordinates 𝐱(𝐭) are

𝐱(𝑡) = 𝐒 {𝐫 𝑡 }

0.164 −0.3685 −0.2955 0.656𝑐𝑜𝑠 0.44𝑡

𝒙(𝑡) = 0.2955 −0.164 0.3685 −1.474𝑐𝑜𝑠(1.247𝑡)
0.3685 0.2955 −0.164 −1.182𝑐𝑜𝑠(1.8019𝑡)

0.1075𝑐𝑜𝑠 0.44𝑡 + 0.544𝑐𝑜𝑠(1.247𝑡) + 0.3492𝑐𝑜𝑠(1.8019𝑡)

𝒙(𝑡) = 0.1938𝑐𝑜𝑠 0.44𝑡 + 0.2417𝑐𝑜𝑠(1.247𝑡) − 0.4355𝑐𝑜𝑠(1.8019𝑡)
0.2417𝑐𝑜𝑠 0.44𝑡 − 0.4355𝑐𝑜𝑠(1.247𝑡) + 0.1934𝑐𝑜𝑠(1.8019𝑡)