Advanced Structural Dynamics and Vibration Control

Continuous Structure

Kishor Timsina, PhD

CONTINUOUS STRUCTURES: PARTIAL DIFFERENTIAL EQUATIONS OF MOTION

Dynamics of some simple continuous structures (strings, beams, rods) will be studied in this Chapter.

For these structures, it is possible to formulate ‘exact’ governing equations of motion in the form of
Partial Differential Equations (PDE). It is also possible to find ‘exact’ analytical solutions (response)
to some dynamic loadings.

More ‘physical insight’ of dynamics of structures can be gained.

Moreover, in some loading cases, especially where the duration of loading is very short compared
with the fundamental period, many higher modes may be excited significantly in such a way that
‘modal analysis’ may not be an appropriate analysis approach.

Example: an axial vibration in a pile due to a drop hammer.

For such cases, ‘wave propagation analysis’ which is based upon PDE may be more appropriate.
The wave solutions may be derived from PDE in a straight forward manner.
CONTINUOUS STRUCTURES: PARTIAL DIFFERENTIAL EQUATIONS OF MOTION:

Discrete model Continuous model

Matrix equations (ODE) PDE

Finite degree of freedom ∞ DOF

Finite Eigenvalues ∞ Eigenvalues
Finite Eigenvectors ∞ Eigenfunctions
TRANSVERSE VIBRATION OF STRING:
𝑝(𝑥, 𝑡)

𝑚(𝑥): mass per unit length (kg/m)

A Free-Body Diagram:
𝑇(𝑥, 𝑡): String tension (N) An Element of String-Length 𝑑𝑥

𝑝(𝑥, 𝑡): distributed force per unit length (N/m)

(No gravity load)
TRANSVERSE VIBRATION OF STRING:

D’ Alembert’s direct equilibrium of dynamic forces in vertical direction:

𝜕𝑇 𝜕𝑢 𝜕 𝑢 𝜕𝑢 𝜕 𝑢
𝑇(𝑥, 𝑡) + 𝑑𝑥 + 𝑑𝑥 − 𝑇(𝑥, 𝑡) + 𝑝(𝑥, 𝑡) 𝑑𝑥 − 𝑚(𝑥) 𝑑𝑥 =0
𝜕𝑥 𝜕𝑥 𝜕𝑥 𝜕𝑥 𝜕𝑡
𝑝(𝑥, 𝑡)
For small vibration, it is reasonable to assume that

𝑇 𝑥, 𝑡 = 𝑇(𝑥) = initial static tension

(that is, the dynamic change of tension is negligible)

For a uniform string, 𝑇(𝑥) = 𝑇, 𝑚(𝑥) = 𝑚

We can obtain: 𝜕 𝑢 𝜕 𝑢
𝑚 − 𝑇 = 𝑝 𝑥, 𝑡
𝜕𝑡 𝜕𝑥
for 0 < 𝑥 < 𝐿 (1)
TRANSVERSE VIBRATION OF STRING:

The Partial Differential Equation of motion (PDE) for transverse vibration of string:

𝜕 𝑢 𝜕 𝑢
𝑚 −𝑇 = 𝑝 𝑥, 𝑡 for 0 < 𝑥 < 𝐿 (1)
𝜕𝑡 𝜕𝑥

By the condition that the two ends of string are fixed ends:

𝑢(0, 𝑡) = 𝑢(𝐿, 𝑡) = 0 (2)

The associated boundary conditions
TRANSVERSE VIBRATION OF BEAM:
𝑝 𝑥, 𝑡 𝑑𝑥
𝑝(𝑥, 𝑡)

D’ Alembert’s direct equilibrium of dynamic forces in vertical direction:

𝐸𝐼 (𝑥): Bending rigidity 𝜕𝑉 𝜕 𝑢
𝑉 (𝑥, 𝑡) + 𝑑𝑥 − 𝑉 (𝑥, 𝑡) + 𝑝(𝑥, 𝑡)𝑑𝑥 − 𝑚(𝑥)𝑑𝑥 =0
𝑀(𝑥, 𝑡): Bending moment 𝜕𝑥 𝜕𝑡
𝑉 (𝑥, 𝑡): Shear force
𝜕 𝑢 𝜕𝑉
𝑚𝑥 − = 𝑝(𝑥, 𝑡) (3)
𝜕𝑡 𝜕𝑥
TRANSVERSE VIBRATION OF BEAM:

Assuming that rotational angle is so small that the inertia torque is negligible, the moment equilibrium
about N-A at right hand face at the segment is

𝜕𝑀 𝑑𝑥 𝜕 𝑢 𝑑𝑥
𝑀 𝑥, 𝑡 + 𝑑𝑥 − 𝑀 𝑥, 𝑡 + 𝑉 𝑥, 𝑡 𝑑𝑥 − 𝑝 𝑥, 𝑡 𝑑𝑥 − 𝑚 𝑥 𝑑𝑥 = 0
𝜕𝑥 2 𝜕𝑡 2

second-order terms
𝑝 𝑥, 𝑡 𝑑𝑥

𝜕𝑀
= −𝑉 (4)
𝜕𝑥
TRANSVERSE VIBRATION OF BEAM:

From the mechanics of beam (Bernoulli - Euler beam):

𝜕 𝑢
𝑀(𝑥, 𝑡) = 𝐸𝐼 (𝑥) (5)
𝜕𝑥

Introducing Eq (4) and Eq (5) into Eq (3), we obtain

𝜕 𝑢 𝜕 𝜕 𝑢
𝑚𝑥 + 𝐸𝐼 𝑥 = 𝑝(𝑥, 𝑡) for 0 < 𝑥 < 𝐿
𝜕𝑡 𝜕𝑥 𝜕𝑥 (6)

Eq (6) is the partial differential equation governing the motion 𝑢 (𝑥, 𝑡) of the beam, which is
subjected to the external dynamic forces 𝑝(𝑥, 𝑡).
TRANSVERSE VIBRATION OF BEAM:

For a uniform beam, 𝑚 𝑎𝑛𝑑 𝐸𝐼 would be constant:

𝜕 𝑢 𝜕 𝑢
𝑚 + 𝐸𝐼 = 𝑝(𝑥, 𝑡) for 0 < 𝑥 < 𝐿
𝜕𝑡 𝜕𝑥 (7)

Boundary conditions:
𝑢 0, 𝑡 = 𝑢 𝐿, 𝑡 = 0
(8)
𝜕 𝑢 𝜕 𝑢
(0, 𝑡) = (𝐿, 𝑡) = 0 Hinged-Roller:
𝜕𝑥 𝜕𝑥 displacement and moments are zeros

𝑢 0, 𝑡 = 𝑢 𝐿, 𝑡 = 0
(9)
𝜕𝑢 𝜕𝑢
0, 𝑡 = 𝐿, 𝑡 = 0 Clamped ends:
𝜕𝑥 𝜕𝑥
TRANSVERSE VIBRATION OF BEAM:

Boundary conditions:

Cantilever beam:

𝜕𝑢
𝑢 0, 𝑡 = 0, (0, 𝑡) = 0
𝜕𝑥
𝜕 𝑢
𝐿, 𝑡 = 0 ← Moment (10)
𝜕𝑥
𝜕 𝑢
𝐿, 𝑡 = 0 ← Shear
𝜕𝑥
AXIAL VIBRATION OF ROD:

: axial displacement

𝜕 𝑢
𝑓 = 𝑚 𝑑𝑥
𝜕𝑡

𝜕 𝑢 𝜕𝑁
Dynamic equilibrium: 𝑚 − =0 (11)
𝜕𝑡 𝜕𝑥
No external force

𝜕𝑢
From the mechanics of rod: 𝑁 = 𝜎 𝐴 = 𝐸 𝜖 𝐴 = 𝐸𝐴 (12)
𝜕𝑥
𝜕 𝑢 𝜕 𝑢
Eq (12)  Eq (11): 𝑚 − 𝐸𝐴 =0 for 0 < 𝑥 < 𝐿
𝜕𝑡 𝜕𝑥 (13)
P.D.E for the axial vibration of rod.
AXIAL VIBRATION OF ROD:

P.D.E for the axial vibration of rod.

𝜕 𝑢 𝜕 𝑢
𝑚 − 𝐸𝐴 =0 for 0 < 𝑥 < 𝐿 (13)
𝜕𝑡 𝜕𝑥

Boundary conditions:

Fixed end: 𝑢(0, 𝑡) = 0

𝜕𝑢
Free-end: (𝐿, 𝑡) = 0 when 𝑃 (𝑡) = 0
𝜕𝑥
(14)
𝜕𝑢 𝑃𝑡
or (𝐿, 𝑡) = when 𝑃 (𝑡) ≠ 0
𝜕𝑥 𝐸𝐴
Natural Frequencies and Mode Shapes: Transverse vibration of string

𝜕 𝑢 𝜕 𝑢
PDE: 𝑚 −𝑇 = 0 for 0 < 𝑥 < 𝐿
𝜕𝑡 𝜕𝑥 (1)

BC: 𝑢 0, 𝑡 = 𝑢 𝐿, 𝑡 = 0 (2)

Assuming that the solution can be

expressed in a separated variable form:
𝑢(𝑥, 𝑡) = 𝜙(𝑥) 𝑞(𝑡) (3)

𝜕 ( )
Substituting Eq (3) to Eq (1) leads to: 𝑚 𝑞̈ (𝑡) 𝜙(𝑥) = 𝑇 𝑞 (𝑡) 𝜙 (𝑥) where ̈ =
𝜕𝑡
𝜕 ( )
( ) =
Rearrange the above equation: 𝑞̈ (𝑡) 𝑇 𝜙 (𝑥) 𝜕𝑥
= (4)
𝑞 (𝑡) 𝑚 𝜙(𝑥)

a function of 𝑡 a constant a function of 𝑥

Natural Frequencies and Mode Shapes:

Define 𝑐 ≡ 𝑇⁄𝑚

Eq (4) is satisfied for any arbitrary 𝑥 and 𝑡 only if L.H.S and R.H.S are equal to a constant:

𝑞̈ (𝑡) 𝜙 (𝑥)
= 𝑐 = constant = −𝜔
𝑞 (𝑡) 𝜙(𝑥)

That is: 𝑞̈ 𝑡 + 𝜔 𝑞 𝑡 = 0 (5)

Eq (5) and Eq (6) are ODE
in 𝑡 and 𝑥, respectively.
𝜔
and 𝜙′′(𝑥) + 𝜙(𝑥) = 0 (6)
𝑐

𝜔
Define 𝜆 = Eq. 6 can be rewritten as: 𝜙 (𝑥) + 𝜆 𝜙(𝑥) = 0 (6)
𝑐
Natural Frequencies and Mode Shapes:

𝑞̈ 𝑡 + 𝜔 𝑞 𝑡 = 0 (5)

𝜙 𝑥 +𝜆 𝜙 𝑥 = 0 (6)

The general solutions of Eqs (5) and (6) are, respectively,

𝑞 𝑡 = 𝐴 sin 𝜔𝑡 + 𝐵 cos 𝜔𝑡 (7)

𝜙 𝑥 = 𝐶 sin 𝜆𝑥 + 𝐷 cos 𝜆𝑥 (8)

where 𝐴, 𝐵, 𝐶, 𝐷 are arbitrary constants (to be determined from BC.)

𝜔 and 𝜆 are constants but they not ‘arbitrary’ constants.

Natural Frequencies and Mode Shapes:

Let’s consider BC: 𝑢 0, 𝑡 = 𝜙 0 𝑞 𝑡 = 0 ∴ 𝜙0 = 0

(9)
𝑢 𝐿, 𝑡 = 𝜙 𝐿 𝑞 𝑡 = 0 ∴ 𝜙(𝐿) = 0

Applying BC (Eq (9)) to Eq (8), we obtain

𝜙 0 = 𝐶 sin 𝜆. 0 + 𝐷 cos 𝜆. 0 = 𝐶. 0 + 𝐷. 1 = 0
𝜙 𝐿 = 𝐶 sin 𝜆𝐿 + 𝐷 cos 𝜆𝐿 = 0

In matrix form: 0 1 𝐶 0
= (10)
sin 𝜆𝐿 cos 𝜆𝐿 𝐷 0
𝑨𝒙=𝟎
In order to obtain non-zero 𝐶 & 𝐷 vector (non-zero 𝜙(𝑥)), the determinant of the matrix A
should vanish. This leads to
sin 𝜆𝐿 = 0 (11)
Natural Frequencies and Mode Shapes:

sin 𝜆𝐿 = 0 (11)

Eq. (11) is called “frequency equation”. There are ∞ solutions for this equation:
𝑛𝜋
𝜆𝐿 = 𝜋, 2𝜋, 3𝜋, … That is 𝜆 = for 𝑛 = 1, 2,3, … (12)
𝐿

𝑛𝜋 𝑇 for 𝑛 = 1, 2,3, …
The corresponding 𝜔 are given by 𝜔 = 𝜆 𝑐 = (13)
𝐿 𝑚

Substituting Eq. (12) back to Eq. (10), we obtain 𝐶 𝐶 𝐶 𝑖𝑠 𝑎 𝑛𝑜𝑛 − 𝑧𝑒𝑟𝑜 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡
= (14)
𝐷 0
𝑛𝜋𝑥
The corresponding mode shapes 𝜙 (𝑥) are: 𝜙 (𝑥) = 𝐶 sin for 𝑛 = 1, 2,3, …
𝐿
Natural Frequencies and Mode Shapes:

𝑆𝑖𝑛𝑐𝑒 𝐶 can be any value, for convenience, let 𝐶 = 1:

𝑛𝜋𝑥
𝜙 (𝑥) = sin for 𝑛 = 1, 2,3, … (15)
𝐿
In summary, the mode shapes and natural frequencies of a string are given by

𝜋𝑥 𝜋 𝑇
𝑛 =1, 𝜙 (𝑥) = sin , 𝜔 =
𝐿 𝐿 𝑚

2𝜋𝑥 2𝜋 𝑇
𝑛 =2, 𝜙 (𝑥) = sin , 𝜔 =
𝐿 𝐿 𝑚

3𝜋𝑥 3𝜋 𝑇
𝑛 =3, 𝜙 (𝑥) = sin , 𝜔 =
𝐿 𝐿 𝑚
Natural Frequencies and Mode Shapes:

The solution of 𝑢 (𝑥, 𝑡) given by Eq. (3) is:

𝑢 𝑥, 𝑡 = 𝜙 𝑥 𝑞 𝑡 = 𝜙 (𝑥) 𝐴 sin 𝜔 𝑡 + 𝐵 cos 𝜔 𝑡

The “general” solution (Free Vibration Response) is:

𝑖𝜋𝑥
𝑢(𝑥, 𝑡) = (𝐴 sin 𝜔 𝑡 + 𝐵 cos 𝜔 𝑡) sin (16)
𝐿
Natural Frequencies and Mode Shapes:

𝑖𝜋𝑥
𝑢(𝑥, 𝑡) = (𝐴 sin 𝜔 𝑡 + 𝐵 cos 𝜔 𝑡) sin (16)
𝐿

𝑖𝜋𝑥
At initial state (𝑡 = 0), 𝑢(𝑥, 0) = 𝐵 sin (17)
𝐿

Multiply Eq. (17) by sin and integrate over the length of the string,

𝑗𝜋𝑥 𝑖𝜋𝑥 𝑗𝜋𝑥 𝐵 𝐿

𝑢(𝑥, 0) sin 𝑑𝑥 = 𝐵 sin sin 𝑑𝑥 =
𝐿 𝐿 𝐿 2

2 𝑗𝜋𝑥
𝐵 = 𝑢(𝑥, 0) sin 𝑑𝑥 for 𝑗 = 1, 2, 3, … (18)
𝐿 𝐿
 Natural Frequencies and Mode Shapes:

𝑖𝜋𝑥
In similar manner, 𝑢̇ (𝑥, 𝑡) = (𝜔 𝐴 cos 𝜔 𝑡 − 𝜔 𝐵 sin 𝜔 𝑡) sin
𝐿

𝑖𝜋𝑥
𝑢̇ (𝑥, 0) = (𝜔 𝐴 ) sin
𝐿

𝑗𝜋𝑥 𝜔𝐴 𝐿
𝑢̇ (𝑥, 0) sin 𝑑𝑥 =
𝐿 2

2 𝑗𝜋𝑥 (19)
∴ 𝐴 = 𝑢̇ (𝑥, 0) sin 𝑑𝑥
𝐿𝜔 𝐿

By Eqs (18) and (19), the constants 𝐴 and 𝐵 can be determined when the initial 𝑢(𝑥, 0) and 𝑢̇ (𝑥, 0) are known.
So, we obtain the free vibration response 𝑢 (𝑥, 𝑡) in a complete manner as given by Eq. (16).
Orthogonality of Mode Shapes:

In analogy to MDOF discrete systems

where, 𝝓 𝑴𝝓 = 0
for 𝑖 ≠ 𝑗
𝝓 𝑲𝝓 = 0

It can be shown that

𝑚 𝑥 𝜙 𝑥 𝜙 𝑥 𝑑𝑥 = 0 for 𝑖 ≠ 𝑗
(20)
𝑚 𝑥 𝜙 𝑥 𝑑𝑥 = 𝜇 for 𝑖 = 𝑗

Mode shape functions are orthogonal to each other with respect to the mass distribution function.
Orthogonality of Mode Shapes:

and
0 for 𝑖 ≠ 𝑗
𝑇(𝑥) 𝜙 ′(𝑥) 𝜙 ′(𝑥) 𝑑𝑥 = (21)
𝜇 𝜔 for 𝑖 = 𝑗

Mode shape functions are orthogonal to each other with respect to the stiffness distribution function.

Another way of writing this orthogonality relationship:

0 for 𝑖 ≠ 𝑗
− 𝑇(𝑥) 𝜙 ′′(𝑥) 𝜙 (𝑥) 𝑑𝑥 = (22)
𝜇 𝜔 for 𝑖 = 𝑗
Orthogonality of Mode Shapes: Proof

Start from Eigen Equation: 𝜙 +𝜆 𝜙 =0 (a)

and 𝜙 +𝜆 𝜙 =0 (b)

Multiply Eq (a) by 𝜙 : 𝜙 𝜙 𝑑𝑥 + 𝜆 𝜙 𝜙 𝑑𝑥 = 0 (c)

Multiply Eq (b) by 𝜙 : 𝜙 𝜙 𝑑𝑥 + 𝜆 𝜙 𝜙 𝑑𝑥 = 0 (d)

Integration by part (Eq (c)): 𝜙 𝜙 𝑑𝑥 = 𝜙 𝜙 − 𝜙 𝜙 𝑑𝑥 (e)

zero

Integration by part (Eq (d)): 𝜙 𝜙 𝑑𝑥 = 𝜙 𝜙 − 𝜙 𝜙 𝑑𝑥 (f)

zero
Orthogonality of Mode Shapes:
∴ 𝜙 𝜙 𝑑𝑥 = 𝜙 𝜙 𝑑𝑥 (g)

Eq (𝑐) − Eq (𝑑): (𝜆 −𝜆 ) 𝜙 𝜙 𝑑𝑥 = 0 (h)

for 𝑖 ≠ 𝑗 ; 𝜆 ≠𝜆 ⇒ 𝜙 𝜙 𝑑𝑥 = 0 (i)

for 𝑖 = 𝑗 ; 𝜙 𝑑𝑥 = 0 ⇒ + 𝑑𝑒𝑓𝑖𝑛𝑖𝑡𝑒 (j)

Eq. (i) ⇒ Eq. (c) & Eq (e): 𝜙 𝜙 𝑑𝑥 = 0 𝑓𝑜𝑟 𝑖 ≠ 𝑗

1
Eq. (j) ⇒ Eq. (c) & Eq (e): 𝜙 𝑑𝑥 = 𝜙 𝑑𝑥 𝑓𝑜𝑟 𝑖 = 𝑗
𝜆

∴ 𝑚 𝜙 𝑑𝑥 𝜔 = 𝑇 𝜙 𝑑𝑥
 Forced Vibration: Modal Analysis

𝜕 𝑢 𝜕 𝑢 (23)
𝑚 − 𝑇 = 𝑝(𝑥, 𝑡)
𝜕𝑡 𝜕𝑥

The solution 𝑢(𝑥, 𝑡) is

expressed in the form of 𝑢(𝑥, 𝑡) = 𝜙 (𝑥) 𝑞 (𝑡) (24)

Introduce Eq (24) into Eq (23): 𝑚 𝜙 𝑥 𝑞̈ (𝑡) − 𝑇 𝜙 𝑥 𝑞 (𝑡) = 𝑝(𝑥, 𝑡) (25)

Multiply Eq (25) by 𝜙 (𝑥) and integrate the Equation over the whole length (𝑓𝑟𝑜𝑚 0 𝑡𝑜 𝐿)

𝑚 𝜙 (𝑥) 𝜙 (𝑥) 𝑞̈ (𝑡) 𝑑𝑥 − 𝑇 𝜙 𝑥 𝜙 𝑥𝑞 𝑡 𝑑𝑥 = 𝑝 𝑥, 𝑡 𝜙 (𝑥) 𝑑𝑥

(26)
Forced Vibration: Modal Analysis

Multiply Eq (25) by 𝜙 (𝑥) and integrate the Equation over the whole length (𝑓𝑟𝑜𝑚 0 𝑡𝑜 𝐿)

𝑚 𝜙 (𝑥) 𝜙 (𝑥) 𝑞̈ (𝑡) 𝑑𝑥 − 𝑇 𝜙 𝑥 𝜙 𝑥𝑞 𝑡 𝑑𝑥 = 𝑝 𝑥, 𝑡 𝜙 (𝑥) 𝑑𝑥

(26)

Employing the orthogonality of mode shapes (Eqs. (20 & 22))

𝜇 𝑞̈ + 𝜇 𝜔 𝑞 = 𝑄 (𝑡) for 𝑗 = 1, 2, … , ∞ (27)

modal mass modal stiffness

modal force 𝑄 (𝑡) = 𝑝(𝑥, 𝑡) 𝜙 (𝑥) 𝑑𝑥

Introduce modal damping

term into Eq (27):
𝜇 𝑞̈ + 2 𝜉 𝜔 𝜇 𝑞̇ + 𝜇 𝜔 𝑞 = 𝑄 (𝑡) (28)

PDE ∞ ODEs
Example of Modal Analysis

A steel cantilever pipe as shown in the figure1 is subjected to a harmonic force.

𝑃 sin 2𝜋𝑓𝑡

Figure 1 t = 0.02 m

10 m 10 m Circular Tube

The mass density of steel is 7800 kg/m3. 1m

The Young’s modulus of steel is 2 X 1011 N/m2. Figure 2 pipe’s cross section
Modal damping ratio is assumed to be 0.01 (every mode).

𝑃 = 10000 𝑁
 𝑃 sin 2𝜋𝑓𝑡

Figure 1

10 m 10 m

Based on this problem statement, the steady-state response amplitude at the beam’s tip,
denoted by “𝑎 ”, will be a function of the excitation frequency “𝑓 ” , that is,

𝑎 = 𝑎(𝑓) 𝑎(𝑓)

Show the relationship between 𝑎 and 𝑓 by plotting the

frequency response curve for

0.01 𝐻𝑧 < 𝑓 < 60 𝐻𝑧 in “log-scale graph”.

𝑓
Answer:

The equations of motion of the beam in terms of the modal coordinates:

𝜇 𝑞̈ + 2 𝜇 𝜉 𝜔 𝑞̇ + 𝜇 𝜔 𝑞 = 𝑄 (𝑡) for 𝑖 = 1,2, … . . , ∞ (1)

Where 𝜇 is the 𝑖𝑡ℎ modal mass: 𝜇 = 𝑚 𝜙 𝑑𝑥 (2)

𝜉 is the 𝑖𝑡ℎ modal damping ratio: 𝜉 = 0.01 for all 𝑖 (3)

𝜔 is the 𝑖𝑡ℎ modal (circular) frequency.

𝐸𝐼
For the case of a uniform cantilever beam 𝜔 =𝛽 (4)
𝑚𝐿
where 𝛽 = 1.8752 = 3.516; 𝛽 = 4.6942 = 22.03; 𝛽 = 7.8552 = 61.701
𝑑
𝐸𝐼 = Bending rigidity
𝐷 −𝑑
𝐸 = 2 × 10 𝑁/𝑚 , 𝐼= = 7.5 × 10 𝑚 (5)
20 𝐷
 𝜋𝐷 𝜋𝑑
𝑚 = mass per unit length = 7800 × − = 480 𝑘𝑔⁄𝑚
4 4
𝐿 = 20 𝑚

With all these values, by Eq (4) we obtain:

𝜔 = 15.536 rad/sec; 𝜔 = 97.38 rad/sec; 𝜔 = 272.7 rad/sec (6)

Note that the modal frequencies of the higher modes can also be calculated, but in this case
where the frequency response below 60 Hz is of interest, only the three modal responses would
be sufficient.
Next, consider the modal force: 𝑄 𝑡 = 𝜙 𝑃 sin(2𝜋𝑓𝑡) (7)

Where 𝜙 = 𝜙 𝑥 = 0.5𝐿 ; 𝜙 is the 𝑖𝑡ℎ mode shape

Beam’s tip response: 𝑦(𝑡) 𝑦(𝑡) = 𝜙 𝐿𝑞 𝑡 ≈ 𝜙 𝑞 (𝑡) (8)

Where 𝜙 =𝜙 𝑥=𝐿
Mode shapes of a uniform centiliver beam:
cos 𝑎 𝐿 + cosh 𝑎 𝐿 sin 𝑎 𝑥 − sinh 𝑎 𝑥
𝜙 (𝑥) = cos 𝑎 𝑥 − cosh 𝑎 𝑥 − (9)
sin 𝑎 𝐿 + sinh 𝑎 𝐿

𝑎 = 1.875/𝐿

𝑥/𝐿

𝜙 = −0.679
𝜙
𝜙 = −2.000
Mode shapes of a uniform centiliver beam:
cos 𝑎 𝐿 + cosh 𝑎 𝐿 sin 𝑎 𝑥 − sinh 𝑎 𝑥
𝜙 (𝑥) = cos 𝑎 𝑥 − cosh 𝑎 𝑥 − (9)
sin 𝑎 𝐿 + sinh 𝑎 𝐿

𝑎 = 4.694/𝐿

𝜙 = −1.427

𝜙 𝑥/𝐿 𝜙 = +2.000
Mode shapes of a uniform centiliver beam:
cos 𝑎 𝐿 + cosh 𝑎 𝐿 sin 𝑎 𝑥 − sinh 𝑎 𝑥
𝜙 (𝑥) = cos 𝑎 𝑥 − cosh 𝑎 𝑥 − (9)
sin 𝑎 𝐿 + sinh 𝑎 𝐿

𝑎 = 7.855/𝐿

𝜙 = −0.04
𝜙 𝑥/𝐿
𝜙 = −2.00
Introducing the mode shapes 𝜙 (𝑥) into Eq. (2):

𝜇 = 𝑚 𝜙 𝑑𝑥 ≅ 𝑚𝐿

𝜇 = 𝑚 𝜙 𝑑𝑥 ≅ 𝑚𝐿 (10)

𝜇 = 𝑚 𝜙 𝑑𝑥 ≅ 𝑚𝐿

∴ 𝜇 ≈ 𝜇 ≈ 𝜇 ≈ 480 × 20 = 9600 𝑘𝑔 (11)

Steady-State Response: 𝑞 𝑡 = 𝐴 cos 2𝜋𝑓𝑡 + 𝐵 sin 2𝜋𝑓𝑡 (12)

or 𝑞 𝑡 = 𝜌 cos(2𝜋𝑓𝑡 + 𝜃 ) (13)

𝑃𝜙 −2𝜉 𝛽
where 𝐴 = (14)
𝜇 𝜔 1−𝛽 + 2𝜉 𝛽

𝑃𝜙 1−𝛽
𝐵 = (15)
𝜇 𝜔 1−𝛽 + 2𝜉 𝛽

𝑃𝜙 1
𝜌 = (16)
𝜇 𝜔
1−𝛽 + 2𝜉 𝛽

2𝜋𝑓
𝛽 = (17)
𝜔
The contribution of the 1st modal 𝑃 𝜙 𝜙 1
response to the beam’s tip
= 𝜙 𝑞 = ×
𝜇 𝜔 2𝜉
response will reach its maximum
when 𝛽 = 1, that is 10000) × (−0.679) × (−2.0 1
= × = 0.293 𝑚
9600 × 15.536 2 × 0.01

In similar manner, the second 𝑃 𝜙 𝜙 1

mode’s max contribution is
= ×
𝜇 𝜔 2𝜉
(𝛽 = 1)
10000) × (−1.427) × (2.0 1
= × = −0.0156 𝑚
9600 × 97.38 2 × 0.01

The third mode’s max contribution is 𝑃 𝜙 𝜙 1

(𝛽 = 1) = ×
𝜇 𝜔 2𝜉
10000) × (−0.04) × (−2.0 1
= × = 5.6 × 10 𝑚
9600 × 272.7 2 × 0.01
The contribution of the 3rd mode is extremely low because 𝜔 ≫ 𝜔 ≫ 𝜔 and 𝜙 is close to zero.
To combine a modal responses into the actual response at beam’s tip, Eqs (14, 15) are
introduced into Eq (8):
𝑦(𝑡) = 𝜙 𝐴 cos 2𝜋𝑓𝑡 + 𝜙 𝐵 sin(2𝜋𝑓𝑡

That is 𝑦 𝑡 = 𝑎 cos 2𝜋𝑓𝑡 + 𝑏 sin(2𝜋𝑓𝑡) (18)

where 𝑎= 𝜙 𝐴 and 𝑏= 𝜙 𝐵 (19)

The amplitude of the tip response can be estimated from

𝑦 𝑡 = 𝑎 cos 2𝜋𝑓𝑡 + 𝜃 (20)

𝑎 = 𝑎 +𝑏 (21)
The amplitude of the tip response can be estimated from

𝑦(𝑡) = 𝑎 cos 2𝜋𝑓𝑡 + 𝜃 (20)

𝑎 = 𝑎 +𝑏 (21)

And it can be shown that 𝑎𝑦 = 𝑎 𝑓 :

𝑎𝑦 ≈ 0.293 𝑚

Log scale
0.0156 𝑚

5.6 × 10 𝑚

𝑓
Log scale
15.536/2𝜋 97.38/2𝜋 272.7/2𝜋
Thank You

42