Vibration Control of A Beam

Nodal Control of Vibrating Structures: Beam
A Thesis

Submitted to the Graduate Faculty of the
Louisiana State University and
Agricultural and Mechanical College
in partial fulfillment of the
requirements for the degree of
Master of Science in Mechanical Engineering
in
The Department of Mechanical Engineering

by
Akshay Nareshraj Singh

B.E. in Mechanical Engineering
Maharaja Sayajirao University, 1999

December 2001

To my parents Alka and Nareshraj Singh
and my younger brother Abhishek

ii
Cataloging Abstract
Vibration control is an important engineering problem and many methods for both
active and passive vibration absorption have been developed. This thesis deals with
developing a method to achieve nodal control at the point of excitation in a Bernoulli-
Euler beam. It is established that, for a uniform Bernoulli-Euler beam, the steady state
motion at the point of excitation can be absorbed by means of a control force
determined from displacement information at the point of application. A closed form
solution for the control gain is presented and a criterion for implementing the control by
active and passive means is developed. The result for the control gain is generalized for
the case of a non-uniform beam. Chapter 4 shows through some examples that the
theory can be also applied to eliminate the steady state motion at any desired location
other than the point of excitation. Analysis is also performed to determine the optimal
control force and investigate the stability of the overall system. Several controllability
graphs are shown and meaningful conclusions are drawn from these graphs.

An experiment is designed to validate the proposed theory and display its practicality.
A uniform steel beam supported at two locations is tested. Modal testing is performed
to extract natural frequencies in order to characterize the system and assist in
formulation of an appropriate mathematical model. The steel beam is then excited by a
known harmonic force supplied by a vibration exciter and a spring, with suitable spring
constant obtained by performing the control gain calculations on the model, is used to
absorb the motion at the free end. It is confirmed, that the theory developed in this
thesis produces accurate results, and that it can serve as a vital tool in developing
practical solutions to structural control problems.

Akshay Nareshraj Singh, B.E., Maharaja Sayajirao University, 1999
Master of Science, Fall Commencement, 2001
Major: Mechanical Engineering
Nodal Control of Vibrating Structures: Beam
Thesis directed by Associate Professor Yitshak Ram
Pages in thesis, 105. Words in abstract, 297

iii
Acknowledgements
The guidance and support provided by my major professor Dr. Yitshak Ram is
acknowledged. Thanks go to Dr. Michael Khonsari and Dr. Su-Seng Pang for serving
on my graduate committee and evaluating my thesis.

The intellectual input of my colleagues Kumar Vikram Singh, Jaeho Shim, Sumit
Singhal and Madhulika Sathe, as well as the assistance of Mr. Ed Martin in constructing
the experiment is also acknowledged.

Last but not the least, my gratitude goes to my aunty Mrs. Anita Singh and uncle Dr.
Vijay Singh who have always been there for me, and to my parents Alka and Nareshraj
Singh who have made me what I am today.

The research was supported in part by a National Science Foundation research grant
CMS-9978786.

iv
Table of Contents
Dedication..ii
Cataloging Abstract.iii
Acknowledgements........................................................................................................ iv
Table of Contents............................................................................................................ v
List of Tables................................................................................................................. vii
List of Figures .............................................................................................................. viii
List of symbols ................................................................................................................ x
Abstract ......................................................................................................................... xii
Chapter 1: Introduction................................................................................................. 1
Chapter 2: Literature Survey and Background .......................................................... 6
2.1 Introduction ...................................................................................................... 6
2.2 Passive vibration control .................................................................................. 6
2.3 Active vibration control.................................................................................... 9
Chapter 3: Beam: Theory and Background .............................................................. 11
3.1 Introduction .................................................................................................... 11
3.2 Equation of motion for a non-uniform beam.................................................. 12
3.3 Natural frequencies and modeshapes ............................................................. 15
3.4 Steady state response and natural frequencies for a uniform clamped
cantilever beam............................................................................................... 17
3.5 Summary......................................................................................................... 23
Chapter 4: The Control Gain...................................................................................... 24
4.1 Introduction .................................................................................................... 24
4.2 Dynamic absorption in a uniform beam and a formula for
the control gain............................................................................................... 24
4.3 Analysis of results .......................................................................................... 29
4.4 Some illustrations ........................................................................................... 36
4.5 Summary......................................................................................................... 44
Chapter 5: Stability and Optimality........................................................................... 45
5.1 Introduction .................................................................................................... 45
5.2 Stability analysis............................................................................................. 45
5.3 Optimality....................................................................................................... 48
5.4 Summary......................................................................................................... 50
Chapter 6: Experimental Verification........................................................................ 52
6.1 Introduction .................................................................................................... 52
6.2 Proposed model for the experiment................................................................ 52
v
6.3 Determination of natural frequencies ............................................................. 53
6.4 Determination of the control gain................................................................... 64
6.5 Control system design .................................................................................... 66
6.6 Procedure........................................................................................................ 70
6.7 Experimental result......................................................................................... 70
6.8 Validation of the proposed theory .................................................................. 70
6.9 Summary......................................................................................................... 72
Chapter 7: Conclusions and Recommendations........................................................ 73
7.1 Conclusions .................................................................................................... 73
7.2 Recommendations for future research............................................................ 75
References ..................................................................................................................... 77
Appendix A.................................................................................................................... 79
Matlab Programs ........................................................................................................ 79
Appendix B.................................................................................................................. 100
Eigenvalues............................................................................................................... 100
Vita............................................................................................................................... 105

vi
List of Tables
TABLE 4.1: Comparison of control gain formulae........................................................ 35
TABLE 4.2: Comparison of control gain formula for static case .................................. 35
TABLE 6.1: Comparison of natural frequencies............................................................ 63
TABLE 1 (Appendix B): Eigenvalues ..................................................................... 100
TABLE 2 (Appendix B): Eigenvalues for a=0.25 .................................................. 101
TABLE 3 (Appendix B): Eigenvalues for a=0.5..................................................... 102
TABLE 4 (Appendix B): Eigenvalues for a=
2
1 ................................................... 103
TABLE 5 (Appendix B): Eigenvalues for a=2/3..................................................... 104

vii
List of Figures
FIGURE 1.1: A schematic for active control ................................................................... 2
FIGURE 1.2: Vibration control of a harmonically excited beam .................................... 4
FIGURE 2.1: Single-degree-of-freedom dynamic absorber ............................................ 7
FIGURE 3.1: Equation of motion for a beam................................................................ 13
FIGURE 3.2: Uniform cantilever beam subject to harmonic excitation........................ 17
FIGURE 3.3: Steady state amplitude for a uniform cantilever beam............................. 22
FIGURE 4.1: Vibration control of a harmonically excited beam .................................. 25
FIGURE 4.2: Plot of control gain against ............................................................. 29
FIGURE 4.3: Plot of control gain against - (superimposed) ................................. 30
FIGURE 4.4: Clamped and clamped-double-hinged uniform beam.............................. 31
FIGURE 4.5: Static deflection of a clamped-hinged beam............................................ 32
FIGURE 4.6: Illustration demonstrating control gain calculations................................ 37
FIGURE 4.7: Controlled uniform beam......................................................................... 39
FIGURE 4.8: Uncontrolled uniform beam..................................................................... 41
FIGURE 4.9: Controlled uniform beam......................................................................... 42
FIGURE 4.10: Implementation of nodal control............................................................ 43
FIGURE 5.1: Passively controlled uniform beam.......................................................... 46
FIGURE 5.2: Stability analysis and equivalent stiffness ............................................... 47
FIGURE 5.3: Plots of control gain, control force and inverse of static deflection
against the beam span x for cases with excitation frequency of
(a) 10 = , (b) 20 = , (c) 30 = and (d) 80 = ................................ 49
FIGURE 6.1: Mathematical model of the test beam used in the experiment................. 53
FIGURE 6.2: Dimensions of the test beam used in the experiment............................... 57
viii
FIGURE 6.3: Clamping details ...................................................................................... 58
FIGURE 6.4: Impact Hammer ....................................................................................... 59
FIGURE 6.5: Accelerometer .......................................................................................... 59
FIGURE 6.6: Modal Analysis ........................................................................................ 61
FIGURE 6.7: VirtualBench DSA display ...................................................................... 63
FIGURE 6.8: Controlled test beam................................................................................ 64
FIGURE 6.9: Control gain and control force variation along the beam span ................ 65
FIGURE 6.10: Test beam modeshape before and after control ..................................... 66
FIGURE 6.11: Spring housing configuration................................................................. 67
FIGURE 6.12: Attaching the vibration exciter to the beam........................................... 67
FIGURE 6.13: Schematic of the experimental setup ..................................................... 68
FIGURE 6.14: Attaching the springs and the shaker ..................................................... 69
FIGURE 6.15: Experimental setup................................................................................. 69
FIGURE 6.16: Active Vibration Control ....................................................................... 71

ix
List of symbols
) ( , x A A area of cross-section
i i i i
D C B A and , , , constant coefficients for each i
A square matrix
b force vector
control gain
static deflection
) ( , x E E the Youngs modulus of elasticity
p
e p
th
unit vector
) (t f harmonic force
non-dimensional parameter
) ( , x I I moment of inertia
p
k spring stiffnes of the primary system
s
k spring stiffnes of the secondary system
L length of the beam
eigenvalues of clamped-hinged beam
p
m mass of the primary system
s
m mass of the secondary system
) , ( t x M bending moment
eigenvalues of clamped-double-hinged beam
) ( , x density
x
B
S bending stiffness
) (t u control force
) , ( t x V shear force
) (x v shape function
) , ( t x w deflection
frequency of excitation
n
natural frequencies
z displacement vector

xi
Abstract
Vibration control is an important engineering problem and many methods for both
active and passive vibration absorption have been developed. This thesis deals with
developing a method to achieve nodal control at the point of excitation in a Bernoulli-
Euler beam. It is established that, for a uniform Bernoulli-Euler beam, the steady state
motion at the point of excitation can be absorbed by means of a control force
determined from displacement information at the point of application. A closed form
solution for the control gain is presented and a criterion for implementing the control by
active and passive means is developed. The result for the control gain is generalized for
the case of a non-uniform beam. Chapter 4 shows through some examples that the
theory can be also applied to eliminate the steady state motion at any desired location
other than the point of excitation. Analysis is also performed to determine the optimal
control force and investigate the stability of the overall system. Several controllability
graphs are shown and meaningful conclusions are drawn from these graphs.

An experiment is designed to validate the proposed theory and display its practicality.
A uniform steel beam supported at two locations is tested. Modal testing is performed
to extract natural frequencies in order to characterize the system and assist in
formulation of an appropriate mathematical model. The steel beam is then excited by a
known harmonic force supplied by a vibration exciter and a spring, with suitable spring
constant obtained by performing the control gain calculations on the model, is used to
absorb the motion at the free end. It is confirmed, that the theory developed in this
xii
thesis produces accurate results, and that it can serve as a vital tool in developing
practical solutions to structural control problems.

xiii
Chapter 1: Introduction
Most mechanical components are subject to vibrations, which, depending on
circumstances may be desirable or undesirable. On one hand, vibrations of guitar
strings produce wonderful music and on the other hand, vibrations in an automobile
may cause excessive discomfort and fatigue to the driver.

This thesis is concerned with the case of undesirable vibrations and in particular in
developing methods for elimination of steady state response. Components are designed
to withstand definite levels of vibrations. Design, in vibrations, is used to denote an
educated method of choosing and adjusting the physical parameters of a vibrating
system in order to obtain a more favorable response [1]. Modifications of physical
parameters namely mass, damping, and stiffness, in order to improve the vibrational
response of the system fall in the category of passive control. The most common
passive control device is a vibration absorber, which manifests in the form of layers of
damping material added to vibrating structures. Passive control may also involve
changing values of mass and stiffness and hence is also referred as redesign.

Chapter 2 describes a passive control device called single-degree-of-freedom dynamic
absorber. A single-degree-of-freedom dynamic absorber is made up of single mass and
spring and may or may not have a damper. Essentially, the dynamic vibration absorber
introduces additional degree of freedom in the original dynamic system, which results
in a different steady state response. The values of mass, stiffness and damping can be
modified to tune the response of the resulting system to desired levels.
1
However, altering the physical parameters of the system may not always yield a desired
response. In these situations, one has to try implementing active control. Active
control uses external active device, called an actuator, which assists in shaping the
system response. The actuator (e.g. a piezoelectric device, a hydraulic piston, or rack
and pinion) is capable of applying control force to the system under consideration. The
control force is determined based on a mathematical rule, which operates on the system
response measured in realtime by a sensor. The mathematical rule used to apply the
force from the sensor measurement is called the control law.

Dynamic System
Actuator
Sensor
Processor Dynamic System
Actuator
Sensor
Processor

FIGURE 1.1: A schematic for active control

The system comprising both, the actuator and the sensor together with the electronic
circuitry that reads the sensor output and calculates corresponding input to the actuator
is called the control system [5]. Figure 1.1 shows a schematic for implementing active
control.

2
Much work has been done on tuning vibrational response of multi degree of freedom
systems by applying results obtained from the study of dynamic absorbers. A single-
degree-of-freedom dynamic absorber is attached to continuous system like a beam to
control single a mode of vibration under the influence of harmonic excitation. Aida,
Toda, Ogawa and Imada in [11] and Kawazoe, Kono, Aida, Aso and Ebisuda in [12]
discuss a beam type vibration absorber capable of suppressing several vibration modes
of beams.

The subject investigated in this thesis is elimination of steady state vibration at a desired
location in beams by means of nodal control. The developed theory will provide a
strong foundation for realizing realistic and convenient methodologies in control
applications in cases like surgical procedures, drilling and turning operations etc.
However, one of the many direct applications of this method is structural vibration
control in an aircraft wing. Several measurements such as vibrational response, air
temperature, wind velocity etc are required in order to monitor flight conditions in an
aircraft. These data also assist the pilot in flying the aircraft. Sensors and data
collection circuitry form an entire network of the electrical wiring all in and around the
airplane body. Data acquisition devices are also located on the wing of the airplane.
Shielding of these devices from undesirable vibration of the wing is critical in order to
avoid noise in the gathered data and prevent damage to electrical wiring. Exclusion of
steady state vibration at the locations of these devices provides the motivation for this
investigation.

3
Consider a non-uniform Bernoulli-Euler beam of length . Suppose that the beam is
excited by a harmonic force
L
( ) t t f cos = , as shown in Figure 1.2(a).

x
( ) t t f cos =
( ) t x w ,
L
(a) Uncontrolled beam
x
( ) t t f cos =
( ) t x w ,
L
(b) Controlled beam
( ) ( ) t a w t u , =
a
x
( ) t t f cos =
( ) t x w ,
L
x
( ) t t f cos =
( ) t x w ,
L
(b) Controlled beam
( ) ( ) t a w t u , =
a

FIGURE 1.2: Vibration control of a harmonically excited beam
The steady state motion of a prescribed point of the beam may be vanished by applying
a concentrated control force u at ) , ( t a a x = as shown in Figure 1.2(b).

The work here focuses on determining a closed form solution for the control gain
that absorbs the motion of the beam at L x = . Chapter 4 describes the method in detail
and also provides criterion to determine the type of the control i.e. active nodal control
or passive nodal control. It also provides an illustration to show that motion can be
4
absorbed not only at the point of excitation but also at any other point along the beam
span. Chapter 5 discusses the criterion for stability of the controlled system and
optimality in relation to the control force. In order to validate the findings, an
experiment is designed exhibiting nodal control by means of a passive element, a
spring, in a steel beam under the influence of harmonic excitation. It is thus shown that
the approach presented in this thesis is highly practical.

5
Chapter 2: Literature Survey and Background
2.1 Introduction
Some fundamental knowledge and some key results found in the literature survey
related to vibration control are presented here. The survey covers topics related to the
design of both active and passive vibration absorbers.

2.2 Passive vibration control
In 1911 Frahm invented a device for stabilization of rocking oscillations of ships [4].
This device is now known as a dynamic absorber. The dynamic absorber is extremely
simple in principle and has large practical applications. For example Lee, Nian, and
Tarng in [6], Al-Bedoor, Moustafa, and Al-Hussain in [9], Yamashita, Seto, and Hara in
[10] describe design of a dynamic vibration absorber for vibration control in turning
operations, synchronous motor-driven compressors, and piping systems, respectively.

Theory of single-degree-of-freedom dynamic absorber
The dynamic vibration absorber is an additional mass-spring system, which is
appropriately chosen to neutralize the steady state force acting on a particular degree of
freedom. Consider the single-degree-of-freedom system shown in Figure 2.1(a), under
the harmonic excitation of ( ) t F t f sin
0
= . Let this system be called the primary
system. Upon a harmonic excitation, the system vibrates with two frequencies, the
frequency of excitation , and the natural frequency of the system,
6
,
p p n
m k =
(2.2.1)

where subscript p denotes the parameters associated with the primary system. The
objective is to eliminate the forced component of vibrations. This is implemented by
attaching additional single-degree-of-freedom mass-spring system, which is called the
secondary system, with the mass and the spring stiffness , to the primary system.
The global system is shown in Figure 2.1(b).
s
m
s
k

(a) Primary system
(b) Global system
( ) t x
p
( ) t F t f sin
0
=
p
m
p
k
( ) t x
s
s
k
s
m
( ) t x
p
( ) t F t f sin
0
=
p
m
p
k
(a) Primary system
(b) Global system
( ) t x
p
( ) t F t f sin
0
=
p
m
p
k
( ) t x
p
( ) t F t f sin
0
=
p
m
p
k
( ) t x
s
s
k
s
m
( ) t x
p
( ) t F t f sin
0
=
p
m
p
k

FIGURE 2.1: Single-degree-of-freedom-dynamic absorber

7
The mathematical model of the global system has the form
, sin
0 0
0
0
t
F
x
x
k k
k k k
x
x
m
m
s
p
s s
s s p
s
p
s
p
|
|
.
|
\
|
=
|
|
.
|
\
|
(
+
+
|
|
.
|
\
|
(
& &
& &

(2.2.2)
which can also be written as
( )
= +
= + +
. 0
sin
0
s s p s s s
s s p s p p p
x k x k x m
t F x k x k k x m
& &
& &

(2.2.3)
Both masses vibrate with forced harmonic vibrations of the form
( )
( )
=
=
, sin
sin
t X t x
t X t x
s s
p p

(2.2.4)
where the constants , are the amplitudes of the forced component of vibration.
Substituting (2.2.4) in (2.2.3) gives
s p
X X ,
( )
= +
= + +
. 0
2
0
2
s s p s s s
s s p s p p p
X k X k X m
F X k X k k X m

(2.2.5)
Since is to be eliminated, substituting
p
X 0 =
p
X in the above set of equations yields
= +
=
. 0
2
0
s s s s
s s
X k X m
F X k

(2.2.6)
The second equation in (2.2.6), gives
,
2
s
s
m
k
=
(2.2.7)
and the first equation in (2.2.6) provides the amplitude of vibration of ,
s
m
.
0
s
s
k
F
X =
(2.2.8)
Hence, it is concluded that the vibratory motion of the primary mass can be eliminated
provided that the stiffness and the mass values for the secondary mass are chosen such
8
that they satisfy (2.2.7). In this case the secondary mass vibrates out of phase to the
external harmonic excitation and the spring force exactly contradicts the harmonic
force, causing the forced motion of to vanish.
p
m

This idea can also be extended to a multi-degree-of-freedom system. Ram and Elhay in
[8], show that a multi-degree-of-freedom dynamic absorber may absorb the steady state
motion associated with the harmonic excitation of several frequencies. There are,
however, some limitations associated with practical implementation of the dynamic
absorber. Firstly, it is not always feasible to attach the absorber to the specific degree of
freedom of which the motion is to be absorbed. Secondly, application of dynamic
absorber increases the dimension of the system, and hence introduces new natural
frequencies that may interfere with other excitations. Thirdly, the theory of dynamic
absorbers for damped systems is not fully developed. It is, therefore, not clear how the
dynamic absorption phenomenon may be used in eliminating the steady state motion of
a damped system that is excited by a harmonic force [7]. Herzog in [16] investigates
the topic of performance degradations of dynamic systems implementing passivity-
based control. He has analyzed the topic in terms of flatness of response of the
controlled system in the vicinity of the natural frequency of the dynamic absorber.

2.3 Active vibration control
As described earlier in Chapter 1, in some cases implementing the active vibration
absorber is imperative. Nishimura, Yoshida and Shimogo in [17] have studied optimal
design method of the active dynamic vibration absorber for multi-degree-of-freedom
9
systems. The method was also validated by performing numerical simulations and an
experiment on a 3-degree-of-freedom building like structure. Aida, Toda, Ogawa and
Imada in [11] and Kawazoe, Kono, Aida, Aso and Ebisuda in [12] demonstrate control
of several mode shapes of a beam by using a beam type vibration absorber with
boundary conditions same as the main beam. It is shown that for specific vibration
modes, mode equations of a beam with beam-type dynamic absorber are approximate
equivalents of the motion of two-degree-of-freedom system. Hence, the dynamic
absorber system can be tuned by Den Hartog method.

Gaudreault, Liebst, Bagley in [18] present four techniques for combining active
vibrational control and passive viscous damping. The motivation behind the work is
some findings, which reveal that the passive damping can reduce the amount of active
damping needed to control structural vibrations. However, inappropriate design of the
passive damping can produce contrary results in that it may increase system reaction
times, reducing control effectiveness.

The partial-pole assignment problem is addressed by Ram in [15]. The paper
determines the force required to place a few poles of the spectrum while leaving the rest
unchanged and the conditions under which the solution is unique. The work of Ram in
[3] lays the foundation for this thesis. The paper provides a closed form control gain
solution for absorbing the harmonic response at a desired location in an axially
vibrating rod and analyzes the stability of the controlled system.
10
Chapter 3: Beam: Theory and Background
3.1 Introduction
In order to determine the dynamic behavior of mechanical systems, one needs to
develop an appropriate mathematical model. Mechanical systems can be modeled as
lumped-parameter systems, where it is assumed that the motion of the system is
governed by the mass of the system concentrated at a particular point. The modeled
system is known as a lumped parameter system, which has finite number of lumped
masses connected to each other by means of springs and dampers. Even though a
discrete model provides an acceptable solution to the system, it is not capable of
accounting for the flexibility of various structures. Engineering problems such as
swaying of tall buildings, torsional and bending vibrations of shafts and vibrations in
wings of aircraft demand an insight into elastic behavior of structural members. These
elastic systems consist of continuous mass and elasticity throughout their span. Hence,
these systems are modeled assuming that the mass of the system is distributed in the
entire system as infinitesimally small elements. Such a model for a mechanical system
is known as a distributed parameter model. There are only few distributed parameter
systems such as beams, bars, strings and plates that have closed form solutions.
However, study of these systems provides understanding of behavior and modeling of
most complex systems, which cannot be solved in a closed form manner.

This chapter deals with study of vibration of continuous beams. The equation of motion
for a beam is described and a closed form solution is provided. Investigation is done in
terms of natural frequency and steady state response by considering the case of a
11
cantilever beam subjected to dynamic excitations. This chapter forms the foundation to
the problem investigated in the thesis. The source of most of the material presented in
Sections 3.2 and 3.3 is [2].

3.2 Equation of motion for a non-uniform beam
Consider a non-uniform Bernoulli-Euler beam of length as shown in Figure 3.1(a).
The transverse vibrations are denoted as
L
( ) t x w , . The cross-sectional area is ( ) x A ,
modulus of elasticity is ( ) x E , density is ( ) x , and moment of inertia is ( ) x I . The
external force applied to the beam per unit length is denoted by . ( ) t x f ,

From strength of materials, the bending moment ( ) t x M , is related to the beam
deflection by ( t x w , )
( ) ( ) ( )
( )
2
2
,
,
x
t x w
x I x E t x M
= . (3.2.1)

One can look on an infinitesimal element of the beam, shown in the Figure 3.1(b), and
determine the model of flexural vibrations. It is assumed that the deformation is small
enough such that the shear deformation is much smaller than . From Newtons
second law in the - direction,
( t x w , )
y
( )
( )
( ) ( ) ( ) ( )
( )
2
2
,
, ,
,
,
t
t x w
dx x A x dx t x f t x V dx
x
t x V
t x V
= +
|
.
|
\
|
+ . (3.2.2)

12

( ) t x f ,
( ) t x w ,
dx x
L
x
y
( ) t x w ,
( ) t x f ,
Undeformed x-axis
x
dx x +
dx
( ) t x M ,
( )
( )
dx
x
t x M
t x M
+
,
,
( ) t x V ,
( )
( )
dx
x
t x V
t x V
+
,
,
(a)
(b)
( ) t x f ,
( ) t x w ,
dx x
L
x
y
( ) t x f ,
( ) t x w ,
dx x
L
x
y
( ) t x w ,
( ) t x f ,
Undeformed x-axis
x
dx x +
dx
( ) t x M ,
( )
( )
dx
x
t x M
t x M
+
,
,
( ) t x V ,
( )
( )
dx
x
t x V
t x V
+
,
,
(a)
(b)

FIGURE 3.1: Equation of motion for a beam
13
Here is the shear force at the left end of the element , ( t x V , ) dx ( ) ( )dx t x V t x V
x
, , + is
the shear force at the right end of the element . The term on the right hand side of
the equality sign is the inertia force of the element. The sum of the moments on the
element yields
dx
( )
( )
( ) ( )
( )
( ) | | 0
2
,
,
, ,
,
, = +
(
+ +
(
+
dx
dx t x f dx dx
x
t x V
t x V t x M dx
x
t x M
t x M . (3.2.3)

Here the right hand side in the equation vanishes, since it is assumed that the rotary
inertia of the element is negligible. Simplification of this expression yields, dx
( )
( )
( ) ( )
( ) 0
2
, ,
,
,
2
=
(
+
(
dx
t x f
dx
x
t x V
dx t x V dx
x
t x M
.
(3.2.4)

Since is small, and hence dx ( )
2
dx is negligible. The above expression takes the form
( )
( )
x
t x M
t x V
=
,
, . (3.2.5)

This expression relates shear force and the bending moment. Substituting (3.2.5) in
(3.2.2) gives
( ) | | ( ) ( ) ( )
( )
2
2
2
2
,
, ,
t
t x w
dx x A x dx t x f dx t x M
x
= +
. (3.2.6)

Substituting (3.2.1) in (3.2.6) and dividing by yields dx
( ) ( )
( )
( ) ( )
( )
( ) t x f
x
t x w
x I x E
x t
t x w
x A x ,
, ,
2
2
2
2
2
2
=
(
.
(3.2.7)

If no external force is applied ( ) t x f , is zero. The equation of motion of beam for
is then given as 0 , 0 > < < t L x
( ) ( )
( )
( ) ( )
( )
0
, ,
2
2
2
2
2
2
=
(
x
t x w
x I x E
x t
t x w
x A x .
(3.2.8)

14
The above expression (3.2.8) is a fourth order partial differential equation, which
governs the vibration of a non-uniform Bernoulli-Euler beam. If the parameters ,
, and
) (x E
) (x A ) (x I ) (x are constant then (3.2.8) can be further simplified to give
( ) ( )
, 0
, ,
4
4
2
2
2
=
x
t x w
c
t
t x w
(3.2.9)

where
.
A
EI
c
= (3.2.10)

3.3 Natural frequencies and modeshapes
The equation of motion (3.2.9) contains four spatial derivatives and two time
derivatives. Hence, in order to determine a unique solution for , four boundary
conditions and two initial conditions are needed. Usually, the values of displacement
and velocity are specified at time
) , ( t x w
0 = t , and so the initial conditions can be given as,
0 ) 0 , ( and , 0 ) 0 , ( = = x w x w & ,
(3.3.1)

where dots denote derivates with respect to time. The method of separation of variables
is used to determine the free vibration solution,
) ( ) ( ) , ( t T x V t x w = .
(3.3.2)

Substituting (3.3.2) in (3.2.9) and rearranging yields
( ) ( )
.
) (
1
) (
2
2
2
4
4 2
= =
dt
t T d
t T dx
x V d
x V
c
(3.3.3)

Here, is a positive constant. The above equation can now be written as two
ordinary differential equations as shown below.
2
15
( )
, 0 ) (
4
4
4
= x V
dx
x V d
(3.3.4)

( )
, 0 ) (
2
2
2
= t T
dt
t T d
(3.3.5)

where,
EI
A
c
2
2
2
4

= = . (3.3.6)

The solution to (3.3.5) can be given as
t B t A t T sin cos ) ( + = .
(3.3.7)

The constants A and B can be evaluated using two initial conditions given by (3.3.1),
and the solution to (3.3.4) is assumed as,
sx
e x V = ) ( . (3.3.8)

Substituting (3.3.8) in (3.3.4) and simplifying furnishes,
0
4 4
= s . (3.3.9)

The roots of (3.3.9) are
i s s = =
4 , 3 2 , 1
, ,
(3.3.10)

hence, the solution to (3.3.4) can be given as
x e x e x e e x V
x i x i x x
+ + + =
4 3 2 1
) ( . (3.3.11)

Equation (3.3.11) can also be expressed alternatively as
x x x x x V cosh sinh cos sin ) (
4 3 2 1
+ + + = ,
(3.3.12)

where are different constants, which can be evaluated using the four
boundary conditions. The natural frequency of the beam can be therefore computed
from (3.3.6) as
4 3 2 1
and , ,
16
( )
4
2
4
2 2
AL
EI
Al
EI
L
A
EI
= = = ,
(3.3.13)

where, is a non-dimensional parameter.

3.4 Steady state response and natural frequencies for a
uniform clamped cantilever beam

The dynamic behavior of a beam can be determined by analyzing the case of a uniform
clamped cantilever beam shown in Figure 3.2.

x
( ) t t f sin =
( ) t b w ,
L
a
x
( ) t t f sin =
( ) t b w ,
L
a

FIGURE 3.2: Uniform cantilever beam subject to harmonic excitation
The boundary conditions at the clamped end are no displacement and no slope, which
can be given as,
( ) 0 , 0 = t w , (3.4.1)

and,
( )
0
, 0
=
x
t w
, (3.4.2)

respectively. The boundary conditions at the free end are no bending moment and no
shear force, represented by
17
( ) ( )
( )
0
,
2
2
=
x
t L w
x I x E , (3.4.3)

and,
( ) ( )
( )
0
,
2
2
=
(
x
t L w
x I x E
x
.
(3.4.4)

The steady state response ( ) t b w , , measured at b x = , of the beam under the influence of
harmonic excitation of ( ) t t f sin = at some other position a x = , is described by a
Green function. The Green function is a function of
Frequency of excitation
Position of excitation a x = , and
Position of interest where the response is to be measured . b x =
The beam is separated in two parts, a x 0 , and a L x < , and denoted by
( )
( )
( )
< <
< <
=
. , ,
0 , ,
,
2
1
L x a t x w
a x t x w
t x w
(3.4.5)

The separated beam can then be described as
0
2
1
2
4
1
4
=
t
w
A
x
w
EI , 0 t , 0 > < < a x . (3.4.6)

0
2
2
2
4
2
4
=
t
w
A
x
w
EI , 0 t , > < < L x a . (3.4.7)

The boundary conditions in (3.4.1) through (3.4.4) can be written as
( ) 0 , 0
1
= t w ,
( )
0
, 0
1
=
x
t w
,
( )
0
,
2
2
2
=
x
t L w
, and
( )
0
,
3
2
3
=
x
t L w
EI . (3.4.8)

18
At a x = , the deflection, the slope and the moment are the same for both parts of the
beam. The shear force differs by EI 1 . These four conditions represent the matching
conditions at a x = and can be described by
( ) ( ) t a w t a w , ,
2 1
= , (3.4.9)

( ) ( )
x
t a w
x
t a w
, ,
2 1
, (3.4.10)
( ) ( )
2
2
2
2
1
2
, ,
x
t a w
x
t a w
, (3.4.11)
and
( ) ( )
EI x
t a w
x
t a w 1 , ,
3
2
3
3
1
3
=
. (3.4.12)

Separation of variables gives
( ) ( ) t x v t x w
i i
sin , = , 2 , 1 = i .
(3.4.13)

Substitution of (3.4.13) in (3.4.6) and (3.4.7) yields
0
1
2
1
= v A v EI , a x < < 0 , (3.4.14)

0
2
2
2
= v A v EI , L x a < < , (3.4.15)

with the boundary conditions
( ) ( ) ( ) 0 0 0
2 1 1
= = = L v v v , ( ) 0
2
= L v , (3.4.16)

( ) ( ) ( ) 0 0 0
2 1 1
= = = L v v v , ( ) 0
2
= L v , (3.4.17)

and the matching conditions
( ) ( ) a v a v
2 1
= , ( ) ( ) a v a v
2 1
= , ( ) ( ) a v a v
2 1
= , ( ) ( )
EI
a v a v
1
2 1
= . (3.4.18)

19
The equations (3.4.14) and (3.4.15) can be written as
0
1
4
1
= v v , a x < < 0 , (3.4.19)
and
0
2
4
2
= v v , L x a < < , (3.4.20)
where, again
EI
A

2 4
= . (3.4.21)
The general solution of (3.4.19), (3.4.20) is given by
( ) x D x C x B x A x v
i i i i i
cosh sinh cos sin + + + = , 2 , 1 = i .
(3.4.22)

The four boundary conditions and the four matching conditions can be used to evaluate
the eight constants , 2
i i i i
D C B A and , , , , 1 = i . The matrix representation of the eight
equations obtained from the eight conditions is
b Az = , (3.4.23)

where,
= A
(
(
(
(
(
(
(
(
(
(
(

a sh a ch a s a c a sh a ch a s a c
a ch a sh a c a s a ch a sh a c a s
L sh L ch L s L c
L ch L sh L c L s

3 3 3 3 3 3 3 3
2 2 2 2 2 2 2 2
3 3 3 3
2 2 2 2
0 0 0 0
0 0 0 0
0 0 0 0 0 0
0 0 0 0 1 0 1 0
,
| |
T
D C B A D C B A
2 2 2 2 1 1 1 1
= z ,
and
8
e b
EI
1
= .
20
The determinant of is determined for different values of A . The values of which
make singular are designated as A
n
. Then the natural frequencies of the beam
n

are
( )
4
2
4
2 2
AL
EI
AL
EI
L
A
EI
n n n
= = = , (3.4.24)

where is a dimensionless parameter. The fundamental natural frequency of a
cantilever beam leads to 875 . 1
1
= . Denoting EI , the bending stiffness of the beam,
by , the first natural frequency for the cantilever beam can be expressed as,
B
S
4 1
5156 . 3
AL
S
B
= . (3.4.25)

Now,
z A b
1
= (3.4.26)

allows determination of the values for constants for i .
i i i i
D C B A and , , , , 2 , 1 =
The steady state amplitude at any point other than the point of excitation is given by the
Green function as below
( )
( )
( )
< <
< <
=
. ,
0 ,
, ,
2
1
L b a b v
a b b v
b a G
(3.4.27)

Figure 3.3 shows plots of the steady state amplitude against , for the following four
cases:
(a) , 1 and , 1 = = b a
(b) , 1 and , 5 . 0 = = b a
(c) , 5 . 0 and , 5 . 0 = = b a
(d) , 25 . 0 and , 75 . 0 = = b a
21
where a is the point of harmonic excitation and b is the point of measurement of the
steady state amplitude. The system parameters and , , , L I A E can be chosen
arbitrarily. The figure shows 875 . 1
1
= , 694 . 4
2
= , 854 . 7
3
= , and 995 . 10
4
= .
Hence, for known values of and , , , L I A E the first four natural frequencies of the
cantilever beam can be given by (3.4.24).
(e)
7.85 7.855 7.86
0
) (b w
0 2 4 6 8 10 12
0
) (b w
0 2 4 6 8 10 12
0
) (b w
(b)
(d)
0 2 4 6 8 10 12
0
) (b w
0 2 4 6 8 10 12
0
) (b w
(a)
(c)
(e)
7.85 7.855 7.86
0
) (b w
0 2 4 6 8 10 12
0
) (b w
0 2 4 6 8 10 12
0
) (b w
0 2 4 6 8 10 12
0
) (b w
(b)
(d)
0 2 4 6 8 10 12
0
) (b w
0 2 4 6 8 10 12
0
) (b w
(a)
(c)

FIGURE 3.3: Steady state amplitude for a uniform cantilever beam
22
However, the cantilever beam, being a distributed parameter system, has infinite
number of natural frequencies, which can be approximated by the formula
( )
2
1 2 n
for
[5]. 5 > n

Cases (a) and (c) represent the response at a collocated point, i.e., the point of excitation
and the point of measurement are the same. In cases (b) and (d) the response is at a
non-collocated point, i.e., the point of excitation and the point of measurement are
different.

In Figure 3.3(c) the pole in the neighborhood of 854 . 7
3
= is not observed. This is
because the pole and the zero are very close to each other. A magnified view of the
region marked by circle in Figure 3.3(c) is shown in Figure 3.3(e). Here, it is clear that
the pole and the zero are extremely close to each other.

3.5 Summary
One popular approach in controls is pole-zero cancellation. Here, the idea is to place
some zeros on some poles to reduce vibration in the rod. However, a major drawback
of the method is that one has to be extremely careful in placing the zeros, because the
slightest error in positioning of a zero can lead to a non-vanishing pole, which will
make the system unstable.

23
Chapter 4: The Control Gain
4.1 Introduction
This chapter deals with development of the theory for elimination of steady state
response at a prescribed location for the case of a harmonically excited Bernoulli-Euler
cantilever beam. A closed form expression for the control gain is established for the
uniform cantilever beam and the results are then generalized to the case of a non-
uniform beam. Several graphs indicating the control gain requirement with the change
in excitation frequency are shown. Investigation is performed on the obtained results to
draw meaningful conclusions and provide groundwork for further development of the
proposed theory.

4.2 Dynamic absorption in a uniform beam and a formula
for the control gain

Consider a uniform beam of length , modulus of elasticity L E , density , cross-
sectional area , and moment of inertia A I . The axial distance from the supported end
of the beam is x . Suppose that the beam is excited by a harmonic force ( ) t t f cos = ,
as shown in Figure 4.1(a).

24

FIGURE 4.1: Vibration control of a harmonically excited beam
x
( ) t t f cos =
( ) t x w ,
L
x
( ) t t f cos =
( ) t x w ,
L
(b) Controlled beam
( ) ( ) t a w t u , =
a
x
( ) t t f cos =
( ) t x w ,
L
x
( ) t t f cos =
( ) t x w ,
L
(b) Controlled beam
( ) ( ) t a w t u , =
a
Then as shown in Chapter 3, the steady state transverse vibrations of the beam are
governed by the differential equation
0
2
2
4
4
=
t
w
A
x
w
EI , L x < < 0 , . 0 > t (4.2.1)
and the boundary conditions are
( ) 0 , 0 = t w , (4.2.2)
( )
0
, 0
=
x
t w
, (4.2.3)
( )
0
,
2
2
=
x
t L w
, (4.2.4)
25
and
( )
t
x
t L w
EI cos
,
3
3
=
. (4.2.5)
The conditions (4.2.2) and (4.2.3) describe no-displacement and no-slope at 0 = x ,
condition (4.2.4) imposes no bending moment at L x = and condition (4.2.5) shows that
shear induced by the exciting force is t cos . The steady state motion of a prescribed
point of the beam may be vanished by applying a concentrated control force at a x = ,
( ) ( ) t a w t a u , , = , (4.2.6)
where is a constant. The partial differential equation for the controlled system is then
( ) w a x
t
w
A
x
w
EI =
2
2
4
4
, L x < < 0 , , 0 > t (4.2.7)
with the same boundary conditions as in (4.2.2) through (4.2.5) and ) ( a x is dirac
function. Using the method of separation of variables described in Chapter 3, the steady
state response of the system shown in (4.2.7) is given by
( ) ( ) ( ) t x v t x w cos , = , (4.2.8)
The equations (4.2.7) and (4.2.2) through (4.2.5) can be reduced to set of ordinary
differential equations
( )v a x v v =
4
, (4.2.9)
( ) ( ) ( ) 0 0 0 = = = L v v v , ( ) 1 = L v EI , (4.2.10)
where as before EI A
2 4
= , and primes denote derivatives with respect to x . In
order to eliminate the inhomogeneity of (4.2.9), the beam is separated into two parts,
, and , denoted as a x 0 L x < a
26
( )
( )
( )
< <
< <
=
. ,
0 ,
2
1
L x a x v
a x x v
x v
(4.2.11)
Then (4.2.9) and (4.2.10) can be written in an equivalent form
0
1
4
1
= v v , a x < < 0 , (4.2.12)
0
2
4
2
= v v , L x a < < , (4.2.13)
( ) ( ) ( ) 0 0 0
2 1 1
= = = L v v v , ( ) 1
2
= L v EI , (4.2.14)
( ) ( ) a v a v
2 1
= , ( ) ( ) a v a v
2 1
= , ( ) ( ) a v a v
2 1
= , (4.2.15)
and
( ) ( ) ( ) a v a v EI a v EI
1 2 1
= . (4.2.15)
The conditions in (4.2.15) describe the matching conditions at a x = . The first three
conditions describe the continuity in terms of displacement, slope and bending moment,
and the fourth condition shows that the difference in the shear force on either side of
control force differs by the magnitude of the control force.

The general solution of (4.2.12) and (4.2.13) is given by
i i i i i
cosh sinh cos sin + + + = , 2 , 1 = i .
(4.2.16)
If is known then the eight constants , 2
i i i i
D C B A , , , , 1 = i , can be determined from
the eight conditions in (4.2.14) and (4.2.15). Here is unknown. However, the ninth
condition is generated from the desired objective, i.e. no steady state displacement at the
free end, mathematically described as
( ) 0 , = t L w , (4.2.17)
or equivalently,
27
( ) 0
2
= L v . (4.2.18)
Hence, the four conditions in (4.2.14), the first three conditions of (4.2.15), and the
condition (4.2.18), can be written in matrix form
b Az = , (4.2.19)
where
= A
(
(
(
(
(
(
(
(
(
(
(

L sh L ch L s L c
L ch L sh L c L s
L ch L sh L c L s

3 3 3 3
2 2 2 2
2 2 2 2 2 2 2 2
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0 0 0
0 0 0 0 1 0 1 0
,
| |
T
D C B A D C B A
2 2 2 2 1 1 1 1
= z , (4.2.20)
8
1
e b
EI
= , (4.2.21)
where is the p
p
e
th
unit vector, and s, c, sh, and ch represent sine, cosine, hyperbolic
sine and hyperbolic cosine, respectively. Once is found the shape functions z ( ) x
i
v ,
, are determined by (4.2.16). The fourth condition expressed by (4.2.15) can then
be resolved for
2 , 1 = i
, i.e.
( )
( ) ( )
( )
|
|
.
|
\
|
=
a v
a v a v
EI a
1
2 1
, .
(4.2.22)
The values of the control gain for various points of excitation a ,
as function of the exciting frequency
8 . 0 , 6 . 0 , 4 . 0 , 2 . 0 =
are shown in Figures 4.2(a)-4.2(d).
L A I E and , , , can be chosen arbitrarily.

28
(b)
(a)
0 2 4 6 8 10 12
-8
-4
0
4
8
0 2 4 6 8 10 12
-8
-4
0
4
8
(c) (d)
0 2 4 6 8 10 12
-8
-4
0
4
8
0 2 4 6 8 10 12
-8
-4
0
4
8
3
10
3
10
3
10
3
10
(b)
(a)
0 2 4 6 8 10 12
-8
-4
0
4
8
0 2 4 6 8 10 12
-8
-4
0
4
8
(c) (d)
0 2 4 6 8 10 12
-8
-4
0
4
8
0 2 4 6 8 10 12
-8
-4
0
4
8
(b)
(a)
0 2 4 6 8 10 12
-8
-4
0
4
8
0 2 4 6 8 10 12
-8
-4
0
4
8
(c) (d)
0 2 4 6 8 10 12
-8
-4
0
4
8
0 2 4 6 8 10 12
-8
-4
0
4
8
3
10
3
10
3
10
3
10

FIGURE 4.2: Plot of control gain against
4.3 Analysis of results
Large amount of information can be derived from the above graphical results. The plots
for are singular at certain values of . Also, the positions of these singularities
change with the change in control location a x = . However, Figure 4.3 shows the
above four graphs superimposed on each other. It can be seen from these graphs that
the zeros of ( ) , a are invariant of the control location. Further analysis is performed
in order to understand this behavior.
29

0 2 4 6 8 10 12
-8000
-4000
0
4000
8000
0 2 4 6 8 10 12
-8000
-4000
0
4000
8000
0 2 4 6 8 10 12
-8000
-4000
0
4000
8000

FIGURE 4.3: Plot of control gain against - (superimposed)
Consider the clamped-hinged uniform beam shown in Figure 4.4(a). The natural
frequencies of this beam are indeed the zeros of ( ) , a . Also, the singularities in
( ) , a are the natural frequencies of the beam shown in Figure 4.4(b) which is
clamped in the left end and hinged both at a x = and L x = .

From the above arguments and from [19] and [3], the control gain ) , ( a can also be
determined from the following formula
( )
=
|
|
.
|
\
|
|
|
.
|
\
|
=
1
2
1
2
1
1
,
i i
i i
c a
,
(4.3.1)

30

(a) Clamped-hinged uniform beam
L
(b) Clamped-double- hinged uniform beam
L
a
(a) Clamped-hinged uniform beam
L
(b) Clamped-double- hinged uniform beam
L
a
L
a

FIGURE 4.4: Clamped and clamped-double-hinged uniform beam
where is a constant, c
i
denotes the eigenvalues of the clamped-hinged beam shown
in Figure 4.4(a) and
i
denotes the eigenvalues of the clamped-double-hinged beam in
Figure 4.4(b). Constant c can be evaluated by considering the static case i.e. when
0 = , applied to the case shown in Figure 4.4(a). Substituting 0 = in (4.3.1) gives
( ) c a = 0 , . (4.3.2)
Denote by the static deflection at a x = of the clamped-hinged beam shown in
Figure 4.5(a) due to a collocated unit static load 1 = U . This is a static problem where
the deflection of any point on the beam is no longer a function of time , and
can be denoted by . Hence the mathematical model of the beam shown in Figure
4.5(a) can be described by
) , ( t x w
( w
t
) x
31
L x U a x
dx
w d
< < = 0 , ) (
4
4
(4.3.3)
and the boundary conditions are

(a)
(b)
L
F
F
L
1 = U
a
a
(a)
(b)
L
F
F
L
1 = U
aa
aa

FIGURE 4.5: Static deflection of a clamped-hinged beam
0 ) 0 ( = w ,
(4.3.4)
0
) 0 (
=
dx
dw
, (4.3.5)
0 ) ( = L w ,
(4.3.6)
and
32
0
) (
2
2
=
dx
L w d
. (4.3.7)
The beam is separated in two parts as described below.
< < =
< < =
L x a
dx
w d
a x
dx
w d
, 0
0 , 0
4
2
4
4
1
4
.
(4.3.8)
The boundary conditions are,
( ) ( ) ( ) ( ) 0 and 0 , 0 0 , 0 0
2 2 1 1
w L w w w = = = . (4.3.9)
and the matching conditions are
( ) ( ) ( ) ( ) a w a w a w a w
2 1 2 1
, = = ,
( ) ( ) ( ) ( )
EI
a w a w a w a w
1
and
2 1 2 1
= = .
(4.3.10)
The solution for (4.3.3) can be given by
2 1 , ) (
2 3
, i D x C x B x A x w
i i i i i
= + + + = . (4.3.11)
The arbitrary constants 2 , 1 for , , , , = i D C B A
i i i i
can be evaluated by solving the eight
equations derived by imposing the eight conditions together with (4.3.9) and (4.3.10) on
(4.3.11). The equations can be represented in the matrix
b Az = , (4.3.12)
where
(
(
(
(
(
(
(

=
0 0 2 6 0 0 0 0
1 0 0 0 0
0 0 0 6 0 0 0 6
0 0 2 6 0 0 2 6
0 1 2 3 0 1 2 3
1 1
0 0 0 0 0 1 0 0
0 0 0 0 1 0 0 0
2 3
2 2
2 3 2 3
L
L L L
a a
a a a a
a a a a a a
A ,
(4.3.13)
33
| |
T
D C B A D C B A
2 2 2 2 1 1 1 1
= z , (4.3.14)
6
1
e b
EI
= . (4.3.15)
For each value of the constants can be evaluated from . Then the static
deflection
a b A z
1
=
can be given as
2 , 1 ,
2 3
= + + + = i D a C a B a A
i i i i
. (4.3.16)
By the linearity of the problem (4.3.3), the static deflection due to a force applied to
the beam shown in Figure 4.5(b) is
F
F . Hence, invoking the control law (4.2.6) gives
( ) F a F 0 , = , (4.3.17)
which yields
1
= c , (4.3.18)
by virtue of (4.3.2). The control gain is therefore
( )
=
|
|
.
|
\
|
|
|
.
|
\
|
=
1
2
1
2
1
1
,
i i
i i
a
.
(4.3.19)
Note that the control gain given by (4.3.19) is identical to that expressed in (4.2.22).
Table 4.1 shows the comparison between the values of control gain as determined
from (4.3.19), using different numbers of eigenvalues and , and that determined
from (4.2.22), for . 10 =

34
TABLE 4.1: Comparison of control gain formulae
64.76082 71.97431 64.52690 245.73544 7
67.24491 74.81875 66.70656 254.08160 1
64.92478 72.14979 64.68074 246.36542 3
64.74890 71.97368 64.52439 245.69129
as determined
from (4.2.22)
64.78035 72.01305 64.55730 245.82132 5
a=2/3 a= a=0.5 a=0.25
as determined from (4.3.1)
No. of
eigenvalues
64.76082 71.97431 64.52690 245.73544 7
67.24491 74.81875 66.70656 254.08160 1
64.92478 72.14979 64.68074 246.36542 3
64.74890 71.97368 64.52439 245.69129
as determined
from (4.2.22)
64.78035 72.01305 64.55730 245.82132 5
a=2/3 a= a=0.5 a=0.25
as determined from (4.3.1)
No. of
eigenvalues

( )
5 . 0
5 . 0

It is evident that with increase in number of eigenvalues used to calculate from
(4.3.19) the comparison converges to the exact value determined from (4.2.22). Again,
the accuracy of the numerical value of from (4.3.19) is largely influenced by the
accuracy in the numerical value of eigenvalues and . In calculations for in
Table 4.1 the eigenvalues and have been determined by solving the transcendental
eigenvalue problem as shown in [13]. The eigenvalues are listed in Appendix B. This
argument can be further strengthened by comparing values of for the static case
calculated at . 25 . 0 = a
TABLE 4.2: Comparison of control gain formula for static case
364.0889 364.0927
as determined from (4.3.1) as determined from (4.2.22)
364.0889 364.0927
as determined from (4.3.1) as determined from (4.2.22)

35
Here, 364.0889 is the exact value for . The inaccuracy in the value determined from
(4.2.22) is due to singularity of in (4.2.19) for A 0 = . Hence, 364.0927 is the value
of control gain for 00001 . 0 = . Also, it is important to note that the boundary
condition for the beam shown in Figure 4.5(a) is clamped-hinged. This is because the
desired objective is to achieve no steady state displacement at the free end in the beam
shown in Figure 4.1(a).

4.4 Some illustrations
Some simple examples depicting the calculation procedure are presented for better
understanding of the theory developed in the previous sections. Example 1 shows the
stepwise numerical calculations of the control gain, the control force and the steady
state displacement at some positions of interest. Example 2 demonstrates that control
can be achieved in order to eliminate the steady state motion at a point other than the
free end by making a small modification to the theory. Simulated mode shapes are
shown in example 3, which display beam mode shapes before and after implementation
of the nodal control.

Example 1:
A uniform Bernoulli-Euler beam is described by
0 3 5
2
2
4
4
=
t
w
x
w
, , 2 0 < < x , t , 0 > (4.4.1)
( ) 0 , 0 = t w ,
( )
0
, 0
=
x
t w
,
( )
0
, 2
2
2
=
x
t w
, and
( )
t
x
t w
10 sin 2
, 2
3
3
= 5
(4.4.2)
Hence, EI = bending stiffness = 5, and A = 3. Amplitude of harmonic force = 2.
36

x
( ) t t f 10 sin 2 =
( ) t x w ,
L
x
( ) t t f 10 sin 2 =
( ) t x w ,
L

FIGURE 4.6: Illustration demonstrating control gain calculations
Determine control gain and control force needed to eliminate steady state displacement
at the free end.
From separation of variables
( ) ( ) t x v t x w 10 sin , = , (4.4.3)
and,
60 5 / 3 10 /
2 2 4
= = = EI A . (4.4.4)
Hence, (4.4.1) gives
0 60 = v v , (4.4.5)
and, (4.4.2) gives
( ) ( ) ( ) 0 2 0 0 = = = v v v , ( ) 5 2 2 = v . (4.4.6)
The beam is separated into two parts, 0 75 . 0 x , and 2 75 . 0 < x , denoted as
( )
( )
( )
< <
< <
=
. 2 75 . 0 ,
75 . 0 0 ,
2
1
x x v
x x v
x v
(4.4.7)
Hence,
0 60
1 1
= v v , 75 . 0 0 < < x , (4.4.8)
0 60
2 2
= v v , 2 75 . 0 < < x , (4.4.9)
37
Boundary condition are
( ) ( ) ( ) 0 2 0 0
2 1 1
= = = v v v , ( ) 5 2 2
2
= v , (4.4.10)
and
( ) ( ) 75 . 0 75 . 0
2 1
v v = , ( ) ( ) 75 . 0 75 . 0
2 1
v v = , ( ) ( ) 75 . 0 75 . 0
2 1
v v = , (4.4.11)
and
( ) ( ) ( ) ( ) 75 . 0 75 . 0 75 . 0 5
1 2 1
v v v = . (4.4.12)
The conditions in (4.4.11) and (4.4.12) are the matching conditions at . The
general solution of (4.4.8) and (4.4.9) is given as
75 . 0 = x
( ) ( ) ( ) ( ) ( ) x D x C x B x A x v
i i i i i
4
1
4
1
4
1
4
1
60 cosh 60 sinh 60 cos 60 sin + + + = , (4.4.13)
where i . The desired objective is to achieve no displacement at the free end, i.e., 2 , 1 =
( ) 0 , 2 = t w , (4.4.14)
or equivalently,
( ) 0 2
2
= v . (4.4.15)
The following problem is solved to determine the coefficients . 2 . 1 for , , , , = i D C B A
i i i i

b Az = , (4.4.16)
where
= A
(
(
(
(
(
(
(
(
(
(
(

364 . 2818 447 . 2818 164 . 14 252 . 16 0 0 0 0
680 . 1012 650 . 1012 839 . 5 089 . 5 0 0 0 0
736 . 130 732 . 130 754 . 0 657 . 0 0 0 0 0
710 . 31 750 . 30 826 . 3 735 . 6 710 . 31 750 . 30 826 . 3 735 . 6
049 . 11 394 . 11 420 . 2 375 . 1 049 . 11 394 . 11 420 . 2 375 . 1
094 . 4 970 . 3 494 . 0 0.870 - 4.094 3.970 0.494 - 0.870
0 0 0 0 0 783 . 2 0 2.783
0 0 0 0 1 0 1 0
,
| |
T
D C B A D C B A
2 2 2 2 1 1 1 1
= z , (4.4.17)
38
8
5
2
e b = . (4.4.18)
, 0144 . 0 , 0032 . 0 , 0054 . 0 , 0032 . 0 , 0054 . 0
2 1 1 1 1
= = = = = A D C B A
0686 . 0 and , 0687 . 0 , 0125 . 0
2 2 2
= = = D C B ,
(4.4.19)
( ) ( ) ( ) 75 . 0 60 cos 0032 . 0 75 . 0 60 sin 0054 . 0 75 . 0
4 / 1 4 / 1
1
+ = v
( ) ( ) 75 . 0 60 cosh 0032 . 0 75 . 0 60 sinh 0054 . 0
4 / 1 4 / 1
+ ,
(4.4.20)
( ) 0.0022 75 . 0
1
= v . (4.4.21)
The control gain is determined as
( )
( ) ( )
( )
. 5 . 1781
75 . 0
75 . 0 75 . 0
5 10 , 75 . 0
1
2 1
=

=
v
v v

(4.4.22)
Since is negative, the control can be implemented by means of spring and the system
is stable.
The control force is given by
( ) ( ) ( ) t v t w t u 10 sin 75 . 0 5 . 1781 , 75 . 0 5 . 1781 , 75 . 0
1
= = . (4.4.23)
( ) t t u 10 sin -3.9018 , 75 . 0 = . (4.4.24)

x
( ) t t f 10 sin 2 =
( ) t x w ,
L
5 . 1781 = K
x
( ) t t f 10 sin 2 =
( ) t x w ,
L
5 . 1781 = K
FIGURE 4.7: Controlled uniform beam
The steady state displacement at the free end is as required and is given as
39
( ) ( ) ( ) 2 60 cos 0125 . 0 2 60 sin 0144 . 0 2
4 / 1 4 / 1
2
= v
( ) ( ) 2 60 cosh 0686 . 0 2 60 sinh 0687 . 0
4 / 1 4 / 1
+ .
(4.4.25)
( ) ( ) 0 10 sin 10 1.7764 10 sin 2 ) , 2 (
-15
2
= = t t v t w .
(4.4.26)

Example 2:
The control can also be implemented so as to eliminate steady state motion at any other
prescribed point, for example at 5 . 0 = x . This can be mathematically expressed as
( ) 0 , 5 . 0 = t w , (4.4.27)
or equivalently,
( ) 0 5 . 0
1
= v . (4.4.28)
The sixth row in matrix , in A b Az = changes to accommodate for the condition in
(4.4.26). The new matrix is given by
= A
(
(
(
(
(
(
(
(
(
(
(

364 . 2818 447 . 2818 164 . 14 252 . 16 0 0 0 0
680 . 1012 650 . 1012 839 . 5 089 . 5 0 0 0 0
0 0 0 0 135 . 2 886 . 1 178 . 0 984 . 0
710 . 31 750 . 30 826 . 3 735 . 6 710 . 31 750 . 30 826 . 3 735 . 6
049 . 11 394 . 11 420 . 2 375 . 1 049 . 11 394 . 11 420 . 2 375 . 1
094 . 4 970 . 3 494 . 0 0.870 - 4.094 3.970 0.494 - 0.870
0 0 0 0 0 783 . 2 0 2.783
0 0 0 0 1 0 1 0
,

( ) 0.0046 75 . 0
1
= v . (4.4.29)
The control gain is
( ) . -958.1110 10 , 75 . 0 = (4.4.30)
The control force is given by
( ) ( ) ( ) t v t w t u 10 sin 75 . 0 -958.111 , 75 . 0 -958.111 , 75 . 0
1
= = (4.4.31)
40
( ) t t u 10 sin -4.3685 , 75 . 0 = (4.4.32)
The steady state displacement at 5 . 0 = x is, as desired
( ) ( ) 0 10 sin 10 1.4728 10 sin 5 . 0 ) , 5 . 0 (
-15
1
= = t t v t w .
(4.4.33)

Example 3:
Consider a uniform Bernoulli-Euler beam
0
2
2
4
4
=
t
w
x
w
, , 1 0 < < x , t 0 > (4.4.34)
The position of excitation is at 8 . 0 = = b x . The desired objective is to eliminate steady
state motion at the free end. The mode shapes of the beam before and after
implementing nodal control at position 3 . 0 = = a
, , ,
x
,
are plotted. For the case with the
uncontrolled beam, the coefficients 2 , 1 for = i D C
i i
B A
i i
are determined and the
mode shapes can be determined from (4.4.35) given below.
x
( ) t t f sin =
( ) t x w ,
L
b
x
( ) t t f sin =
( ) t x w ,
L
b

FIGURE 4.8: Uncontrolled uniform beam
i i i i i
cosh sinh cos sin + + + = , i . 2 , 1 = (4.4.35)

41

x
( ) t t f sin =
( ) t x w ,
L
( ) ( ) t a w t u , =
b
a
x
( ) t t f sin =
( ) t x w ,
L
( ) ( ) t a w t u , =
b
a

FIGURE 4.9: Controlled uniform beam
For the case with the controlled beam, the coefficients are
determined and the mode shapes are again determined from (4.4.35). The mode shapes
before and after implementing nodal control are shown in Figure 4.10. It is clear from
the figures that the condition of no motion at the free end is achieved as desired.
3 . 2 , 1 for , , , , = i D C B A
i i i i

42

Before Control After Control
0 0.2 0.4 0.6 0.8 1
-0.008
-0.004
0
0.004
0.008
Mode 2
0 0.2 0.4 0.6 0.8 1
-0.008
-0.004
0
0.004
0.008
Mode 1
x x
) (x v ) (x v
0 0.2 0.4 0.6 0.8 1
-0.004
-0.002
0
0.002
0.004
Mode 3
0 0.2 0.4 0.6 0.8 1
-0.004
-0.002
0
0.002
0.004
Mode 2
x x
) (x v ) (x v
0 0.2 0.4 0.6 0.8 1
-0.004
-0.002
0
0.002
0.004
Mode 4
0 0.2 0.4 0.6 0.8 1
-0.004
-0.002
0
0.002
0.004
Mode 2
x x
) (x v ) (x v
0 0.2 0.4 0.6 0.8 1
-0.08
-0.04
0
0.04
0.08
Mode 1
0 0.2 0.4 0.6 0.8 1
-0.08
-0.04
0
0.04
0.08
Mode 1
x x
) (x v ) (x v
Before Control After Control
0 0.2 0.4 0.6 0.8 1
-0.008
-0.004
0
0.004
0.008
Mode 2
0 0.2 0.4 0.6 0.8 1
-0.008
-0.004
0
0.004
0.008
Mode 1
x x
) (x v ) (x v
0 0.2 0.4 0.6 0.8 1
-0.008
-0.004
0
0.004
0.008
Mode 2
0 0.2 0.4 0.6 0.8 1
-0.008
-0.004
0
0.004
0.008
Mode 1
x x
) (x v ) (x v
0 0.2 0.4 0.6 0.8 1
-0.004
-0.002
0
0.002
0.004
Mode 3
0 0.2 0.4 0.6 0.8 1
-0.004
-0.002
0
0.002
0.004
Mode 2
x x
) (x v ) (x v
0 0.2 0.4 0.6 0.8 1
-0.004
-0.002
0
0.002
0.004
Mode 3
0 0.2 0.4 0.6 0.8 1
-0.004
-0.002
0
0.002
0.004
Mode 2
x x
) (x v ) (x v
0 0.2 0.4 0.6 0.8 1
-0.004
-0.002
0
0.002
0.004
Mode 4
0 0.2 0.4 0.6 0.8 1
-0.004
-0.002
0
0.002
0.004
Mode 2
x x
) (x v ) (x v
0 0.2 0.4 0.6 0.8 1
-0.08
-0.04
0
0.04
0.08
Mode 1
0 0.2 0.4 0.6 0.8 1
-0.08
-0.04
0
0.04
0.08
Mode 1
x x
) (x v ) (x v
0 0.2 0.4 0.6 0.8 1
-0.08
-0.04
0
0.04
0.08
Mode 1
0 0.2 0.4 0.6 0.8 1
-0.08
-0.04
0
0.04
0.08
Mode 1
x x
) (x v ) (x v

FIGURE 4.10: Implementation of nodal control
43
4.5 Summary
The problem of absorbing steady state motion at the free end of a cantilever beam under
the influence of harmonic excitation is investigated. It has been shown that for a fixed
excitation frequency and point of application of control force there exists a unique
control gain that absorbs the harmonic motion at the free end. A closed form
solution for the control gain is given. This result is better than the one given by Ram in
[3] because it does not require infinite product of eigenvalues.

44
Chapter 5: Stability and Optimality
5.1 Introduction
The concept of stability explains the limitations on motion of a vibrating system. A
stable system vibrates in specific bounds while an unstable system has an unbounded
motion. The study of stability also assists in tuning the response of a dynamic system
so that it remains in desired limits. Control systems design involves determining
various parameters like and . Desired overshoot, settling time, response levels,
system robustness, cost of implementing control and simplicity of implementing and
monitoring the control are some of many important factors governing the choice of
these parameters. Study of optimality provides an insight into smart selection of these
physical parameters. Conditions ensuring the stability of the controlled system are
determined in this chapter. Graphs demonstrating regions of active and passive control,
stability and optimal control locations are shown and discussed.
c m, k

5.2 Stability analysis
If the control gain is negative then the control can be implemented by means of a
spring of constant = , which should be attached between point and the ground
as shown in Figure 5.1.
a

45
x
( ) t t f cos =
( ) t x w ,
L

FIGURE 5.1: Passively controlled uniform beam
In this case the system is stable. Note that there are many frequency intervals for which
the control gain is negative. Since the beam in Figure 4.4(b) is obtained by imposing a
single constraint to the beam shown in Figure 4.4(a), the eigenvalues
i
interlace the
eigenvalues
i
in the sense that
L
3 2 2 1 1
< <
(5.2.1)
Hence, it follows from (4.3.19) and (5.2.1) that

Corollary 1
Define 0
0
= .
Then for ,... 3 , 2 , 1 = k
0 > if
k k
< <
1
,
(5.2.2)
and
0 < if
k k
< < .
(5.2.3)
The system is stable whenever 0 < , since the control can be implemented by means
of a passive element, a spring. If 0 > then the system may be either stable or
unstable.
46
Corollary 2
The controlled system is stable if and only if
1
<
(5.2.4)

(a) Controlled beam
F
x
( ) t x w ,
L

x

1
x
F

(b) Equivalent lumped parameter model
for displacement at x = a
FIGURE 5.2: Stability analysis and equivalent stiffness
Proof: By definition from (4.3.16), static deflection is positive. Hence, inequality
(5.2.4) holds for negative . This is the passively controlled case where the control can
be implemented by attaching a spring as discussed above, for which the system is
obviously stable. Instability may arise only for the case where is positive. Consider
47
now the case when is positive. From the linearity of the differential operator in
(4.2.1) it follows that there is a linear relation between force and deflection in the beam
i.e. without the control, ) , ( t a u F = . Hence the deflection at a x = in the controlled
beam can be considered as governed by the sum of two springs, one negative with
constant applied by the control, the other positive with constant 1 applied by the
flexibility of the beam, as shown in Figure 5.2(a) and Figure 5.2(b), where x is an
infinitesimal element of the beam. The system is stable if and only if the equivalent
spring is positive, i.e., 0 > 1 , and otherwise unstable. This is precisely the
inequality (5.2.4).

5.3 Optimality
The topic of optimality discussed here is based on the minimal control force
requirement. Several graphs are shown in Figure 5.3. Considerable information can be
drawn from these graphs in terms of understanding stability, optimality, and type of
control. The graphs are a plot of control force u , control gain ) , ( t a and inverse of
static deflection i.e. over the span of beam described by (4.2.1) through (4.2.5) for
four different values of excitation . Here, L I E , Aand , , can be chosen arbitrarily.

48

x
0 0.2 0.4 0.6 0.8 1
-2
0
4
8
-10
-5
0
5
10
0 0.2 0.4 0.6 0.8 1
x
0 0.2 0.4 0.6 0.8 1
-2
0
2
4
x
0 0.2 0.4 0.6 0.8 1
-5
0
5
x
(a) (b)
(d) (c)
control force control gain
/ 1
3
10
3
10
3
10
3
10
x
0 0.2 0.4 0.6 0.8 1
-2
0
4
8
-10
-5
0
5
10
0 0.2 0.4 0.6 0.8 1
x
0 0.2 0.4 0.6 0.8 1
-2
0
2
4
x
0 0.2 0.4 0.6 0.8 1
-5
0
5
x
(a) (b)
(d) (c)
control force control force control gain control gain
/ 1 / 1
3
10
3
10
3
10
3
10

FIGURE 5.3: Plots of control gain, control force and inverse of static deflection
against the beam span x for cases with excitation frequency of (a) 10 = ,
(b) 20 = , (c) 30 = and (d) 80 =
The following observations can be made from the graphs.
1. The inverse of static deflection is always positive (consistent with the
definition).
Static deflection is positive by definition and hence the inverse of static deflection is
clearly positive.
49
2. Inverse of static deflection is unbounded at 0 = x and . 1 = x
Static deflection is 0 at (clamped end) and at 0 = x 1 = x (desired objective). Hence,
the inverse is .
3. Control force is unbounded near 0 = x .
At location just next to 0 = x the control gain requirement is very large and the
displacement is almost zero. Hence the control force is large. At the control
force should be zero because the displacement is zero.
0 = x
4. The control force is -1 at 1 = x .
Since the excitation amplitude at 1 = x is unity, the control force at is 1. 1 = x
5. The control gain is positive in some regions and negative in some others.
The control gain follows Corollary 1. Again if 0 < , the control can be achieved by
means of a spring, and if 0 > , active control has to be implemented.
6. The control gain is unbounded at 0 = x and 1 = x .
It is not possible to determine control gain at 0 = x numerically because in (4.2.19)
becomes singular. However, using pseudo inverse one can calculate the control gain. It
is obvious that infinite amount of control gain is needed to achieve the state of no steady
state motion at the free end. Also, if control is implemented at the end
A
1 = x , the
control gain required is infinity because the displacement is zero and the control force
which is the product of control gain and steady state amplitude at the end has to be 1.

5.4 Summary
The sign of the control gain i.e., positive or negative, determines the type of control to
be implemented i.e., active or passive. Equations (5.2.2) and (5.2.3) in Corollary 1
50
establish the conditions for the sign of the control gain . If < 0, the control can be
realized by a passive element i.e., spring of constant = k . However, if > 0, then
the control has to be implemented by active means. Corollary 2, through (5.2.4),
establishes the conditions for the controlled system to be stable. The optimal control
force requirements may be obtained from the graphs shown in Figure 5.2.

The issue of robustness is more critical than the issue of optimality of control force,
because at the free end the control force requirement is minimum, but the control gain
required is infinity. A passively controlled system is more robust in comparison to a
system controlled by active means. It is obvious from the graphs that the flatness of the
control gain graph over the beam span is a direct indication of robustness. Also with
increase in the frequency of excitation the flatness of the control gain graphs goes on
decreasing and also there are few locations where control can be implemented by
passive means and active control is impending.

51
Chapter 6: Experimental Verification
6.1 Introduction
The analysis so far has been purely theoretical and has shown that the steady state
motion at a prescribed location on a beam can be eliminated by describing the problem
appropriately and then performing simple mathematical manipulations. The analysis
also helps in determining the value of necessary the control gain, the magnitude of the
control force, the type of control i.e. active or passive and the location of control. The
ultimate objective however, is to be able to implement the theory in practical
applications. A small experiment is carried out in order to validate the method
developed and results derived. This chapter describes the mathematical model, the
experimental setup and the results obtained from the experiment.

6.2 Proposed model for the experiment
In order to validate the theory a simple experiment is constructed so as to mimic the
configuration shown in Figure 6.1. The beam is under the influence of a harmonic
excitation at c x = . The objective is to implement nodal control by using a spring so as
to eliminate the steady state displacement at the free end i.e. at . This is a simple
configuration and can be easily modeled from the information available from previous
analysis and illustrations.
L x =

52

( ) t t f sin =
x
L
( ) t x w ,
c
b
( ) t t f sin =
x
L
( ) t x w , ( ) t x w ,
c
b

FIGURE 6.1: Mathematical model of the test beam used in the experiment
It is imperative that there is significant resemblance in terms of dynamics between the
experimental model constructed and the schematic model shown in Figure 6.1. For this
reason modal analysis is performed on a model of a double-simply-supported beam, as
shown in Figure 6.6. First few natural frequencies of the beam are extracted and
compared to natural frequencies of the beam in Figure 6.1 obtained analytically.

6.3 Determination of natural frequencies
(A) Analytical determination of natural frequencies.
The natural frequencies for the double-simple-supported beam can be extracted
analytically by solving transcendental eigenvalue problem from [13]. However, more
simply, one can write the equation to the beam and impose the boundary conditions (in
this case double-simple-supports and free end) and solve for natural frequencies by
solving a linear algebraic problem as demonstrated in Chapter 3, section 3.4. The
results are not as accurate as one obtained by solving the transcendental eigenvalue
53
problem, but provide significantly close approximation. The governing differential
equation for the beam is given by
0
2
2
4
4
=
t
w
A
x
w
EI , L x < < 0 , t . 0 >
(6.3.1)
Boundary conditions at the simply-supported end are
( ) 0 , 0 = t w ,
(6.3.2)
and

( )
0
, 0
2
2
=
x
t w
.
(6.3.3)
Boundary conditions at the free end are
( )
0
,
2
2
=
x
t L w
,
(6.3.4)
and

( )
0
,
3
3
=
x
t L w
.
(6.3.5)
Dividing the beam in two parts and treating each part independently,
0
2
1
2
4
1
4
=
t
w
A
x
w
EI , 0 , 0 > < < t a x ,
(6.3.6)
and

0
2
2
2
4
2
4
=
t
w
A
x
w
EI , 0 , > < < t L x a .
(6.3.7)
The boundary conditions are,
( ) 0 , 0
1
= t w ,
( )
0
, 0
2
1
2
=
x
t w
,
( )
0
,
2
2
2
=
x
t L w
,
( )
0
,
3
2
3
=
x
t L w
,
(6.3.8)
54
( ) 0 ,
1
= t a w , , ( ) 0 ,
2
= t a w
( ) ( )
x
t a w
x
t a w
, ,
2 1
,
( ) ( )
2
2
2
2
1
2
, ,
x
t a w
x
t a w
.
(6.3.9)
From separation of variables
( ) ( ) t x v t x w
i i
sin , = , 2 , 1 = i .
(6.3.10)
Direct substitution of (6.3.10) in (6.3.6) and (6.3.7) gives,
0
2
=
i i
Av v EI , 2 , 1 = i .
(6.3.11)
Again,
EI
A

2 4
= .
(6.3.12)
Hence,
0
4
=
i i
v v , 2 , 1 = i .
(6.3.13)
Accordingly, the boundary conditions are
( ) ( ) ( ) ( ) , 0 , 0 , 0 0 , 0 0
2 2 1 1
= = = = L v L v v v
(6.3.14)
( ) ( ) ( ) ( ) ). ( ), ( , 0 , 0
2 1 2 1 2 1
a v a v a v a v a v a v = = = =
(6.3.15)
The solution to (6.3.13) is
i i i i i
cosh sinh cos sin + + + = , 2 , 1 = i
(6.3.16)
where, for , , ,
i i i
C B A ,
i
D 2 , 1 = i are constants. Equation in (6.3.16) and the conditions
in (6.3.14) and (6.3.15) together give a set of algebraic equations, which are represented
here in matrix form
b Az =
(6.3.17)
where,
55
(
(
(
(
(
(
(
(
(
(
(
=
a a a a a a a a
a a a a a a a a
a a a a
a a a a
L L L L
L L L L

cosh sinh cos sin cosh sinh cos sin
sinh cosh sin cos sinh cosh sin cos
cosh sinh cos sin 0 0 0 0
0 0 0 0 cosh sinh cos sin
sinh cosh sin cos 0 0 0 0
cosh sinh cos sin 0 0 0 0
0 0 0 0 1 0 1 0
0 0 0 0 1 0 1 0
A
,
, | |
T
D C B A D C B A
2 2 2 2 1 1 1 1
= z
(6.3.18)
and
| |
T
0 0 0 0 0 0 0 0 = b .
(6.3.19)
Clearly there are infinite numbers of solutions for for i for every
position
, , ,
i i i
C B A ,
i
D 2 , 1 =
a x = of the simple support. The values of , which make the matrix
singular, give the eigenvalues
Z
for the Bernoulli-Euler beam by the following
relationship
A
EI

4
= .
(6.3.20)
The natural frequencies of the beam are then given by taking the square root of the
eigenvalues and dividing them by 2 , i.e.,
2
=
n
f .
(6.3.21)
The physical parameters namely, modulus of elasticity E , the cross-sectional area ,
density
A
and moment of inertia I are assumed to be invariant along the beam span.
Parameters and A I are determined from the physical dimensions of the beam.

56

h
w
AISI 1005
Steel
L
h
w
AISI 1005
Steel
L

FIGURE 6.2: Dimensions of the test beam used in the experiment
Beam Dimensions:
mm 0 . 1727 = L ,
mm 8 . 50 = w ,
mm 0 . 6 = h .
Hence,
2
mm 8 . 304 6 8 . 50 = = = h w A ,
(6.3.22)
and
4
3 3
mm 914.40
12
6 8 . 50
12
=
=
h w
I .
(6.3.23)
The beam is made up of AISI 1005 Steel.
Beam properties:
GPa 200 = E .
-3
m - Kg 7870 = .
57
Simulation is performed in MATLAB (Appendix A Program 11) for the double-simple-
supported case discussed above and also for the clamped-simple-supported case.
(B) Determination of natural frequencies from modal test.
A test where the structure or component is vibrated with a known excitation, often out
of its normal service environment, which includes both the data acquisition, and its
subsequent analysis, is called modal testing [14]. It is performed in order to extract the
natural frequencies of the test beam. The complete experimental setup is shown in
Figure 6.6. The following material gives a brief description of the setup, necessary
equipment, procedure of performing the test and some necessary precautions.
(i) Test Beam
The test beam is fastened to a heavy metal frame as shown in Figure 6.4 using clamps

Bolt
Nut
Test Beam
Clamps
Bolt
Nut
Test Beam
Clamps

FIGURE 6.3: Clamping details
at two locations. The details of the clamping are shown in Figure 6.3.
(ii) Impact Hammer
Model 291M78-086C05 from PCB Piezotronics is used to cause an impact. It consists
of an integral ICP quartz force sensor mounted on the striking end of the hammerhead.
58
The frequency range is approximately 5 kHz and hammer range is approximately
22000N.

FIGURE 6.4: Impact Hammer
Its resonant frequency is near 28 KHz. Figure 6.4 shows a picture of the hammer and
the beam.
(iii) Accelerometer

FIGURE 6.5: Accelerometer
Quartz shear ICP accelerometer, Model 353B33, from piezoelectric is used in the test.
The range of frequencies is from 2 Hz to 4000 Hz with % 5 and voltage sensitivity is
100mV per g .

59
(iv) Data Acquisition
NI-4551 from National Instrument is used for data collection. BNC 2140 also from
National Instruments is used to provide ICP power to the accelerometer and the hammer
and connect them to the computer via NI 4551. VirtualBench DSA (Dynamic Signal
Analyzer) is used for signal processing and monitoring.
(v) Connecting Cables
The connecting wires for hammer and accelerometer are recommended to have
impedance in the range of 50~75.

Modal analysis technique
The natural frequencies of a system are a function of the physical parameters, material
properties and boundary conditions of the system. Hence, the frequency response
function of the test beam should remain invariant under ideal conditions irrespective of
the point of measurement and the point of excitation. The accelerometer is mounted on
the beam as shown in Figure 6.5. An impact is made at several locations along the span
of the beam. Again, frequency response is also obtained for several different locations
of accelerometer for same position of impact. The schematic shown in Figure 6.6 gives
a clear picture of the setup and data acquisition for modal analysis. The clamping of the
beam is very critical. Repeated impacts on the beam may cause loosening of the bolts
in the clamps. Also the heavy metal frame should be as rigid as possible and should not
move. These factors change the characteristics of the system and influence the
frequency response function.
60
T
e
s
t

B
e
a
m
I
m
p
a
c
t

H
a
m
m
e
r
A
c
c
e
l
e
r
o
m
e
t
e
r
I
C
P

P
o
w
e
r

U
n
i
t
P
C

w
i
t
h

D
y
n
a
m
i
c

S
i
g
n
a
l
A
n
a
l
y
z
e
r
H
e
a
v
y
M
e
t
a
l
F
r
a
m
e
T
e
s
t

B
e
a
m
I
m
p
a
c
t

H
a
m
m
e
r
A
c
c
e
l
e
r
o
m
e
t
e
r
I
C
P

P
o
w
e
r

U
n
i
t
P
C

w
i
t
h

D
y
n
a
m
i
c

S
i
g
n
a
l
A
n
a
l
y
z
e
r
H
e
a
v
y
M
e
t
a
l
F
r
a
m
e
T
e
s
t

B
e
a
m
I
m
p
a
c
t

H
a
m
m
e
r
A
c
c
e
l
e
r
o
m
e
t
e
r
I
C
P

P
o
w
e
r

U
n
i
t
P
C

w
i
t
h

D
y
n
a
m
i
c

S
i
g
n
a
l
A
n
a
l
y
z
e
r
H
e
a
v
y
M
e
t
a
l
F
r
a
m
e

FIGURE 6.6: Modal Analysis
61
Some important considerations for the Modal Test
It is essential to mount accelerometer securely on the structure to avoid relative
motion between the accelerometer and the structure. Insecure mounting of
accelerometer can cause noise in the collected data for the system. In this test,
an accelerometer is first attached on the beam using a double sided scotch tape
as seen in Figure 6.5 and it is wrapped using a tape around the beam and the
accelerometer for secure mounting.
The location of accelerometer is decided after observation of the dynamic
motion of beam from the side. The mode shape of slow natural frequency can
be observed looking at the motion from the side. The mounting position is then
decided avoiding the nodal points between free end and clamped support.
The impact of the hammer should be quick and sharp. Also for low frequency
range analysis a soft hammer tip should be used. It should be insured that the
amplitude spectrum for the hammer is approximately flat in the frequency range
to be analyzed. This is shown in Display 2 of Figure 6.7.

Figure 6.7 shows the display of the VirtualBench DSA. Sharp peaks in the frequency
response shown in Display 3 are the natural frequencies of the test beam. Time
waveform of the accelerometer and the hammer is shown in Display 1 and the amplitude
spectrum is shown in Display 2.

62

FIGURE 6.7: VirtualBench DSA display
Table 6.1 shows a comparison of the natural frequencies for the two cases with natural
frequencies derived from modal analysis on the test beam.
TABLE 6.1: Comparison of natural frequencies

11.26
73.85 70.95 68.80 4
44.13
23.17
3.6
Modal
Analysis of
Test Beam
(Hz)
55.16
22.28
3.31
Analytical
Clamped
Simple
support
(Hz)
3.13 1
21.14 2
40.76 3
Analytical
Double
Simple
support
(Hz)
Natural
frequency
11.26
73.85 70.95 68.80 4
44.13
23.17
3.6
Modal
Analysis of
Test Beam
(Hz)
55.16
22.28
3.31
Analytical
Clamped
Simple
support
(Hz)
3.13 1
21.14 2
40.76 3
Analytical
Double
Simple
support
(Hz)
Natural
frequency

63
The double-simple-supported case is chosen because the approximation of natural
frequencies is closer to the experimental result. The frequency of 11.26 Hz is believed
to be coming from the overall system comprising the test beam and the heavy metal
frame.

6.4 Determination of the control gain
The control gain calculations are performed as discussed in Chapter 4, Section 4.2.
Actual test beam parameters and dimensions are used. A harmonic force of 45N and
frequency of 3.75 Hz acts exactly in the middle of the test beam span between the two
supports.

( ) t t f 56 . 23 sin 45 =
x
m 7272 . 1
0.327m
( ) ( ) t a w t u , =
0.654m
a
( ) t t f 56 . 23 sin 45 =
x
m 7272 . 1
0.327m
( ) ( ) t a w t u , =
0.654m
a

FIGURE 6.8: Controlled test beam
Graph of control gain variation with changing location of control along the beam span is
shown in Figure 6.9. It is clear from the graphs that is negative between the two
supports (shown by triangles in the graph). Hence the control can be implemented by
means of a spring.
64

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6
-100
-80
-60
-40
-20
0
20
40
60
80
100
control force in N
control gain in N/mm
x
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6
-100
-80
-60
-40
-20
0
20
40
60
80
100
control force in N
control gain in N/mm
x

FIGURE 6.9: Control gain and control force variation along the beam span
In order to verify the accuracy of the calculations, simulation is done to plot the
modeshape of the test beam (Appendix A Program 14 and Program 15) before and after
implementing control. The Figure 6.10 shows that the steady state motion at the free
end is eliminated after implementing control. However, Figure 6.10(b) shows that there
is discontinuity in the shape of the beam. This helps in understanding the huge control
force requirement in the range 0.2m 0.3m and also 0.6m 0.7m.

65

0 0.4 0.8 1.2 1.6
-2.5
-1.5
-0.5
0
0.5
x
) (x v
2
10
0 0.4 0.8 1.2 1.6

-0.8
-0.4
0
0.4
0.8
x
) (x v
2
10
(a) Before Control (b) After Control

0 0.4 0.8 1.2 1.6
-2.5
-1.5
-0.5
0
0.5
x
) (x v
0 0.4 0.8 1.2 1.6
-2.5
-1.5
-0.5
0
0.5
0 0.4 0.8 1.2 1.6
-2.5
-1.5
-0.5
0
0.5
x
) (x v
2
10
0 0.4 0.8 1.2 1.6

-0.8
-0.4
0
0.4
0.8
x
) (x v
2
10
0 0.4 0.8 1.2 1.6

-0.8
-0.4
0
0.4
0.8
0 0.4 0.8 1.2 1.6
-0.8
-0.4
0
0.4
0.8
x
) (x v
2
10
(a) Before Control (b) After Control

FIGURE 6.10: Test beam modeshape before and after control

6.5 Control system design
The graph in Figure 6.9 shows that the most optimal place for placing the spring is at
0.4138m from the first support. This is because before 0.4138m the control force
requirement increases sharply. Also after 0.4138m the control gain requirement
increases sharply even though the control force requirement decreases. The necessary
controlling spring should have a spring constant of 59.34N/mm.

In practical situation a spring of 59.34N/mm positioned at 0.4138m from the first
support may not necessarily eliminate the motion of the free end. However, this
simulation serves as a guideline in choosing both the spring and the location of the
spring on the test beam. A very simple configuration is designed which is extremely
flexible, both in terms of housing springs of variable length and diameter, and also, in
66
positioning the springs at different locations along the span of the test beam between the
two supports. The configuration is shown in Figure 6.11.
Beam Beam

FIGURE 6.11: Spring housing configuration
The photograph shown in Figure 6.14(a) shows how the device can be easily clamped
on the heavy metal frame. The vibration exciter is attached to the test beam by the
configuration shown in Figure 6.12. This is also shown in Figure 6.14(b).

Beam
Vibration
Exciter
Beam
Vibration
Exciter

FIGURE 6.12: Attaching the vibration exciter to the beam
The complete experimental setup is shown in Figure 6.13. Photograph of the same is
shown in Figure 6.15.
67

3
.
6
5
0

H
Z
S
p
r
i
n
g

S
u
p
p
o
r
t
F
u
n
c
t
i
o
n

G
e
n
e
r
a
t
o
r
C
h
a
r
g
e
A
m
p
l
i
f
i
e
r
B
e
a
m

S
u
p
p
o
r
t
s
S
p
r
i
n
g
s
V
i
b
r
a
t
i
o
n
E
x
c
i
t
e
r
T
e
s
t

B
e
a
m
H
e
a
v
y
M
e
t
a
l
F
r
a
m
e
3
.
6
5
0

H
Z
3
.
6
5
0

H
Z
S
p
r
i
n
g

S
u
p
p
o
r
t
F
u
n
c
t
i
o
n

G
e
n
e
r
a
t
o
r
C
h
a
r
g
e
A
m
p
l
i
f
i
e
r
B
e
a
m

S
u
p
p
o
r
t
s
S
p
r
i
n
g
s
V
i
b
r
a
t
i
o
n
E
x
c
i
t
e
r
T
e
s
t

B
e
a
m
H
e
a
v
y
M
e
t
a
l
F
r
a
m
e

FIGURE 13: Schematic of the experimental setup
68

(b) Shaker attachment
(a) Spring Housing
(b) Shaker attachment
(a) Spring Housing

FIGURE 6.14: Attaching the springs and the shaker

FIGURE 6.15: Experimental setup

69
6.6 Procedure
The test beam is excited by a harmonic force of 45N at a frequency of 3.75Hz. The
shaker supplies displacement and not force. Hence, the shaker is attached to the beam
in a way that the beam presses against the shaker. This minimizes the displacement and
the input from the shaker can be treated as a harmonic force. After steady state is
reached significant amplitude of vibration is observed at the free end. The amplitude of
vibrations is of the order of 20mm approximately. Then springs are placed as shown in
Figure 6.11. After steady state is reached the displacement at the free end disappears.

6.7 Experimental result
Two compression springs each of spring constant 127N/mm are placed in series as
shown in the Figure 6.11. This gives a resulting spring constant of 63.5N/mm. These
springs are positioned at a distance of 0.374m from the first support.

6.8 Validation of the proposed theory
From the graph shown in Figure 6.9, a spring of spring constant 63.5N/mm should be
positioned at a distance of 0.4187m from the first support. Hence, the error in the
position is 11.95%. However, from the graph, a spring at a distance of 0.374m from the
support should have a spring constant of about 35.4N/mm. The error in the spring
constant is 79.38%. However this is because the graph for control gain drops sharply in
the vicinity of 0.3m-0.4m that leads to large error. Figure 6.16 shows a schematic of the
experimental setup essential for implementing an active control on the test beam.
70

T
e
s
t

B
e
a
m
A
c
c
e
l
e
r
o
m
e
t
e
r
C
o
m
p
u
t
e
r
9
0
.
0
0

H
Z
V
i
b
r
a
t
i
o
n
E
x
c
i
t
e
r
C
h
a
r
g
e

A
m
p
l
i
f
i
e
r
F
u
n
c
t
i
o
n

G
e
n
e
r
a
t
o
r
D
A
Q

C
a
r
d
H
e
a
v
y
M
e
t
a
l
F
r
a
m
e
T
e
s
t

B
e
a
m
A
c
c
e
l
e
r
o
m
e
t
e
r
C
o
m
p
u
t
e
r
9
0
.
0
0

H
Z
9
0
.
0
0

H
Z
V
i
b
r
a
t
i
o
n
E
x
c
i
t
e
r
C
h
a
r
g
e

A
m
p
l
i
f
i
e
r
F
u
n
c
t
i
o
n

G
e
n
e
r
a
t
o
r
D
A
Q

C
a
r
d
H
e
a
v
y
M
e
t
a
l
F
r
a
m
e

FIGURE 6.16: Active Vibration Control
71
6.9 Summary
The chapter covered details of experimental validation of the theory developed. An
experimental setup was proposed and an appropriate mathematical model was
formulated based on the characterization done by natural frequency identification from
the analytical solution and from modal testing. Control gain calculations were
performed. A simple control system was designed and the experiment was performed.
Comparison is made between the calculated values for spring constant and location of
control force and the results show good agreement.

72
Chapter 7: Conclusions and Recommendations
7.1 Conclusions
Vibration is an inherent phenomenon in dynamic mechanical systems. The work done
in this thesis is to develop a methodology to suppress the undesirable effects of
vibration in a vibrating beam. Large amount of literature is available which covers
several techniques proposed and implemented in order to tune the vibration response of
mechanical systems. The solution to the vibration control problem may involve a great
deal of complexity and may require multidisciplinary knowledge. The designer often
has to strike a trade off between the acceptable levels of vibration amplitude and costs
involved in implementing the control. There may not always be a unique method to
achieve the control and hence, the designer faces the challenge of determining the one,
which is the most optimal.

Singh and Ram in [7] have shown that under certain conditions that have been
characterized in [7] the steady state motion of a certain degree of freedom in a
harmonically excited conservative system may be absorbed by both passive and active
means. Ram in [3] has developed a method to eliminate the steady state motion of a
prescribed location in a continuous system like rod under the influence of a harmonic
excitation. He has presented a closed form solution for the control gain in terms of
infinite product of eigenvalues. This thesis extends the approach in [3] to achieve nodal
control for suppressing vibration at prescribed location in beams and provides a simpler
formula for the control gain in terms of eigenfunctions. Chapter 1 describes in brief the
motivation behind the work done here and introduces the problem investigated in this
73
thesis. Chapter 2 provides a short literature survey of the past work done in the area of
vibration control and presents the theory of a single-degree-of-freedom dynamic
absorber. In Chapter 3, the fundamental theory of vibration in the case of beams is
developed and several graphs, for the steady state response of a clamped cantilever
beam subject to a harmonic excitation plotted against the non-dimensional parameter ,
are shown. This chapter provides the necessary groundwork for investigating the
problem at hand.

Chapter 4 describes in detail the methodology for elimination of steady state response at
a prescribed location for the case of a harmonically excited Bernoulli-Euler cantilever
beam. It is shown by (4.2.22) that for a unique value of excitation frequency and
location of the control force there exists and unique control gain to eliminate steady
state motion at the free end. Several graphs showing the variation of the control gain
requirements with variation in the control location and the excitation frequency are
plotted against the non-dimensional parameter . Another explicit formula for the
control gain involving infinite product of eigenvalues is given by (4.3.1) and compared
with the formula for the control gain in (4.2.22) in Table 4.1 and it shows a good
agreement between the two. The chapter shows some illustrations, which demonstrate
that the steady state motion can be eliminated for any point other than the free end as
well.

In Chapter 5, the conditions for active and passive control are established. Corollary 1
shows the conditions for <0 and >0. If the control gain is negative, the control
74
can be realized by a passive element i.e., spring of spring constant = k . However, if
>0, then the control has to be implemented by active means i.e., by using a single
displacement sensor and a single actuator. Stability criterion is also investigated.
Corollary 2 establishes that the controlled system is stable if and only if < / 1 . A
stable system has a bounded response for arbitrary initial conditions. Several graphs
shown in the chapter may serve as important tools for a vibration control designer in
that they plot the control force requirements and picture the regions where the
implemented control will lead to a stable system.

Chapter 6 describes in detail an experiment conducted to validate the method developed
in Chapter 4 and Chapter 5. An analytical model for the experiment is developed and
control gain calculations are performed. Graphs for the control gain and the control
force requirements along the span of the test beam are plotted to determine the region in
which the control can be implemented by passive means. A spring of the spring
constant value obtained from the graph is used to eliminate steady state motion of the
free end of the beam. There is a good agreement between the theoretical and
experimental results.

7.2 Recommendations for future research
Several openings for performing research remain open. The study of optimality in
Chapter 5 is from the graphs. One needs to address the problem of deriving analytical
expression establishing criterion for realizing optimal control. Once can explore the
75
modeling of a dynamic system controlled by active means. Significant complexities are
introduced in the model because of noise and time delay.
76
References
1. Daniel J. Inman, Vibrations with Control, Measurement, and Stability, Prentice
Hall, 1989.
2. Singiresu S. Rao, Mechanical Vibrations, 2
nd
edition, Addison Wesley, 1990.
3. Ram Y. M., Nodal Control of a Vibrating Rod, Mechanical Systems and Signal
Processing, 2002 (to appear).
4. Den Hartog J. P., Mechanical Vibrations, 4
th
edition, McGraw Hill, 1956.
5. Daniel J. Inman, Engineering Vibration, Prentice Hall, 1994.
6. Lee E. C., Nian C. Y., Tarng Y.S., Design of Dynamic Vibration Absorber
Against Vibrations in Turning Operations, Journal of Materials Processing
Technology, Vol. 108, pp. 278-285, January 2001.
7. Singh K. V., Ram Y. M., Dynamic Absorption by Passive and Active Control,
Journal of Vibration and Acoustics, Vol. 122 (4), pp. 429-433, October 2000.
8. Ram Y. M., Elhay S., The Theory of a Multi Degree of Freedom Dynamic
Absorber, Journal of Sound and Vibration, pp. 607-615, Vol. 195, 1996.
9. Al-Bedoor B. O., Moustafa K. A., Al-Hussain K. M., Dual Dynamic Absorber
For The Torsional Vibrations of Synchronous Motor-Driven Compressors,
Journal of Sound and Vibration, Vol. 220 (4), pp. 729-748, March 4 1999.
10. Yamashita S, Seto K, Hara F, Vibration Control In Piping System by Dual
Dynamic Absorber (3-Dimensional Vibration Analysis And Absorber Design
Using Transfer-Matrix Method), JSME International Journal Series III-Vibration
Control Engineering For Industry, Vol. 32 (4), pp. 531-537, December 1989.
77
11. Aida T, Toda S, Ogawa N, Imada Y, Vibration control of Beams by Beam-type
Dynamic Vibration Absorbers, Journal of Engineering Mechanics, Vol. 118 (2),
pp. 248-258, February 1992.
12. Kawazoe K, Kono I, Aida T, Aso T, Ebisuda K, Beam-type Dynamic Vibration
Absorber Comprised of Free-Free Beam, Journal of Engineering Mechanics.
Vol, 124 (4), pp. 476-479, April 1998.
13. Singh K. V., Ram Y. M., Numerical Algorithm For The Transcendental
Eigenvalue Problem, AIAA Journal (submitted).
14. Ewins D. J., Modal Testing: Theory and Practice, Research Studies Press.
15. Ram. Y. M., Pole Assignment For The Vibrating Rod, Journal of Mechanics and
Applied Mathematics, Vol. 53, Pt. 3, pp. 461-476, 1998.
16. Herzog R., Active Versus Passive Vibration Absorbers, Journal of Dynamic
Systems, Measurement, and Control, Vol. 116, pp. 367-371, September 1994.
17. Nishimura H., Yoshida K., Shimogo T., Optimal Active Dynamic Vibration
Absorber for Multi-Degree-of-Freedom Systems (Feedback and Feedforward
Control Using a Kalman Filter), JSME International Journal, Series III, Vol. 33
(4), 1990.
18. Gaudreault M. L. D., Liebst B. S., Bagley R. L., Simultaneous Design of Active
Vibration Control and Passive Viscous Damping, Journal of Guidance, Control,
and Dynamics, Vol. 16 (4), pp. 654-661, July-August 1993.
19. Elhay S., Gladwell G.M.L. Golub, G.H. and Ram Y.M., On some Eigenvalue-
Eigenvector Relations, SIAM Journal on Matrix Analysis and Applications,
20(3), pp. 563-574, 1999
78
Appendix A
Matlab Programs
1. Program to determine natural frequencies of clamped cantilever beam in
terms of non-dimensional parameter . 1 = = = = = L A I E .

clear all;
n=10000;
for i=1:n,
/////////////////// Defining ////////////////////////////////////////
bet=20;
bet=bet*0.0001*(i-1);
/////////////////// Defining non-dimensional parameter L = ////////////
Beta_L(i,1)=bet;
A=[0 1 0 1;
1 0 1 0;
-sin(bet) -cos(bet) sinh(bet) cosh(bet);
-cos(bet) sin(bet) cosh(bet) sinh(bet)];
Z(i,1)=det(A);
end
/////////////////// Plotting the determinant to determine singularities ///////
/////////////////// The singularities are the natural frequencies ////////////
/////////////////// of the beam ////////////////////////////////////////
figure;
semilogy(Beta_L,abs(Z));
zoom on;

2. Program to determine steady state amplitude for a uniform clamped
cantilever beam with different values of excitation in terms of non-
dimensional parameter . 1 = = = = = L A I E

clear all;
n=2001;
/////////////////// a is the point of excitation ///////////////////////////
/////////////////// b is the point of measurement ///////////////////////
a=1;b=1;
for i=1:n
/////////////////// w is , the frequency of excitation ///////////////////
w=150;
w=(i-1)*0.001*w;
W(i,1)=w;
/////////////////// Defining ////////////////////////////////////////
bet=sqrt(w);
bet=sqrt(W(i,1));
Beta_L(i,1)=bet;
79
Z=[0 1 0 1 0 0 0 0;
1 0 1 0 0 0 0 0;
0 0 0 0 -sin(bet) -cos(bet) sinh(bet) cosh(bet);
0 0 0 0 -cos(bet) sin(bet) cosh(bet) sinh(bet);
sin(bet*a) cos(bet*a) sinh(bet*a) cosh(bet*a)...
-sin(bet*a) -cos(bet*a) -sinh(bet*a) -cosh(bet*a);
cos(bet*a) -sin(bet*a) cosh(bet*a) sinh(bet*a)...
-cos(bet*a) sin(bet*a) -cosh(bet*a) -sinh(bet*a);
-sin(bet*a) -cos(bet*a) sinh(bet*a) cosh(bet*a)...
sin(bet*a) cos(bet*a) -sinh(bet*a) -cosh(bet*a);
-cos(bet*a) sin(bet*a) cosh(bet*a) sinh(bet*a)...
cos(bet*a) -sin(bet*a) -cosh(bet*a) -sinh(bet*a)];
f=[0 0 0 0 0 0 0 -1]';
/////////////////// Solving for the co-efficients //////////////////////////
y=inv(Z)*f;
/////////////////// The co-efficients ///////////////////////////////////
A1=y(1);B1=y(2);C1=y(3);D1=y(4);
A2=y(5);B2=y(6);C2=y(7);D2=y(8);
/////////////////// Steady state amplitude if b a ///////////////////////
v1(i,1)=A1*sin(bet*b)+B1*cos(bet*b)+C1*sinh(bet*b)+D1*cosh(bet*b);
/////////////////// Steady state amplitude if b ///////////////////////// a >
v2(i,1)=A2*sin(bet*b)+B2*cos(bet*b)+C2*sinh(bet*b)+D2*cosh(bet*b);
end
/////////////////// Plotting steady state amplitude against /////////////////
/////////////////// non-dimensional parameter L = ///////////////////
plot(Beta_L,v1)
/////////////////// Labeling the graphs /////////////////////////////////
xlabel(' Beta_L ');
ylabel('Steady state amplitude')
/////////////////// Placing the right axis ////////////////////////////////
axis([0 12 -15 15]);
/////////////////// Introducing text and grid on graphs ///////////////////
gtext('a=1,b=1')
grid

3. Program to determine control gain for a uniform clamped cantilever
beam at four different locations along the length of the beam and with
different values of excitation in terms of non-dimensional parameter .

clear all;
/////////////////// 1 = = = = = L A I E ////////////////////////////////
L=1;
Egp=1;
r=1;
I=1;
Ar=1;
/////////////////// Four location for placing control //////////////////////
for i=1:4
a=0.2*i;
for j=1:1000
80
/////////////////// Defining ////////////////////////////////////////
bet=13.75;
bet=bet*0.001*(j-1);
Beta_L(j,1)=bet;
Z=[0 1 0 1 0 0 0 0;
bet*1 0 bet*1 0 0 0 0 0;
sin(bet*a) cos(bet*a) sinh(bet*a) cosh(bet*a) -sin(bet*a) ...
-cos(bet*a) -sinh(bet*a) -cosh(bet*a);
cos(bet*a)*bet -sin(bet*a)*bet cosh(bet*a)*bet
sinh(bet*a)*bet...
-cos(bet*a)*bet sin(bet*a)*bet -cosh(bet*a)*bet -
sinh(bet*a)*bet;
-sin(bet*a)*bet^2 -cos(bet*a)*bet^2 sinh(bet*a)*bet^2
cosh(bet*a)*bet^2 ...
sin(bet*a)*bet^2 cos(bet*a)*bet^2 -sinh(bet*a)*bet^2 -
cosh(bet*a)*bet^2 ;
0 0 0 0 sin(bet*L) cos(bet*L) sinh(bet*L) cosh(bet*L);
0 0 0 0 -sin(bet*L)*bet^2 -cos(bet*L)*bet^2 sinh(bet*L)*bet^2
cosh(bet*L)*bet^2;
0 0 0 0 -cos(bet*L)*bet^3 sin(bet*L)*bet^3 cosh(bet*L)*bet^3
sinh(bet*L)*bet^3];
f=[ 0 0 0 0 0 0 0 1/Egp/I]';
y=inv(Z)*f;
A=y(1);B=y(2);C=y(3);D=y(4);
E=y(5);F=y(6);G=y(7);H=y(8);
/////////////////// The control gain /////////////////////////////////
alfa(j,1)=Egp*I*(((A-E)*bet^3*cos(bet*a)+(F-B)*bet^3*sin(bet*a)...
+(G-C)*bet^3*cosh(bet*a)+(H-D)*bet^3*sinh(bet*a))...
/(A*sin(bet*a)+B*cos(bet*a)+C*sinh(bet*a)+D*cosh(bet*a)));
end
/////////////////// Storing the control gain for each ///////////////////
/////////////////// location and plotting it against non-dimensional /////////
/////////////////// parameter in different figures //////////////////////
ALFA(:,i)=alfa;
figure(i);
plot(Beta_L,alfa)

axis([0 13.75 -4000 4000])
grid on
end
/////////////////// Plotting control gain for each //////////////////////
/////////////////// location on the same figure against ////////////////////
/////////////////// non-dimensional parameter ////////////////////////
figure(i+1);
grid on
plot(Beta_L,ALFA)
axis([0 13.75 -4000 4000])

81
4. Program to determine eigenvalues of clamped-simply supported uniform
beam FIGURE 4.4(a) in terms of non-dimensional parameter by solving
transcendental eigenvalue problem [**].
4
=

clear all;
for i=1:200,
k=0;
epsi=sqrt(3)*1*0.01;
L=1;
b0=i*0.25;
bet=b0;

while 1e-12<abs(epsi)
bet=b0;
k=k+1;

A=[0 1 0 1;
bet 0 bet 0;
sin(bet*L) cos(bet*L) sinh(bet*L) cosh(bet*L);
-sin(bet*L) -cos(bet*L) sinh(bet*L) cosh(bet*L)];

B=-[0 0 0 0;
1 0 1 0;
cos(bet*L)*L -sin(bet*L)*L cosh(bet*L)*L sinh(bet*L)*L;
-cos(bet*L)*L sin(bet*L)*L cosh(bet*L)*L sinh(bet*L)*L];

s=eig(A,B);
[dum,j]=min(abs(s));
epsi=s(j);
b0=b0+epsi;
if (k>1000)
break;
end

end
B0(i,1)=b0;
B0=real(B0);
K(i,1)=k;
end
plot(B0,'*')

5. Program to determine eigenvalues of clamped-double simply supported
uniform beam FIGURE 4.4(b) in terms of non-dimensional parameter by
solving transcendental eigenvalue problem [**].
4
=

clear all;
for i=1:300,
k=0;
epsi=sqrt(3)*1*0.01;
L=1;a=2/3;
b0=i*0.25;
bet=b0;
82

while 1e-12<abs(epsi)
bet=b0;
k=k+1;
A=[0 1 0 1 0 0 0 0;
bet 0 bet 0 0 0 0 0;
0 0 0 0 -sin(bet*L) -cos(bet*L) sinh(bet*L) cosh(bet*L);
sin(bet*a) cos(bet*a) sinh(bet*a) cosh(bet*a) 0 0 0 0;
0 0 0 0 sin(bet*a) cos(bet*a) sinh(bet*a) cosh(bet*a);
cos(bet*a) -sin(bet*a) cosh(bet*a) sinh(bet*a)...
-cos(bet*a) sin(bet*a) -cosh(bet*a) -sinh(bet*a);
-sin(bet*a) -cos(bet*a) sinh(bet*a) cosh(bet*a)...
sin(bet*a) cos(bet*a) -sinh(bet*a) -cosh(bet*a)];

B=-[0 0 0 0 0 0 0 0;
1 0 1 0 0 0 0 0;
0 0 0 0 cos(bet*L)*L -sin(bet*L)*L cosh(bet*L)*L sinh(bet*L)*L;
0 0 0 0 -cos(bet*L)*L sin(bet*L)*L cosh(bet*L)*L sinh(bet*L)*L;
cos(bet*a)*a -sin(bet*a)*a cosh(bet*a)*a sinh(bet*a)*a 0 0 0 0;
0 0 0 0 cos(bet*a)*a -sin(bet*a)*a cosh(bet*a)*a sinh(bet*a)*a;
-sin(bet*a)*a -cos(bet*a)*a sinh(bet*a)*a cosh(bet*a)*a...
sin(bet*a)*a cos(bet*a)*a -sinh(bet*a)*a -cosh(bet*a)*a;
-cos(bet*a)*a sin(bet*a)*a cosh(bet*a)*a sinh(bet*a)*a...
cos(bet*a)*a -sin(bet*a)*a -cosh(bet*a)*a -sinh(bet*a)*a];

s=eig(A,B);
[dum,j]=min(abs(s));
epsi=s(j);
b0=b0+epsi;
if (k>1000)
break;
end
end
B0(i,1)=b0;
B0=real(B0);
K(i,1)=k;
end
plot(B0,'*')

6. Program to determine determine static deflection at different values of a

clear all;
/////////////////// This is for the case of
3
2
= a //////////////////////////
a=2/3;
/////////////////// 1 = = = = = L A I E ////////////////////////////////
Z=[0 0 0 1 0 0 0 0;
0 0 1 0 0 0 0 0;
a^3 a^2 a 1 -a^3 -a^2 -a -1;
3*a^2 2*a 1 0 -3*a^2 -2*a -1 0;
0 0 0 0 1 1 1 1;
0 0 0 0 6 2 0 0;
83
6*a 2 0 0 -6*a -2 0 0;
6 0 0 0 -6 0 0 0];
f=[0 0 0 0 0 0 0 -1]';
y=inv(Z)*f;
/////////////////// Static deflection //////////////////////////////////
de

l=y(1)*a^3+y(2)*a^2+y(3)*a+y(4);

7. Programs to compare the gain derived from (4.2.22) and (4.3.1)

//////////////////////////////////////////////////////////////////////
A. Control gain derived from (4.2.22)
This is for the case of
3
2
= a
//////////////////////////////////////////////////////////////////////

clear all
/////////////////// 1 = = = = = L A I E ////////////////////////////////
L=1;
Egp=1;
r=1;
I=1;
Ar=1;
/////////////////// Location of control force /////////////////////////////
a=2/3;
/////////////////// Excitation frequency ////////////////////////////////
w=10;
/////////////////// Defining ///////////////////////////////////////
bet=sqrt(sqrt(w^2*r*Ar/Egp/I));
Beta_L=bet;

Z=[0 1 0 1 0 0 0 0;
bet*1 0 bet*1 0 0 0 0 0;
cos(bet*a)*bet -sin(bet*a)*bet cosh(bet*a)*bet sinh(bet*a)*bet...
sinh(bet*a)*bet;
cosh(bet*a)*bet^2 ;
cosh(bet*L)*bet^2;
sinh(bet*L)*bet^3];

f=[ 0 0 0 0 0 0 0 1/Egp/I]';
y=inv(Z)*f;
84
A=y(1);B=y(2);C=y(3);D=y(4);
E=y(5);F=y(6);G=y(7);H=y(8);

/////////////////// The control gain from (4.2.22) ////////////////////////
alfa=Egp*I*(((A-E)*bet^3*cos(bet*a)+(F-B)*bet^3*sin(bet*a)...

//////////////////////////////////////////////////////////////////////
B. Control gain derived from (4.3.1)
This is for the case of
3
2
= a
//////////////////////////////////////////////////////////////////////
clear all;
w=10;
/////////////////// Static deflection //////////////////////////////////
del=0.00914494741655;
/////////////////// Eigenvalues in terms of non-dimensional ////////////
/////////////////// parameter . ///////////////////////////////
4
=
eta_lamda=[3.92660231204792
7.068582745628730
10.21017612281303
13.35176877775409
16.49336143134641
19.63495408493621
22.77654673852600
25.91813939211580
29.05973204570559
32.20132469929538
35.34291735288517
38.48451000647497
41.62610266006476
44.76769531365456
47.90928796724435];
/////////////////// Eigenvalues in terms of non-dimensional ////////////
/////////////////// parameter . ///////////////////////////////
4
=
/////////////////// here
3
2
= a ////////////////////////////////////////
eta_mu=[6.44629453935
10.06139347353
11.77980694026
15.85734089578
19.48566603658
21.20581648246
25.18626731570
28.27433388231
37.69911184308
47.12388980385
56.54866776462];
lamda=eta_lamda.^4;
85
mu=eta_mu.^4;
/////////////////// The control gain from (4.3.1) /////////////////////////
/////////////////// using 7 values of and ///////////////////////////
for j=1:7
K(j,1)=j;
alf=1;
for i=1:j
alf=alf*(1-w^2/lamda(i))/(1-w^2/mu(i)4);
end
ALFA(j,1)=alf/d;
end
/////////////////// Graphical comparison showing //////////////////////
/////////////////// the speed with which the solution converges /////////////
ALFA_EXACT=64.74890912522000*ones(size(ALFA));
plot(K,ALFA,'-*',K,ALFA_EXACT,'.-')

8. Program for Example 1

clear all
//////////////////// 2 , 1 , 3 , 1 , 5 = = = = = L A I E ////////////////////////
L=2;
Egp=5;
r=3;
I=1;
Ar=1;
/////////////////// Location of control /////////////////////////////////
a=0.75;
/////////////////// Defining ///////////////////////////////////////
bet=60^0.25;;

Z=[0 1 0 1 0 0 0 0;
bet*1 0 bet*1 0 0 0 0 0;
-cos(bet*a)*bet sin(bet*a)*bet -cosh(bet*a)*bet -sinh(bet*a)*bet;
cosh(bet*a)*bet^2 ;
cosh(bet*L)*bet^2;
sinh(bet*L)*bet^3];
f=[ 0 0 0 0 0 0 0 2/Egp/I]';
y=inv(Z)*f;
A=y(1);B=y(2);C=y(3);D=y(4);
E=y(5);F=y(6);G=y(7);H=y(8);
86
/////////////////// The maximum amplitude at point /////////////////// a
v=A*sin(bet*a)+B*cos(bet*a)+C*sinh(bet*a)+D*cosh(bet*a);
/////////////////// Steady state amplitude at free end /////////////////////
v_obj=E*sin(bet*L)+F*cos(bet*L)+G*sinh(bet*L)+H*cosh(bet*L);
/////////////////// Control force //////////////////////////////////////
u_at=alfa*v;


clear all
//////////////////// 2 , 1 , 3 , 1 , 5 = = = = = L A I E ////////////////////////
L=2;
Egp=5;
r=3;
I=1;
Ar=1;
/////////////////// Location of control /////////////////////////////////
a=0.75;
/////////////////// Location where motion is to be eliminated //////////////
b=0.50;
/////////////////// Defining ///////////////////////////////////////
bet=60^0.25;;

Z=[0 1 0 1 0 0 0 0;
bet*1 0 bet*1 0 0 0 0 0;
-cos(bet*a)*bet sin(bet*a)*bet -cosh(bet*a)*bet -sinh(bet*a)*bet;
cosh(bet*a)*bet^2 ;
sin(bet*b) cos(bet*b) sinh(bet*b) cosh(bet*b) 0 0 0 0;
cosh(bet*L)*bet^2;
sinh(bet*L)*bet^3];

f=[ 0 0 0 0 0 0 0 2/Egp/I]';
y=inv(Z)*f;
A=y(1);B=y(2);C=y(3);D=y(4);
E=y(5);F=y(6);G=y(7);H=y(8);
87
/////////////////// The maximum amplitude at point a ///////////////////
v=A*sin(bet*a)+B*cos(bet*a)+C*sinh(bet*a)+D*cosh(bet*a);
/////////////////// Steady state amplitude at /////////////////////////// b
v_obj=A*sin(bet*b)+B*cos(bet*b)+C*sinh(bet*b)+D*cosh(bet*b);
/////////////////// Control force //////////////////////////////////////
u_at=alfa*v;


//////////////////////////////////////////////////////////////////////
A. Without control
Excitation at 8 . 0 = b
//////////////////////////////////////////////////////////////////////
/////////////////// Four different excitation frequencies ///////////////////

%w=10,35,70,121
clear all;
/////////////////// 1 = = = = = L A I E ////////////////////////////////
L=1;
Egp=1;
r=1;
I=1;
Ar=1;
w=10;
/////////////////// Excitation at 8 . 0 = b ////////////////////////////////
b=0.8;
/////////////////// Building vector for beam span ////////////////////////
x1=[0:0.01:b]';
x2=[b+0.01:0.01:1]';
/////////////////// Defining ///////////////////////////////////////
bet=sqrt(w*sqrt(r*Ar/Egp/I));

Z=[0 1 0 1 0 0 0 0;
bet*1 0 bet*1 0 0 0 0 0;
0 0 0 0 -bet^2*sin(bet*L) -bet^2*cos(bet*L) bet^2*sinh(bet*L)
bet^2*cosh(bet*L);
0 0 0 0 -bet^3*cos(bet*L) bet^3*sin(bet*L) bet^3*cosh(bet*L)
bet^3*sinh(bet*L);
sin(bet*b) cos(bet*b) sinh(bet*b) cosh(bet*b)...
-sin(bet*b) -cos(bet*b) -sinh(bet*b) -cosh(bet*b);
bet*cos(bet*b) -bet*sin(bet*b) bet*cosh(bet*b)
bet*sinh(bet*b)...
-bet*cos(bet*b) bet*sin(bet*b) -bet*cosh(bet*b) -
bet*sinh(bet*b);
-bet^2*sin(bet*b) -bet^2*cos(bet*b) bet^2*sinh(bet*b)
bet^2*cosh(bet*b)...
bet^2*sin(bet*b) bet^2*cos(bet*b) -bet^2*sinh(bet*b) -
bet^2*cosh(bet*b);
-bet^3*cos(bet*b) bet^3*sin(bet*b) bet^3*cosh(bet*b)
bet^3*sinh(bet*b)...
88
bet^3*cos(bet*b) -bet^3*sin(bet*b) -bet^3*cosh(bet*b) -
bet^3*sinh(bet*b)];

f=[0 0 0 0 0 0 0 1/Egp/I]';
y=inv(Z)*f;
A1=y(1);B1=y(2);C1=y(3);D1=y(4);
A2=y(5);B2=y(6);C2=y(7);D2=y(8);
/////////////////// The modeshapes ///////////////////////////////////
v1=A1*sin(bet*x1)+B1*cos(bet*x1)+C1*sinh(bet*x1)+D1*cosh(bet*x1);
/////////////////// Beam span vector ///////////////////////////////////
x=[x1;x2];
/////////////////// Beam modeshape vector /////////////////////////////
v=[v1;v2];
/////////////////// Plot of beam modeshape /////////////////////////////
plot(x,v)
grid

//////////////////////////////////////////////////////////////////////
B. Without control
Excitation at 8 . 0 = b , control at 3 . 0 = a
//////////////////////////////////////////////////////////////////////
%w=10,35,70,121
clear all
/////////////////// 1 = = = = = L A I E ////////////////////////////////
L=1;
Egp=1;
r=1;
I=1;
Ar=1;
w=10;
/////////////////// Excitation at 8 . 0 = b , control at 3 . 0 = a ////////////////
a=0.3;
b=0.8;
x1=[0:0.01:a]';
x2=[a+0.01:0.01:b]';
x3=[b+0.01:0.01:1]';
/////////////////// Defining ///////////////////////////////////////
bet=sqrt(w*sqrt(r*Ar/Egp/I));

Z=[0 1 0 1 0 0 0 0 0 0 0 0;
bet*1 0 bet*1 0 0 0 0 0 0 0 0 0;
-sin(bet*a) -cos(bet*a) -sinh(bet*a) -cosh(bet*a) 0 0 0 0;
sinh(bet*a)*bet...
sinh(bet*a)*bet 0 0 0 0;
89
cosh(bet*a)*bet^2 0 0 0 0;
0 0 0 0 sin(bet*b) cos(bet*b) sinh(bet*b) cosh(bet*b) -
sin(bet*b) ...
-cos(bet*b) -sinh(bet*b) -cosh(bet*b);
0 0 0 0 cos(bet*b)*bet -sin(bet*b)*bet cosh(bet*b)*bet
sinh(bet*b)*bet...
-cos(bet*b)*bet sin(bet*b)*bet -cosh(bet*b)*bet -
sinh(bet*b)*bet;
0 0 0 0 -sin(bet*b)*bet^2 -cos(bet*b)*bet^2 sinh(bet*b)*bet^2
cosh(bet*b)*bet^2 ...
sin(bet*b)*bet^2 cos(bet*b)*bet^2 -sinh(bet*b)*bet^2 -
cosh(bet*b)*bet^2;
0 0 0 0 -cos(bet*b)*bet^3 sin(bet*b)*bet^3 cosh(bet*b)*bet^3
sinh(bet*b)*bet^3 ...
cos(bet*b)*bet^3 -sin(bet*b)*bet^3 -cosh(bet*b)*bet^3 -
sinh(bet*b)*bet^3;
0 0 0 0 0 0 0 0 sin(bet*L) cos(bet*L) sinh(bet*L) cosh(bet*L);
0 0 0 0 0 0 0 0 -sin(bet*L)*bet^2 -cos(bet*L)*bet^2
sinh(bet*L)*bet^2 cosh(bet*L)*bet^2 ;
0 0 0 0 0 0 0 0 -cos(bet*L)*bet^3 sin(bet*L)*bet^3
cosh(bet*L)*bet^3 sinh(bet*L)*bet^3];

f=[ 0 0 0 0 0 0 0 0 1/Egp/I 0 0 0]';
y=inv(Z)*f;
A=y(1);B=y(2);C=y(3);D=y(4);
E=y(5);F=y(6);G=y(7);H=y(8);
I=y(9);J=y(10);K=y(11);L=y(12);
/////////////////// The modeshapes ///////////////////////////////////
v1=A*sin(bet*x1)+B*cos(bet*x1)+C*sinh(bet*x1)+D*cosh(bet*x1);
v2=E*sin(bet*x2)+F*cos(bet*x2)+G*sinh(bet*x2)+H*cosh(bet*x2);
v3=I*sin(bet*x3)+J*cos(bet*x3)+K*sinh(bet*x3)+L*cosh(bet*x3);
/////////////////// Beam span vector ///////////////////////////////////
x=[x1;x2;x3];
/////////////////// Beam modeshape vector /////////////////////////////
v=[v1;v2;v3];
/////////////////// Plot of beam modeshape /////////////////////////////
plot(x,v);
gr

id on

11. Program for Stability and Optimality analysis along the beam span


%w=10,20,40,80
clear all
/////////////////// 1 = = = = = L A I E ////////////////////////////////
L=1;
90
Egp=1;
r=1;
I=1;
Ar=1;
w=10;
for i=1:1001
a=L;
a=a*0.001*(i-1);
x(i,1)=a;
/////////////////// Defining ///////////////////////////////////////
bet=sqrt(w);

Z=[0 1 0 1 0 0 0 0;
bet*1 0 bet*1 0 0 0 0 0;
sinh(bet*a)*bet...
sinh(bet*a)*bet;
cosh(bet*a)*bet^2 ;
0 0 0 0 sin(bet) cos(bet) sinh(bet) cosh(bet);
0 0 0 0 -sin(bet)*bet^2 -cos(bet)*bet^2 sinh(bet)*bet^2
cosh(bet)*bet^2;
0 0 0 0 -cos(bet)*bet^3 sin(bet)*bet^3 cosh(bet)*bet^3
sinh(bet)*bet^3];

f=[ 0 0 0 0 0 0 0 1/Egp/I]';
y=inv(Z)*f;
A=y(1);B=y(2);C=y(3);D=y(4); % ]constant coefficients
E=y(5);F=y(6);G=y(7);H=y(8); % ]
v(i,1)=A*sin(bet*a)+B*cos(bet*a)+C*sinh(bet*a)+D*cosh(bet*a);
alf(i,1)=Egp*I*(((E-A)*bet^3*cos(bet*a)+(B-F)*bet^3*sin(bet*a)...
+(C-G)*bet^3*cosh(bet*a)+(D-H)*bet^3*sinh(bet*a))...
/////////////////// Static deflection /////////////////////////////////
ZZ=[0 0 0 1 0 0 0 0;
0 0 1 0 0 0 0 0;
a^3 a^2 a 1 -a^3 -a^2 -a -1;
3*a^2 2*a 1 0 -3*a^2 -2*a -1 0;
0 0 0 0 1 1 1 1;
0 0 0 0 6 2 0 0;
6*a 2 0 0 -6*a -2 0 0;
6 0 0 0 -6 0 0 0];

ff=[0 0 0 0 0 0 0 -1]';
yy=inv(ZZ)*ff;
91
del(i,1)=yy(1)*a^3+yy(2)*a^2+yy(3)*a+yy(4); % static deflection
end
/////////////////// Control force //////////////////////////////////////
u_at=alf.*v; % control force
/////////////////// Plot for control force, control gain and inverse //////////
/////////////////// of static deflectioin //////////////////////////////////
figure;
plot(x,[u_at alf 1./del])
xlabel('x');
legend('control force','control gain','static deflection');
axis([0 1 -100000 50000])

12. Program for determining natural frequency of the mathematical model of
the test beam.
(a) double-simple-support and free at the end
(b) clamped-simple-support and free at the end.

clear all
/////////////////// Location of support /////////////////////////////////
a=25.75*0.0254;
/////////////////// Dimensions and properties of test beam ////////////////
L=68*0.0254;
E=2e11;
w=2*0.0254;
t=0.006;
I=t^3*w/12;
A=t*w;
r=7870;
n=1000;
for i=1:n,
/////////////////// Defining ///////////////////////////////////////
b=15;
b=b*0.001*(i-1);
B(i,1)=b;

AA=[0 1 0 1 0 0 0 0;
0 -1 0 1 0 0 0 0;
0 0 0 0 -sin(b*L) -cos(b*L) sinh(b*L) cosh(b*L);
0 0 0 0 -cos(b*L) sin(b*L) cosh(b*L) sinh(b*L);
sin(b*a) cos(b*a) sinh(b*a) cosh(b*a) 0 0 0 0;
0 0 0 0 sin(b*a) cos(b*a) sinh(b*a) cosh(b*a);
cos(b*a) -sin(b*a) cosh(b*a) sinh(b*a)...
-cos(b*a) sin(b*a) -cosh(b*a) -sinh(b*a);
-sin(b*a) -cos(b*a) sinh(b*a) cosh(b*a)...
sin(b*a) cos(b*a) -sinh(b*a) -cosh(b*a)];
Z(i,1)=det(AA);
end
/////////////////// Plotting the determinant to determine singularities ///////
/////////////////// The singularities are the natural frequencies ////////////
92
/////////////////// of the beam ////////////////////////////////////////
figure;
f=((B.^4*(E*I/r/A)).^0.5)/2/pi;
semilogy(f,abs(Z));
zoom on
/////////////////// This is for the clamped case //////////////////////////
AA=[0 1 0 1 0 0 0 0;
1 0 1 0 0 0 0 0;
0 0 0 0 -sin(b*L) -cos(b*L) sinh(b*L) cosh(b*L);
0 0 0 0 -cos(b*L) sin(b*L) cosh(b*L) sinh(b*L);
sin(b*a) cos(b*a) sinh(b*a) cosh(b*a) 0 0 0 0;
0 0 0 0 sin(b*a) cos(b*a) sinh(b*a) cosh(b*a);
cos(b*a) -sin(b*a) cosh(b*a) sinh(b*a)...
-cos(b*a) sin(b*a) -cosh(b*a) -sinh(b*a);
-sin(b*a) -cos(b*a) sinh(b*a) cosh(b*a)...
sin(b*a) cos(b*a) -sinh(b*a) -cosh(b*a)];

13. Program for determination of control gain and control force requirements
for the test beam

clear all
/////////////////// Dimensions and properties of test beam ////////////////
Ln=68*0.0254;
Egp=2e11;
r=7870;
br=2*0.0254;h=0.006;
MI=br*h^3/12;
Ar=br*h;
fr=3.65;
w=2*pi*fr;
/////////////////// Location of support /////////////////////////////////
b=25.75*0.0254;
/////////////////// Location of shaker //////////////////////////////////
c=b/2;
n=1000;
/////////////////// Loop for different values of a x = ////////////////////
/////////////////// i.e. location of control ///////////////////////////////
for i=1:n
a=Ln;
a=a*0.001*i;
x(i,1)=a;
/////////////////// Defining ///////////////////////////////////////
bet=sqrt(w*sqrt(r*Ar/Egp/MI));
/////////////////// Matrix representation of beam equations //////////////
/////////////////// after applying boundary conditions /////////////////
Z=[0 1 0 1 0 0 0 0 0 0 0 0 0 0 0 0;
0 -bet 0 bet 0 0 0 0 0 0 0 0 0 0 0 0;
sin(bet*c) cos(bet*c) sinh(bet*c) cosh(bet*c)...
-sin(bet*c) -cos(bet*c) -sinh(bet*c) -cosh(bet*c)
93
0 0 0 0 0 0 0 0;

bet*cos(bet*c) -bet*sin(bet*c) bet*cosh(bet*c)
bet*sinh(bet*c)...
-bet*cos(bet*c) bet*sin(bet*c) -bet*cosh(bet*c) -
bet*sinh(bet*c) 0 0 0 0 0 0 0 0;

-bet^2*sin(bet*c) -bet^2*cos(bet*c) bet^2*sinh(bet*c)
bet^2*cosh(bet*c)...
bet^2*sin(bet*c) bet^2*cos(bet*c) -bet^2*sinh(bet*c) -
bet^2*cosh(bet*c) 0 0 0 0 0 0 0 0;

-bet^3*cos(bet*c) bet^3*sin(bet*c) bet^3*cosh(bet*c)
bet^3*sinh(bet*c)...
bet^3*cos(bet*c) -bet^3*sin(bet*c) -bet^3*cosh(bet*c) -
bet^3*sinh(bet*c) 0 0 0 0 0 0 0 0;

0 0 0 0 sin(bet*b) cos(bet*b) sinh(bet*b) cosh(bet*b)
0 0 0 0 0 0 0 0;

0 0 0 0 0 0 0 0 sin(bet*b) cos(bet*b) sinh(bet*b)
cosh(bet*b) 0 0 0 0;

0 0 0 0 bet*cos(bet*b) -bet*sin(bet*b) bet*cosh(bet*b)
bet*sinh(bet*b)...
bet*sinh(bet*b) 0 0 0 0;

0 0 0 0 -bet^2*sin(bet*b) -bet^2*cos(bet*b) bet^2*sinh(bet*b)
bet^2*cosh(bet*b) 0 0 0 0;

0 0 0 0 0 0 0 0 sin(bet*a) cos(bet*a) sinh(bet*a)
cosh(bet*a)...

0 0 0 0 0 0 0 0 cos(bet*a)*bet -sin(bet*a)*bet
cosh(bet*a)*bet sinh(bet*a)*bet...
sinh(bet*a)*bet;

0 0 0 0 0 0 0 0 -sin(bet*a)*bet^2 -cos(bet*a)*bet^2
sinh(bet*a)*bet^2 cosh(bet*a)*bet^2 ...
cosh(bet*a)*bet^2;

0 0 0 0 0 0 0 0 0 0 0 0 sin(bet*Ln) cos(bet*Ln)
sinh(bet*Ln) cosh(bet*Ln);

0 0 0 0 0 0 0 0 0 0 0 0 -bet^2*sin(bet*Ln) -
bet^2*cos(bet*Ln) bet^2*sinh(bet*Ln) bet^2*cosh(bet*Ln);

0 0 0 0 0 0 0 0 0 0 0 0 -bet^3*cos(bet*Ln)
bet^3*sin(bet*Ln) bet^3*cosh(bet*Ln) bet^3*sinh(bet*Ln)];
/////////////////// Force vector ///////////////////////////////////////
94
f=[ 0 0 0 0 0 45/Egp/MI 0 0 0 0 0 0 0 0 0 0]';
y=inv(Z)*f;
A=y(1);B=y(2);C=y(3);D=y(4);
E=y(5);F=y(6);G=y(7);H=y(8);
I=y(9);J=y(10);K=y(11);L=y(12);
M=y(13);N=y(14);O=y(15);P=y(16);
v(i,1)=I*sin(bet*a)+J*cos(bet*a)+K*sinh(bet*a)+L*cosh(bet*a);
alf(i,1)=Egp*MI*(((M-I)*bet^3*cos(bet*a)+(J-N)*bet^3*sin(bet*a)...
+(K-O)*bet^3*cosh(bet*a)+(L-P)*bet^3*sinh(bet*a))...
/(I*sin(bet*a)+J*cos(bet*a)+K*sinh(bet*a)+L*cosh(bet*a))); %]
/////////////////// Static deflection //////////////////////////////
ZZ=[0 0 0 1 0 0 0 0 0 0 0 0;
0 1 0 0 0 0 0 0 0 0 0 0;
b^3 b^2 b 1 0 0 0 0 0 0 0 0;
0 0 0 0 b^3 b^2 b 1 0 0 0 0;
3*b^2 2*b 1 0 -3*b^2 -2*b -1 0 0 0 0 0;
6*b 2 0 0 -6*b -2 0 0 0 0 0 0;
0 0 0 0 a^3 a^2 a 1 -a^3 -a^2 -a -1;
0 0 0 0 3*a^2 2*a 1 0 -3*a^2 -2*a -1 0;
0 0 0 0 6*a 2 0 0 -6*a -2 0 0;
0 0 0 0 6 0 0 0 -6 0 0 0;
0 0 0 0 0 0 0 0 Ln^3 Ln^2 Ln 1;
0 0 0 0 0 0 0 0 6*Ln 2 0 0];

ff=[0 0 0 0 0 0 0 0 0 -1 0 0]';

yy=inv(ZZ)*ff;

del(i,1)=yy(5)*a^3+yy(6)*a^2+yy(7)*a+yy(8); % static deflection

end
/////////////////// Control force //////////////////////////////////////
u_at=alf.*v; % control force
/////////////////// Plot for control force and control gain /////////////////
figure(1);
/////////////////// Control gain divided by 1000 /////////////////////////
/////////////////// to convert to N/mm from N/m ////////////////////////
plot(x,[u_at alf/1000])
hold on
/////////////////// Plot to identify location of supports ////////////////////
/////////////////// and location of shaker ///////////////////////////////
plot(0,0,'k^',b,0,'k^',c,0,'^')
xlabel('x');
legend('control force','control gain in N/mm');
axis([0 Ln -100 100])
hold off
zoom on
grid on
/////////////////// Plot of control gain in lb/in ///////////////////////////
figure(2);
95
alf_in=alf*0.00571014587728; %0.2248089/39.37008
del_in=del*39.37008;
plot(x*39.37008,alf_in)
hold on;
plot(0,0,'k^',b*39.37008,0,'k^',c*39.37008,0,'^')
legend('control gain in lb/in');
axis([0 Ln*39.37008 -500 5])
zoom on
grid on

14. Program for plotting modeshape of test beam before control

clear all;
L=68*0.0254;
Egp=2e11;r=7870;
br=2*0.0254;h=0.006;
MI=br*h^3/12;
Ar=br*h;
fr=3.65;
w=2*pi*fr;
b=25.75*0.0254;a=b/2;
x1=[0:0.01:a]';
x2=[a+0.01:0.01:b]';
x3=[b+0.01:0.01:L]';

Z=[0 1 0 1 0 0 0 0 0 0 0 0;
0 -bet 0 bet 0 0 0 0 0 0 0 0;
-sin(bet*a) -cos(bet*a) -sinh(bet*a) -cosh(bet*a) 0 0 0 0;
bet*cos(bet*a) -bet*sin(bet*a) bet*cosh(bet*a)
bet*sinh(bet*a)...
-bet*cos(bet*a) bet*sin(bet*a) -bet*cosh(bet*a) -
bet*sinh(bet*a) 0 0 0 0;
-bet^2*sin(bet*a) -bet^2*cos(bet*a) bet^2*sinh(bet*a)
bet^2*cosh(bet*a)...
bet^2*sin(bet*a) bet^2*cos(bet*a) -bet^2*sinh(bet*a) -
bet^2*cosh(bet*a) 0 0 0 0;

-bet^3*cos(bet*a) bet^3*sin(bet*a) bet^3*cosh(bet*a)
bet^3*sinh(bet*a)...
bet^3*cos(bet*a) -bet^3*sin(bet*a) -bet^3*cosh(bet*a) -
bet^3*sinh(bet*a) 0 0 0 0;
0 0 0 0;
cosh(bet*b);
bet*sinh(bet*b)...
bet*sinh(bet*b);
96
bet^2*cosh(bet*b);
0 0 0 0 0 0 0 0 -bet^2*sin(bet*L) -bet^2*cos(bet*L)
bet^2*sinh(bet*L) bet^2*cosh(bet*L);
0 0 0 0 0 0 0 0 -bet^3*cos(bet*L) bet^3*sin(bet*L)
bet^3*cosh(bet*L) bet^3*sinh(bet*L)];
f=[0 0 0 0 0 45/Egp/MI 0 0 0 0 0 0]';
y=inv(Z)*f;
A1=y(1);B1=y(2);C1=y(3);D1=y(4);
A2=y(5);B2=y(6);C2=y(7);D2=y(8);
A3=y(9);B3=y(10);C3=y(11);D3=y(12);
/////////////////// Determination of modeshape ////////////////////////
x=[x1;x2;x3];
v=[v1;v2;v3];
/////////////////// Plot of modeshape //////////////////////////////////
plot(x,v)
axis([0 1.7272 -0.025 0.005])
grid

15. Program for plotting modeshape of test beam after control

clear all
Ln=68*0.0254;
Egp=2e11;
r=7870;
br=2*0.0254;h=0.006;
MI=br*h^3/12;
Ar=br*h;
fr=3.65;
w=2*pi*fr;
a=14.29*0.0254;
b=25.75*0.0254;
c=b/2;
x1=[0:0.001:c]';
x2=[c+0.001:0.001:a]';
x3=[a+0.001:0.001:b]';
x4=[b+0.001:0.001:Ln]';

Z=[0 1 0 1 0 0 0 0 0 0 0 0 0 0 0 0;

0 -bet 0 bet 0 0 0 0 0 0 0 0 0 0 0 0;

sin(bet*c) cos(bet*c) sinh(bet*c) cosh(bet*c)...
-sin(bet*c) -cos(bet*c) -sinh(bet*c) -cosh(bet*c)
0 0 0 0 0 0 0 0;
bet*cos(bet*c) -bet*sin(bet*c) bet*cosh(bet*c)
bet*sinh(bet*c)...
-bet*cos(bet*c) bet*sin(bet*c) -bet*cosh(bet*c) -
bet*sinh(bet*c) 0 0 0 0 0 0 0 0;
97
-bet^2*sin(bet*c) -bet^2*cos(bet*c) bet^2*sinh(bet*c)
bet^2*cosh(bet*c)...
bet^2*sin(bet*c) bet^2*cos(bet*c) -bet^2*sinh(bet*c) -
bet^2*cosh(bet*c) 0 0 0 0 0 0 0 0;

-bet^3*cos(bet*c) bet^3*sin(bet*c) bet^3*cosh(bet*c)
bet^3*sinh(bet*c)...
bet^3*cos(bet*c) -bet^3*sin(bet*c) -bet^3*cosh(bet*c) -
bet^3*sinh(bet*c) 0 0 0 0 0 0 0 0;

0 0 0 0 0 0 0 0;

cosh(bet*b) 0 0 0 0;

bet*sinh(bet*b)...
bet*sinh(bet*b) 0 0 0 0;

bet^2*cosh(bet*b) 0 0 0 0;

0 0 0 0 0 0 0 0 sin(bet*a) cos(bet*a) sinh(bet*a)
cosh(bet*a)...

0 0 0 0 0 0 0 0 cos(bet*a)*bet -sin(bet*a)*bet
cosh(bet*a)*bet sinh(bet*a)*bet...
sinh(bet*a)*bet;

0 0 0 0 0 0 0 0 -sin(bet*a)*bet^2 -cos(bet*a)*bet^2
sinh(bet*a)*bet^2 cosh(bet*a)*bet^2 ...
cosh(bet*a)*bet^2;

0 0 0 0 0 0 0 0 0 0 0 0 sin(bet*Ln) cos(bet*Ln)
sinh(bet*Ln) cosh(bet*Ln);

0 0 0 0 0 0 0 0 0 0 0 0 -bet^2*sin(bet*Ln) -
bet^2*cos(bet*Ln) bet^2*sinh(bet*Ln) bet^2*cosh(bet*Ln);

0 0 0 0 0 0 0 0 0 0 0 0 -bet^3*cos(bet*Ln)
bet^3*sin(bet*Ln) bet^3*cosh(bet*Ln) bet^3*sinh(bet*Ln)];

f=[ 0 0 0 0 0 45/Egp/MI 0 0 0 0 0 0 0 0 0 0]';
y=inv(Z)*f;
A=y(1);B=y(2);C=y(3);D=y(4);
E=y(5);F=y(6);G=y(7);H=y(8);
I=y(9);J=y(10);K=y(11);L=y(12);
M=y(13);N=y(14);O=y(15);P=y(16);
v1=A*sin(bet*x1)+B*cos(bet*x1)+C*sinh(bet*x1)+D*cosh(bet*x1);
v2=E*sin(bet*x2)+F*cos(bet*x2)+G*sinh(bet*x2)+H*cosh(bet*x2);
98
v3=I*sin(bet*x3)+J*cos(bet*x3)+K*sinh(bet*x3)+L*cosh(bet*x3);
v4=M*sin(bet*x4)+N*cos(bet*x4)+O*sinh(bet*x4)+P*cosh(bet*x4);

x=[x1;x2;x3;x4];
v=[v1;v2;v3;v4];
figure;
plot(x,v);
axis([0 1.7272 -0.01 0.01])
grid on

99
Appendix B
Eigenvalues
Eigenvalues are in terms of non-dimensional parameter
.
TABLE 1 (Appendix B): Eigenvalues

1

3.93

2

7.07

3

10.21

4

13.35

5

16.49

6

19.63

7

22.78

8

25.92

9

29.06

10

32.20

100
TABLE 2 (Appendix B): Eigenvalues for a=0.25

1

4.93

2

8.95

3

12.95

4

16.42

5

18.18

6

21.59

7

25.53

8

28.96

9

30.73

10

33.51

101
TABLE 3 (Appendix B): Eigenvalues for a=0.5

1

6.79

2

8.93

3

13.09

4

15.18

5

19.37

6

21.47

7

31.42

8

33.65

9

37.70

10

43.98

102
TABLE 4 (Appendix B): Eigenvalues for a=
2
1

1

6.16

2

10.17

3

12.28

4

15.18

5

19.29

6

32.18

7

42.90

8

51.50

9

53.63

103
TABLE 5 (Appendix B): Eigenvalues for a=2/3

1

6.45

2

10.06

3

11.78

4

15.86

5

19.49

6

21.21

7

25.19

8

28.27

9

37.70

10

47.12

104
Vita
Akshay Nareshraj Singh was born on October 29, 1977, in New Delhi, India to Alka
and Nareshraj Singh. He did his schooling in Gujarat. After completing his high
school, he joined Faculty of Technology and Engineering, Maharaja Sayajirao
University of Baroda (MSU), Baroda, to do a bachelors degree in Mechanical
Engineering. He decided to pursue graduate program in Mechanical Engineering at
Louisiana State University, Baton Rouge with specialization in vibration and control.
He successfully defended this thesis on his 24
th
birthday i.e. October 29 2001 and will
be graduating with a Masters degree in Mechanical Engineering in December 2001.
He is considering to pursue a Doctoral program at Louisiana State University, Baton
Rouge.
105

Vibration Control of A Beam

Încărcat de

Informații document

Titlu original

Drepturi de autor

Formate disponibile

Partajați acest document

Partajați sau inserați document

Opțiuni de partajare

Vi se pare util acest document?

Este necorespunzător acest conținut?

Drepturi de autor:

Formate disponibile

Vibration Control of A Beam

Încărcat de

Drepturi de autor:

Formate disponibile

Nodal Control of Vibrating Structures: Beam

0 0.4 0.8 1.2 1.6

(a) Before Control (b) After Control

0 0.4 0.8 1.2 1.6

0 0.4 0.8 1.2 1.6

(a) Before Control (b) After Control

S-ar putea să vă placă și