Analytical Dynamics: With An Introduction To Dynamical Systems
Analytical Dynamics: With An Introduction To Dynamical Systems
Analytical Dynamics: With An Introduction To Dynamical Systems
oNewtonian mechanics( F ) : A
->each particle : vector :
oAnalytical dynamics( Energy ) : B
->system as a whole :scalar !
( T , V ; Conservative system) Non-conservative system
% A > B or A = B
Merit or Dis-advantage ?
Vector ? ~ Scalar ?
6 Home works
Mid-term, Final tests
Preface
1. Principles of Dynamics
Mechanics
Basic Principles of Mechanics
Kinematics
Coordinate Transformations
Time Rate of Change of Unit Vector
Kinetics
Work and Energy
Conservative Systems
Systems of Particles
Motion in Non-inertial Reference Frames
Planar Motion of Rigid Bodies
Virtual Work
* Problems
2. Lagrangian Dynamics
Generalized Coordinates
Constraints
Holonomic Systems ~
Kinetic Energy and Generalized Momenta
Generalized Force
Lagrange‟s Equations of Motion
Consevative Systems
Lagrangian Systems
Dissipative Systems
Forces of Constraint
Integrals of Motion
Ignorable Coordinates
Steady Motionand Systems
Lagrange‟s Equations for Impulsive Forces
Electromechanical Analogies
Problems
3. Calculus of Variations
Introduction
Extrema of Functions
Necessary Conditions for an Extremun
Special Cases of the Euler-Lagrange Equations
The Variational Operator
Natural Boundary Conditions
Generalizations
Several Independent Variables
Variational Problems of Constraint
Hamilton‟s Priciple
Problem
4. Dynamics of Rotating Bodies
5. Hamiltonian Systems
6. Stability Theory
Preface:
Principles of analytical mechanics
~mechanicalengr.physics and applied math.
- Topics in Advanced dynamics and elegant variational
approachto formulating problems in mechanics.
Basically, time t is involved !
vSolid mechanics: Stiffness (PDE)
vDynamics: Mass(ODE)
- Solid Mechanics + Dynamics = Real system !
Practice !!!
Chapter 1 PRINCIPLESOF DYNAMICS
F ~a: Inertia property ?
F = ma: Newton law( Principle ) :
(Definition?) : why ? God only knows !
Theory ~ Experiment..
18 th : Laplace:
19 th : Mathematics and Physics ....
20 th : Modern Technologies Age:
:Relativity theory and Quantum mechanics( Einstein..)
Mechanics
Mechanics
KINEMATICS :
Geometric expression only ~coordinate systems
Choice of the reference frame: x , dx/dt , d2x/dt2
Absolute velocity, acceleration
Trajectory ~ path
Speed
Global ~ Local coordinate
Local moving frame of reference
o Normal and tangential coordinates ~ bi-normal
o Cylindrical (polar) coordinates
o Spherical coordinates
COORDINATE TRANSFORMATIONS
TIME RATE OF CHANGE OF A UNIT VECTOR
Rotating coordinate system with angular velocity:
KINETICS :
Force ~ origin of motion !
CONSERVATIVE SYSTEMS
Mechanics
KINEMATICS :
Geometric expression only ~coordinate systems
Choice of the reference frame: x , dx/dt , d2x/dt2
Absolute velocity, acceleration
Trajectory ~ path
Speed
Global ~ Local coordinate
Local moving frame of reference
Rectangular coordinate system ~ Moving coordinate system ;orthogonality !
( x, y, z)
(r , , z )
(r , , )
KINETICS :
Force ~ origin of motion !
CONSERVATIVE SYSTEMS
SYSTEMS OF PARTICLES
Rc?
Newton’s 2nd law
Fi fij mi ai ; ri Fi ri fij ri mi ai
i 1 i 1 j 1 i 1 i 1 i 1 j 1 i 1
= > (1.50) !
Angular momentum ..
Kinetic Energy
T= 2 2 j 1
N .
Only for
m r
j 1
j j 0
MOTION IN NONINERTIAL REFERENC FRAMES
or
Gun ? I.C.B.M. ?
~Non-inertial frame !
Then
vp = v B + ( r rel) + v rel
=?
Acceleration:
v B+ dt ( r rel) +
d d d d
dt
vp = dt dt
v rel
ap = a B + ( r rel) + dt r rel +
d d
dt
v rel
Remember
( r rel) + v rel
d
r
dt rel =
Then,
ap = a B + ( r rel) + dt r rel +
d d
dt
v rel
Finally,
Coriolisforce :
Time is involved !
Mechanics:
~ System as a whole !
Kinetic Energy and Generalized Momenta
configuration space.
: Generalized coordinate !
Physical coordinate:
xi=xi(q1,q2,q3,t)(2.15)
For i = 1
what happens?
1 dxi dxi 1
Next, T= m m( xi , j qj xi , t )( xi , k qk xi , t )
2 dt dt 2
1 1 1
= mxi , j xi , k qj qk m( xi, j xi, t qj xi, k xi, t q k ) mxi, t xi, t)
2 2 2
nd
2 term : index k -> j
Why ?
1 1 1
= jk qj qk 2 m xi , j xi , t qj mxi , t xi , t )
2 2 2
1
= 2 jk qj qk j qj (2.17)
coordinates ! ->
T=T( q, dq/dt , t)
T = T2 + T1 +T0
dependent on t T1=T0=0
Generalized momentum
linear momentum
components of position
Using the definition, px =mdx/dt ,py=.., pz=..
p mr 2 cos2 (angular..momentum)
.
p mr 2 (angular..momentum)
Generalized Force
- Vector mechanics :
force, moment
- Analytical mechanics:
momentum !
Energyconcept !
xi
xi q j ( j 1...n)
q j
(2.29)
Virtual displacement is defined for time is fixed
t 0
generalized coordinates
Kinetic Energy and Generalized Momenta
EX:
(r, , z)and (r, , ) : orthogonal coordinate system
Eqn(2.17) <~>Eqn(2.19)
T=T( q, dq/dt , t)
T = T2 + T1 +T0
Energyconcept !
Chain rule,
xi
xi q j ( j 1...n)
q j
(2.29)
Objective
generalized coordinates.
generalized coordinates qj
x( i )
F ( i ) xx ( i ) F (i ) x q j
q j (2.30)
Individual terms in the summation ~
What is Qj ?
coordinates fixed.
displacement. ( independent !)
coordinates at a time.
inphysical space ~
attached to it.
x x , (2.34)
And x 0 , (2.35)
Note:
to zero.
Therefore, in the computation of generalized forces,
coordinates.
V V ( x1 , y1 , z1 , x2 , y2 , z2 , , xN , yN , zN )
Fi iV
potential function :
N
V V V
W ( xi yi zi ) V (2.44)
i 1 xi yi zi
potential energy.
V 0
Suppose :Single potential function
V V ( x1 , y1 , z1 , x2 , y2 , z2 , , xN , yN , zN )
Fi iV (2.42)
in here
ˆ ˆ ˆ
i i+ j k
xi yi zi
SubstituteEqn.(2.42) into W : Eqn.(2.31) ,
N
V V V
W ( xi yi zi ) V (2.44)
i 1 xi yi zi
(MINI ?MAX ?)
V 0
of a conservative system.
Transforming to generalized coordinates :
V V (q1 , q2 , , qn ) (2.46)
of q is: Eqn(2.47)
j
or
n
V
W q j V , q j q j ( for j=1..n)(2.48)
j 1 q j
For a conservative system:
That is,
V
Qj (2.49)
q j
forholonomic systems
: Lagrangiandynamics !!
coordinates.
there is a transformation :
miai= Fi(2.50)
or
dpi
Fi (2.51)
dt
: linear momentum of the i-th particle as
pi mi ri (2.52)
1 N
T mi ( xi2 yi2 zi2 )
2 i 1 (2.54)
T N
xi yi zi
pk mi ( xi yi zi )
qk i 1 qk qk qk (2.55)
Remember the chain rule :
N
xi x
xi qj i
j 1 q j t
xi xi
qk qk (2.57)
( Remember : ( x y) x y x y )
dt
d T N
xi yi zi
( ) mi ( xi yi zi )
dt qk i 1 qk qk qk
N
d xi d y d z
mi [ xi ( ) yi ( i ) zi ( i )]
i 1 dt qk dt qk dt qk
note that
d xi n
2 xi 2 xi
(
dt qk
)
j 1 q j qk
qj
t qk
n
xi xi
[ qj ] xi x i , k
qk j 1 q j t qk
d T T
( ) Qk k 1, 2,
dt qk qk ,n (2.61)
d T T
( ) Qk k 1, 2,
dt qk qk ,n (2.61)
PDE !
d T
( pk ) Qk
dt qk , k 1, 2, ,n
coordinates q1 , q2 , qn .
transformations.
equations of motion.
generalized velocity as
1 2 1 2 2
T mv ml
2 2
From a previous example, the generalized force
Q mgl sin
d
(ml 2 ) 0 mgl sin
dt
g
sin 0
l
virtual work.
the particle.
The constraint force is in fact the tension in the cable.
Ex :
1 1
T m1q1 m2 q2
2 2
virtual work.
Then
Q1 k1q1 k2 ( q2 q1 ) , Q2 k2 ( q2 q1 )
d T T d T T
( ) Q1 ( ) Q2 (2.62)
dt q1 q1 , dt q2 q2
in standard form as
m2 q2 k2 q1 k2 q2 0
In a Matrix Form ?
CONSERVATIVE SYSTEMS
dynamical system.
generalized forces.
For a conservative system, there exists a potential
V V (q1 , q2 , , qn )
V
Qk
qk (2.63)
generalized coordinates,…
Thus
T (T V )
qk qk
form.
Lagrangian function:
velocities.
conservative system.
1 2 2
Sol) T= 2 mr ( 2
sin 2
) , V= mgr cos , L=T-V
1 2 2
Sol) T= m( x y )
, V= mgy with y 3x 2
2
Insert : y ~ Finally,
4. Text : p.120
LAGRANGIAN SYSTEMS
Suppose :
in Text !
d L(q, q, t ) L(q, q, t )
[ ] 0 (2.67)
dt qk qk
e
e { A v curlA}
c
: e - charge carried by the particle,
scalar potential,
generalized potential
ev A
V (r , r ) e (r )
c
Not all systems are Lagranian, although all generalized
forces~Conservative or Non-conservative
system.
V
Qk Qknc
qk
Construct the Lagrangian function L T V and formulate
d L L
(
dt qk
)
qk
Qknc
, k 1, 2, ,n
potential function.
DISSIPATIVE SYSTEMS
: frictionforce isnon-conservative,
particles
N
1 N 2
(cx xi c y yi cz zi ) qk
2 2
k 1 2 k 1 qk
i i i
1 N
Qknc (cxi xi2 c yi yi2 czi zi2 )
2 i 1 qk
…
Now, define a scalar function forgeneralized velocities
1 N
D (cxi xi2 c yi yi2 czi zi2 )
2 i 1
motionas
d L L D
( ) Qk* : (2.68)
dt qk qk qk
non-conservative system.
1 2
D cq
2
d
(mq) (kq) cq A cos f t
dt
which can be put into standard form as
mq cq kq A cos f t
M f
Instantaneous rate of energy loss
P M f
1
D 2
2
d
(ml 2 ) (mgl sin )
dt
FORCES OF CONSTRAINT
analyzed.
the constraints.
Lagrangianformalism:
analyzed.
coordinates(superfluouscoordinates)corresponding to
displacements becomes
Qk Ak
d T T
(
dt qk
)
qk
Qk Ak , k 1, 2, ,n (2.70)
constraints given by
or equivalently as
coordinate is
J
Qk j Ajk
j 1
d T T J
dt qk Q A jk qk (2.72)
qk
k j
j 1
The set of eqns (2.72) together with the J eqns of
coordinateqk:
J
Rk j Ajk
j 1
Method of Lagrange multipliers may also be applied
f f
Ajk Aj 0
qk t
types.
as
r 0
2br 01
0 1 0 0
z 0
1
L m(r 2 r 2 2 z 2 ) mgz
2
mr mr 2 2b1r
d
(mr 2 ) 2
dt
mz mg 1
r r 2 2br ( z g )
z 2br 2 2brr
rotation is
2 2mrr
Rr 2b1 (t )r (t ) and Rz 1 (t )
Practice !
As an example, (p.120):Consider the dynamics of a
particle constrained to slide on a frictionless wire. This
as
r 0
2br 01
0 1 0 0
z 0
Hence there will be two Lagrange multipliers – one for
1
L m(r 2 r 2 2 z 2 ) mgz
2
mr mr 2 2b1r
d
(mr 2 ) 2
dt
mz mg 1
Now since , there are four unknowns r (t ) z (t ) 1 (t )
, , , and
r r 2 2br ( z g )
z 2br 2 2brr
and so we end up with the single differential equation for
r (t ) as
rotation is
2 2mrr
Rr 2b1 (t )r (t ) and Rz 1 (t )
Practice!
INTEGRALS OF MOTION
eqns of motion.
Generally, eqns of motion are too complicated !
time, then
G(q1 , q2 , , qn ; q1 , q2 , , qn ; t ) C (2.74)
orequivalently
dG
0
dt
system.
system.
d L(q, q, t ) L(q, q, t )
[ ] 0
dt qk qk (2.75)
L d L
qk dt qk
this means
dL n
L n
L L
qk qk
dt k 1 qk k 1 qk t
n
d L L
qk
k 1 dt qk t
Therefore
d n L L
k q L ..(2.76)
dt k 1 qk t
dh L
dt t
n
L
k 1
qk
qk
L h cosnt
T2 T0 V h (2.77)
mechanical energy:
T V h
Suppose :
1 1
T (m M )v02 ml 2 2 mv0 l cos
2 2
Potential energy :
V mgl cos .
Then, Lagrangian :
1 1
L (m M )v02 ml 2 2 mv0 l cos mgl cos
2 2
Kinetic energy is not purely quadratic, but
1 2 2 1
ml (m M )v02 mgl cos h : (2.78)
2 2
1 1
h ml 2 o 2 (m M )v02 mgl cos o
2 2
Hence Eqn (2.78) may also be written as
1 2 2 1
ml mgl cos ml 20 2 mgl cos 0
2 2
It should be note:
IGNORABLE COORDINATES
Lagrangian system (n dof) and generalized coordinates
q1 , q2 , , qn .
d L L
( ) 0, k 1, 2, ,n m
dt qk qk
d L
( ) 0, i n m 1, ,n
dt qi (2.80)
Eqn (2.80) :Last mcoordinates qnm1 , , qn do not
Or inactive coordinates.
L
Ci
qi (2.81)
~ Generalized coordinates and velocities :conserved,
velocities,
V
0
qi
then,
L T
qi qi
pi Ci ..(2.82)
generalized coordinates q1 , q2 , , qn m .
and velocities.
L
Ci , i n m 1, ,n (2.83)
qi
??
d R R
( ) 0, k 1, 2, , n m
dt qk qk (2.86)
of motion Cn m1 , , Cn .
R
qi ...(2.87)
Ci
integratedas
t R
qi (t ) d , i n m 1, , n,,,,(2.88)
t0 C
i
RouthianFunction :
in polar coordinates as
1
L m(r 2 r 2 2 ) V (r )
2
- is ignorable; conjugate momentum is constant,
p mr 2 C
C 1 2 1 2 C 2
R(r , r , C ) C L 2 C { mr mr ( 2 ) V (r )}
mr 2 2 mr
>>
1 2 C2
R(r , r , C ) mr 2
V (r )
2 2mr
C2
mr 3 V '(r ) 0....(2.89)
mr
Eq.(2.89) * dr ~
C2
(mr 3 V '(r ) 0) dr....(2.89)'
mr
>>
1 2 C2
mr 2
V (r ) E0 (2.90)
2 2mr
C2
Veff 2
V (r )
2mr
Hence, construct the phase curves based onEqn (2.90)
1 dr
mr Veff (r ) E0 (r
2
2( E0 Veff (r )) / m :
2 dt
dt ......!
m r dr
2 r0 E0 Veff
t t0
C
Also, 2
mr
thus
t C
(t ) d 0
t0 mr 2
STEADY MOTION
zero.
>>
results in
qi (t ) vit const, i n m 1, , n
substituting
qk 0 and qk 0
qk (t ) qk 0 sk (t )
d R R
( ) 0, k 1, 2, ,n m
dt sk sk
1 2 2 2
L ml ( sin 2 ) mgl cos
2
p ml 2 sin 2 C....(2.93)
: An integral of motion.
The analysis is reduced to a single degree of freedom
C2 1 2 2
R( , , C ) ml mgl cos
2ml sin 2
2 2
C 2 cos
ml 2 3 mgl sin 0 (2.94)
2
ml sin
C2
Veff mgl cos ....(2.95)
2ml sin
2 2
dVeff
ml 2 0
d
C 2 cos
mgl sin 0
ml sin
2 3
or
C 2 cos m2 gl 3 sin 4
0 s(t )
g
s (3cos 0 sec 0 ) s 0
l
interval of impact.
t0 t
F̂=
t0
F(t )dt
t2 t0 t , we have
t2 d T t2 T t2
t1
dt qk
dt t1
qk
dt Qk dt ,
t1
k 1, 2, , n (2.97)
Now letting t 0 ,
T T
Qˆ k k 1, 2, ,n
qk 2 qk 1 (2.98)
pk Qˆ k , k 1, 2, , n (2.99)
individual contributions to Ŵ .
8
T 2mx12 mb2 2
3
16 2
px1 4mx1 , p mb
3
Similar to computing virtual work, we consider the
x1 x1 x1 , 0 and x1 0,
Qˆ x1 Fˆ , Qˆ 2b sin Fˆ
Fˆ 3sin ˆ
x1 , F
4m 8mb
Practice !
1
Virtual work of impulse : W Fˆ ( y c L ) Qˆ xc xc Qˆ yc yc Qˆ
2
1
~ Qˆ xc 0, Qˆ yc Fˆ , Qˆ LFˆ
2
( m y c ) m( y v ) Fˆ (2)
L ˆ 1
( I ) I F (3) : ( I mL2 )
2 3
L L
And .. y c ev y c ev (4)
2 2
2
(4) : (ev yc ) (3) Fˆ ...
L
v 6
: yc (4e v).. (1 e)v
7 7L
ELECTROMECHANICAL ANALOGIES
The Lagrangian formalism is based on energy and
therefore has applicability that goes far beyond simple
mechanical systems (Fig. 2.18). A very practical extension
of the theory is to electrical circuits and combined
electromechanical systems. A direct application of
Lagrangian‟s equations to electrical circuits is based on
the parameters given in Table 2.1. Energy carried by an
inductor coil is
CHAPTER THREE:
CALCULUS OF VARIATIONS
INTRODUCTION
EXTRMA OF FUNCTIONS
NECESSARY CONDITIONS FOR AN EXTREMUM
SPECIAL CASES OF THE EULR-LAGRANGE EQN
THE VARIATIONAL OPERATOR
NATURAL BOUNDARY CONDITIONS
GENERALIZATIONS
SEVERAL INDEPENDENT VARIABLES
VARIATIONAL PROBLEMS WITH CONSTRAINTS
HAMILTON‟S PRINCIPLE
INTRODUCTION
Lagrangian formulation of eqns of motion:
~ Energy of a system and work done by external forces.
: Dynamical system with inertial properties–
: Kinetic energy
Also,conservative generalized forces ~Derivable from
: Potential energy
*Non-conservative part ~
Applying virtual workprinciple !
Most systems can be synthesized by the construction
ofa Lagrangian ~
Theoretical & Experimentally Verified !
oPrimary feature of Lagrangian dynamics ~
Independent of any coordinate system
( x1 ,......xN ) f ( x1 ,...xN ) R.
Graph of a function of one variable is a curve
~ Graph of a function of two variables is a surface.
Generalization of a function is called a functional.
An assignment of a real value to a point, to avector, or
to an entire function.
Now concern here with functionalsthat are defined on
some suitable space or set of functions.
: Functionals~Integral functionals.
- Functionals defined by the integration of some
expression involving an input function:
f ( x) I [ f ( x)]
Integral functional I [..] may be of the form (3.1).
Ex: integral functionals include area under the graph of a
function and arclength of a curve between two points.
Ex might look like …
where p( x) and q( x) are specified.
Integral functionals may be defined as integrals over
some interval or as integrals over some region in space.
Integration interval is the domain of the input function.
Argument or input of an integral functional may be a
single function or several functions.
It all depends on the context of the problem. It is
important to distinguish between the domain of the
input function and the domain of the functional itself,
which is comprised of some class of admissible
functions.
Consider afunctionals of the form
x1
I [ y( x)] F ( x, y( x), y( x))dx (3.2)
x0
EXTREMA OF FUNCTIONS
An extremum problem consists of finding the largest or
smallest value of a quantity. For functions of one or
two variables, the function can be graphed and we
immediately see where the function attains its
extreme values. These may be inside of a domain, or
the extrema may be located at points on the boundary.
For functions of more than two variables, graphing is
not possible, so we must report to performing a
comparison of the values at a point with neighboring
values. That is, we examine the local rate of change of
a function. These ideas carry over directly to finding
extrema of integral functionals.
To find the extrema of a function inside an interval, we
look for local stationary behavior. At a point x where
a function attains a local extremum, given an
infinitesimal change, the value of the function should
remain the same; otherwise we do not have an
extremum. This is the same as examing the local
linearization of the function. At an extremum, the
function should be flat. The rate of change in every
possible direction must be zero. Since there are only
two directions, this is easy. Analytically, this means
that the differential of of the function is equal to zero.
Hence, a necessary condition that the function f (x) be
stationary at a point x is that the derivative f (x ) is
0 0
infinitesimal displacement:
Now the function
f ( r, ) f ( x x, y y)
0 0
That is,
**(3.7)
The condition (3.7) give a necessary and sufficient
condition for local stationary behavior of ** at the
point **.
These conditions extend to higher dimensions for
functions of n variables **, as
**
We now generalize these concepts to establish criteria for
stationary values of scalar quantities that depend on
entire functions as arguments. Before formally
developing the theory, let us consider a simple
motivational example.
Remember : , d ,
Lagrangianeqn in hybrid form : Page.115
x1
I [ y( x)] F ( x, y( x), y( x))dx (3.2)
x0
EXTREMA OF FUNCTIONS
NECESSARY CONDITIONS FOR AN EXTREMUM
Establishing local stationary behavior of a functional is a
generalization of locating the critical points of a function.
:Local maxima and minima are found by setting the
derivative of the function w.r.t independent variable
equals to 0 ~ slop =0 (Fig.3.4) !
-Value of the function is stationary.
G( x) :Continuous function in [ x1 , x2 ]
x2
G( x) ( x)dx 0
x1
~>
x1
I [ y * ] F ( x, y * , y * )dx (3.8)
x0
F d F
( ) 0....(3.9)
y dx y
~An ordinary differential equation that represents the
necessary condition for an admissible y( x) to be a
minimizing function.
: Euler-Lagrange equation
Back to the future ! P. 115
x t
y , y q , q !
Ex : (3.10) : p.169
HAMILTON‟S PRINCIPLE
THE VARIATIONAL OPERATOR(?)
NATURAL BOUNDARY CONDITIONS
F
…. y x2
0...(3.15)
y
x1
as in Fig.3.6
Then, at x for I 0 , Euler-Lagrange Eqn. is satisfied
2
F
with y =0 : Natural BC or Force BC.(ex: ICBM…)
( y =0 :Essential BC or Geometric BC)
GENERALIZATIONS
F F ( x, y, y) : most.simple!
~> For more generality with several functions !
As in Eqn.(3.16)
y =( y 1 , ..., yn ): independent Ftns
Ex : !
~ Integral constraint
Eqn(3.20)
Isoperimetric problem
~Eqns of constraint
(Non~) holonomic !
G ( x, y1..., yn ) 0.....(3.25)
or
Taking.variation,
G, yi yi 0
THE VARIATIONAL OPERATOR(?)
NATURAL BOUNDARY CONDITIONS
GENERALIZATIONS
y ( x, t ) y ( x, t )
4 2
1 y( x, t ) 2 1 2 y( x, t ) 2
~ T m( ) ,V EI ( )
2 t 2 x 2
HAMILTON‟S PRINCIPLE
Aim:
? Lagrangian formulation ~ Calculus of Variation?
Remember !
d 1
(mi xi xi ) mi xi xi mi xi xi W [ mi xi 2 ]..(i 1,,3N )
dt 2
Eqn(3.28):
T W (mi xi xi ),t
HAMILTON‟S PRINCIPLE