Lagrangian Mechanics

ide
Sk Jahiruddin*

gu *Assistant Professor
Sister Nibedita Govt. College, Kolkata
ics
Author was the topper of IIT Bombay M.Sc Physics 2009-2011 batch
He ranked 007 in IIT JAM 2009 and 008 (JRF) in CSIR NET June 2011

He has been teaching CSIR NET aspirants since 2012

ys
Ph

1
©Sk Jahiruddin, 2020 Lagrangian Mechanics

Contents

1 Introduction 4

ide
2 Constrains 4

3 Generalize coordinates 6

gu
4 Lagrangian 6
ics
5 Non-uniqueness of Lagrangian: Gauge 13

6 Velocity dependent potential 14

ys

7 charged particle in em field 15

Ph

8 Hamilton’s Principle 15

8.1 Configuration Space . . . . . . . . . . . . . . 16

8.2 Action . . . . . . . . . . . . . . . . . . . . . . 16

[email protected] 2 Physicsguide
©Sk Jahiruddin, 2020 Lagrangian Mechanics

8.3 Hamilton’s Principle . . . . . . . . . . . . . . 17

9 Symmetry and Conservation 18

ide
10 Energy conservation and Hamiltonian 22

gu
ics
ys
Ph

[email protected] 3 Physicsguide
©Sk Jahiruddin, 2020 Lagrangian Mechanics

1 Introduction

Newton’s Equation of Motion predicted the motion of any

ide
object(s) from the force. So, in principle, If you know the
force then you can integrate it and get position as a function
of time.

But there are many problems

gu
ˆ The force may not be known, and it it known, then also
there comes many constrain forces like Tension, Reaction
ics
force etc.

ˆ Forces may depend on time, location, or even velocity.

ˆ There are always three coordinates for a point object, 3N

ys

coordinate for N particles.

Ph

2 Constrains

Constrains are the conditions which impose restrictions on a

dynamic system.

Holonomic Constrains: Constrains which can be ex-

[email protected] 4 Physicsguide
©Sk Jahiruddin, 2020 Lagrangian Mechanics

pressed as f (r1 , r2 , r3 , . . . , t) = 0 is called a holonomic equa-

tion. Examples could be (1) Particle on the x -y plane −→
z = 0 . (2) Rigid body −→ (ri − rj )2 − c2ij = 0 etc.

ide
Nonholonomic Constrains: All other cases are called
nonholonomic.

It means ”we don’t really want to mess with it”. It may

be inequalities such as z ≥ 0 Or it may depend on deriva-

gu
tives such as ṙi . However we will deal only with holonomic
constraints.
ics
A holonomic constraint reduces the number of indepen-
dent variables by ONE.

If z = 0, you are left with only x and y

ys

You may be able to solve the constraint for one variable

Ph

f (r1 , r2 , r3 , . . . , t) = 0 −→ x1 = g (y1 , z1 , r2 , r3 , . . . , t)

Then you can drop this variable. Sometimes you may have
to switch to a different set of variables. Suppose for a particle
on a sphere x2 + y 2 + z 2 = c2 a good choice is (θ, φ). This new
choice of variables are called generalised coordinates.

[email protected] 5 Physicsguide
©Sk Jahiruddin, 2020 Lagrangian Mechanics

Remember D’Alembert’s Principle which states that the

sum of work done due to forces of constrains is ZERO. So we
can forget the constrain forces.

ide
3 Generalize coordinates

gu
Any s quantities q1 , q2 , . . . , qs which completely define the po-
sition of a system with s degrees of freedom are called gen-
eralised co-ordinates of the system, and the derivatives q̇i are
called its generalised velocities.
ics
There are many ways to describe a system. Suppose a
pendulum can be described as
ys

x = L cos θ, y = −L sin θ, z = 0, θ = θ(t)

So the generalised coordinate is θ only.

Ph

4 Lagrangian

Do L = T − V is always? Nope. L = T − V is true only

when there is no velocity dependent potential is there. That
is, V = V (q, t)

[email protected] 6 Physicsguide
©Sk Jahiruddin, 2020 Lagrangian Mechanics

The Lagrangian follows the equation which called Lagrangian

equation of motion when these assumptions are hold

ide
ˆ Constraints are holonomic

ˆ Constraint forces do no work

ˆ Applied forces are conservative

gu
ˆ Potential V does not depend on q̇j
ics
 
d ∂L ∂L
− =0 (4.1)
dt ∂ q̇j ∂qj
ys
Ph

[email protected] 7 Physicsguide
©Sk Jahiruddin, 2020 Lagrangian Mechanics

Example 4.1: In the figure below we show a box of

mass m sliding down a ramp(triangular box) of mass
M. The ramp moves without friction on the horizontal

ide
plane and is located by coordinate x1 . The box also slides
without friction on the ramp and is located by coordi-
nate x2 with respect to the ramp. Find the equation of
motion of both the variables x1 and x2

gu
ics
ys

Solution: The kinetic energy of the ramp TM is given by:

1
Ph

TM = M ẋ21
2
The coordinate of the box is x = x1 + x2 cos θ, and y = h −
x2 sin θ. Now calculate 21 m ẋ2 + ẏ 2 and see that the kinetic

energy of the box Tm is given by:

1
Tm = m ẋ21 + ẋ22 + 2ẋ1 ẋ2 cos θ


2

[email protected] 8 Physicsguide
©Sk Jahiruddin, 2020 Lagrangian Mechanics

The velocity of the box is obviously derived from the vector

sum of the velocity relative to the ramp and the velocity of
the ramp.
The potential energy of the system is just the potential

ide
energy of the box and that leads to:

U = −mgx2 sin θ

and finally the Lagrangian is given by:

1
gu
L (x1 , ẋ1 , x2 , ẋ2 ) = TM + Tm − U
1
ics
= M ẋ21 + m ẋ21 + ẋ22 + 2ẋ1 ẋ2 cos θ + mgx2 sin θ


2 2
The equations of motion are:
d ∂L ∂L d ∂L ∂L
= and =
ys

dt ∂ ẋ1 ∂x1 dt ∂ ẋ2 ∂x2

The x1 variable equation gives
Ph

d
[m (ẋ1 + ẋ2 cos θ) + M ẋ1 ] = 0
dt
The RHS is zero because the Lagrangian does not explicitly
depend on x1 . The quantity in the brackets is the total mo-
mentum in the horizontal direction which is a constant since
there are no forces on the system in this direction.
The x2 variable equation gives

[email protected] 9 Physicsguide
©Sk Jahiruddin, 2020 Lagrangian Mechanics

d
(ẋ2 + ẋ1 cos θ) = g sin θ
dt
Carrying out the time derivatives in both the equations

ide
leads to two linear equations in ẍ1 and ẍ2 . These two equa-
tions in two unknowns can be solved to yield:

−g sin θ cos θ
ẍ1 =
(m + M )/m − cos2 θ

gu and ẍ2 =
g sin θ
1 − m cos2 θ/(m + M )
The RHS of both of the above equations are constants - we
can easily find x1 (t) and x2 (t)
ics
Example 4.2: A simple pendulum consisting of a string
of length r and a bob of mass m that is attached to a
support of mass M. The support moves without friction
ys

on the horizontal plane as shown in the figure. Find the

Lagrangian and equation of motion.
Ph

[email protected] 10 Physicsguide
©Sk Jahiruddin, 2020 Lagrangian Mechanics

Solution: Look at the figure carefully the position coor-

dinates of the bob is xbob = x + r sin θ and ybob = r cos θ The x
component of the velocity of the bob is given by ẋ + rθ̇ cos θ
and the y component by rθ̇ sin θ. So the overall kinetic energy

ide
of the system is given by:
M 2 mh 2 2 2
i
T = ẋ + ẋ + r θ̇ + 2ẋrθ̇ cos θ
2 2

gu
The potential energy U = −mgr cos θ. hence the Lagrangian
is

1 2 1
 
ics
2 2
L(x, ẋ, θ, θ̇) = (m+M )ẋ + r θ̇ + 2ẋrθ̇ cos θ +mgr cos θ
2 2
Now apply the Lagrange’s equation to get
d
ys

[(m + M )ẋ + mrθ̇ cos θ] = 0

dt
and
d h  2
Ph

i
m r θ̇ + ẋr cos θ = −m(ẋrθ̇ + gr) sin θ
dt
The first equation (x variable equation ) leads to a con-
served quantity − the momentum along the x direction and
the second equation (θ variable equation) gives us:
ẍ g
θ̈ + cos θ + sin θ = 0
r r
[email protected] 11 Physicsguide
©Sk Jahiruddin, 2020 Lagrangian Mechanics

Example 4.3: Example: A mass is attached to a

spring which is fixed at a point. The mass can move
on a massless rail which moves about the fixed point of

ide
the spring with angular velocity α as shown in the fig-
ure. Calculate the Lagrangian, construct the equation of
motion and discuss the solution.

gu
ics
Solution: Transformation functions:
(
x = (l + r) cos αt
ys

y = (l + r) sin αt

Kinetic energy
Ph

m 2 m 2
ẋ + ẏ 2 = ṙ + (l + r)2 α2

T =
2 2
Potential energy
K 2
V = r
2
m 2 K
L= ṙ + (l + r)2 α2 − r2
2 2
[email protected] 12 Physicsguide
©Sk Jahiruddin, 2020 Lagrangian Mechanics

 
d ∂L ∂L
− = mr̈ − mα2 (l + r) + Kr = 0
dt ∂ ṙ ∂r
Solving the equation of motion

ide
 
d ∂L ∂L
− = mr̈ − mα2 (l + r) + Kr = 0
dt ∂ ṙ ∂r
or
mα2 l
  
2


mr̈ + K − mα r− =0
K − mα2
So ifr
ω=
gu
K > mα2 , the motion is of a harmonic oscillator with
K − mα2
m
If K < mα2 , particle moves away exponentially
ics
If K = mα2 , the velocity is constant
There are many solved examples of Lagrangian calculation
in various books, namely JC Upadhya and YK Lim’s Me-
ys

chanics book. Do some really interesting problems from these

books.
Ph

5 Non-uniqueness of Lagrangian: Gaug

Lagrangian is not unique for a given system. If a Lagrangian
L describes a system, then
dF (q, t)
L0 = L +
dt

[email protected] 13 Physicsguide
©Sk Jahiruddin, 2020 Lagrangian Mechanics

This new lagrangian L0 also describes the system equally well.

So that means if we add total time derivative of any function of
generalised coordinate and time with a Lagrangian the equa-
tion of motion remains unchanged. You can prove this very

ide
easily by
    
d ∂ dF ∂ dF dF ∂F ∂F
− = 0 using = q̇ +
dt ∂ q̇ dt ∂q dt dt ∂q ∂t

6
gu
Velocity dependent potential
If U = U (qj , q̇j , t) is a velocity dependent potential then gen-
ics
eralised force is is
 
∂U d ∂U
Qj = − +
∂qj dt ∂ q̇j
ys

Example 6.4: A particle is moving under the action of

1 + q̇
a generalized potential V (q, q̇) = . The magnitude
q2
of the generalized force is? [GATE 2011]
Ph

2(1 + q̇) 2(1 − q̇) 2 q̇

(a) (b) (c) (d)
q3 q3 q3 q3

Solution:
 
dU −2 dU 1
= (1 + q̇) ; = 2
dq q3 dq̇ q

[email protected] 14 Physicsguide
©Sk Jahiruddin, 2020 Lagrangian Mechanics

 
d dU 2 2
=− q̇; −→ Q=
dt dq̇ q3 q3

7 charged particle in em field

ide
Force on a charged particle in em field is

F = q[E + (v × B)]

gu
The velocity dependent potential is

U = qφ − qA · v
ics
Direction derivation of this is little difficult. But you can
easily verify (do it from Goldstein) that this potential leads
to the Lorentz force. The Lagrangian is
1
ys

L = mv 2 − qφ + qA · v
2

8 Hamilton’s Principle
Ph

Lagrange’s Equation can be derived from Newton’s Eqn us-

ing a “differential principle”, i.e, D’Alembert’s principle uses
infinitesimal displacements.
It’s also possible to do it with an “integral principle”, i.e,
Hamilton’s Principle.

[email protected] 15 Physicsguide
©Sk Jahiruddin, 2020 Lagrangian Mechanics

8.1 Configuration Space

Generalized coordinates q1 , . . . , qf fully describe the system’s
configuration at any moment.

ide
Now imagine a n dimensional space.
Each point in this space (q1 , . . . , qn ) corresponds to one
configuration of the system
Time evolution of the system → A curve in the configura-
tion space

gu
ics
ys

8.2 Action
Suppose a system is moving as qj = qj (t) j = 1..n where the
Ph

Lagrangian is L(q, q̇, t) = L(q(t), q̇(t), t)

Z t2
Integrate =⇒ I = Ldt (8.1)
t1

I is called ACTION. Action depends on the entire path from

t1 to t2 and the choice of coordinates qj does not matter. And
the action is invariant under coordinate transformation

[email protected] 16 Physicsguide
©Sk Jahiruddin, 2020 Lagrangian Mechanics

8.3 Hamilton’s Principle

Now the Hamilton’s Principle says the action integral of a
physical system is stationary for the actual path. First ques-

ide
tion is What does the term Stationary mean?
Consider two paths that are close to each other where the
difference is infinitesimal

gu
ics
ys

Stationary means that the difference of the action integrals

Ph

is zero to the 1 st order of δq(t)

Z t2 Z t2
δI = L(q + δq, q̇ + δ q̇, t)dt − L(q, q̇, t)dt = 0 (8.2)
t1 t1

δq(t) has to be ZERO ar the starting and the end points,

i.e, at t1 and t2

[email protected] 17 Physicsguide
©Sk Jahiruddin, 2020 Lagrangian Mechanics

You can get the Lagrange’s equation derived from the Hamil-
ton’s principle. The derivation is found any good book like
Goldstein, Landau, Upadhya etc.

ide
9 Symmetry and Conservation
Let’s consider a simple system where potential does not de-
pend on velocity.

gu
L=T −V =
X mi ẋ2i + ẏi2 + żi2

i
2


− V (xi , yi , zi , t)
ics
∂L
Momentum → = mi ẋi = pix ;
∂ ẋi
∂L ∂V
Force → =− = Fix
∂xi ∂xi
ys

Momentum pix conserved if V does not depend on xi . Now

we will generalize from here.
∂L
Let’s call pj ≡ the generalized momentum. It is also
∂ q̇j
Ph

known as canonical or conjugate momentum and equals to

usual momentum for simple x − y − z coordinates
dpj ∂L
Lagrange’s equation becomes − =0
dt ∂qj
pj is conserved if L does not depend explicitly on qj . Such
qi is called cyclic (or ignorable). Generalized momentum as-
sociated with a cyclic coordinate is conserved.

[email protected] 18 Physicsguide
©Sk Jahiruddin, 2020 Lagrangian Mechanics

Generalized momentum may not always look like linear

momentum. Dimension of it may vary, if qj is not a space
coordinate. But one thing is fixed. pj qj always has the di-
mension of action (= work × time ). Form of generalized

ide
momentum may vary if V depends on velocity. For example
in case of particle in e.m field.
1
L = mv 2 − qφ + qA · v =⇒ px = mẋ + qAx
2

gu
Extra term comes due to velocity dependent potential.
We can see if Lagrangian is invariant under space transla-
tion
ics
(x, y, z) → (x + ∆x, y + ∆y, z + ∆z)
Then the linear momentum is conserved. Similarly you can
prove that (See Landau/Goldstein for this) -
ys

Angular momentum is conserved if the system is symmetric

under rotation, i.e, the Lagrangian does not change if one
particular angle is changed.
Ph

We can generalize like this - the following four statements

are equivalent:

ˆ System is symmetric with respect to a generalized coordi-

nate.
ˆ The coordinate is cyclic (does not appear in Lagrangian).
ˆ The conjugate generalized momentum is conserved.

[email protected] 19 Physicsguide
©Sk Jahiruddin, 2020 Lagrangian Mechanics

ˆ The associated generalized force is zero.

Example 9.5: A particle moves in a potential V =

2
x2 + y 2 + z2 . Which component(s) of angular momentum

ide
is/are constant of motion? [NET Dec 2013]
(a) none (b) Lx , Ly & Lz (c) only Lx & Ly
(d) only Lz

gu
Solution: To handle these types of problems you need
to transform the coordinates in a suitable coordinate system
where you can clearly see the cyclic coordinates. Suppose in
this problem if you transform it into cylindrical coordinate
ics
x = ρ cos θ, y = ρ sin θ, z = Z, you will see that the potential
2 z2
becomes ρ + , hence we see that the conjugate momentum
2
corresponding to theta coordinate which is Lz is conserved.
But sometimes you may not be able to transform the co-
ys

ordinates. Then what will you do? Remember If you have

square all three Cartesian coordinates with same factor in the
potential then two angular momentum is conserved, if two
Ph

coordinates with same factor then one angular momentum is

conserved, and if all coordinates with different factors then no

[email protected] 20 Physicsguide
©Sk Jahiruddin, 2020 Lagrangian Mechanics

angular momentum is conserved. I write in mathematically.

V (x, y, z) = k(x2 + y 2 + z 2 )
=⇒ Two ang momentum is conserved

ide
V (x, y, z) = k1 (x2 + y 2 ) + kz 2
=⇒ One ang momentum Lz is conserved
V (x, y, z) = ky 2 + k1 (x2 + z 2 )
=⇒ One ang momentum Ly is conserved

gu
V (x, y, z) = kx2 + k1 (y 2 + z 2 )
=⇒ One ang momentum Lx is conserved
V (x, y, z) = k1 x2 + k2 y 2 + z 2
ics
=⇒ No ang momentum is conserved

Now my question is When V (x, y, z) = k(x2 + y 2 + z 2 ) One

angular momentum which is conserved is Lφ . You can clearly
ys

see it when transforming the potential into spherical polar

coordinates. What is the other angular momentum which
is conserved? Have you ever heard of Laplace-Runge-Lenz
Ph

Vector?

[email protected] 21 Physicsguide
©Sk Jahiruddin, 2020 Lagrangian Mechanics

10 Energy conservation and Hamil-

tonian

ide
Consider time derivative of Lagrangian
dL(q, q̇, t) X ∂L dqj X ∂L dq̇j ∂L
= + + (10.1)
dt j
∂q j dt j
∂ q̇ j dt ∂t

From here Using Lagrange’s equation you can derive

gu d X ∂L
dt j
q̇j
∂ q̇ i
!
−L +
∂L
∂t
=0 (10.2)
ics
 
P ∂L
Now define j q̇j − L as energy function h(q, q̇, t).
∂ q̇i

Does energy function represents total energy?

ys

Energy function equals to the total energy if

ˆ Constraints are time-independent
Ph

ˆ Kinetic energy T is 2nd order homogeneous function of the

velocities
ˆ Potential V is velocity-independent
More importantly Energy function is conserved if La-
grangian does not depend explicitly on time t.

[email protected] 22 Physicsguide