SPA5304 Physical Dynamics Lecture 16

David Vegh
(figures by Masaki Shigemori)

12 February 2019

1 Symmetries and conservation laws

1.1 Energy conservation

Assuming the Lagrangian does not depend explicitly on time, we have seen that

H ≡ p~ · ~q˙ − L

is conserved. H is called the Hamiltonian.

• If L depends explicitly on t, then we ehave to be more careful with the derivation:

 
∂L ∂L ∂L d ∂L ∂L d
δL = L(~q + δ~q, ~q˙ + δ ~q˙, t + δt) − L(~q, ~q˙, t) = · δ~q + · δ ~q˙ + δt = · δ~q + δt = δF
∂~q ˙ ˙
∂ ~q |∂t
{z } dt ∂ ~q ∂t dt
extra term

Thus we get  
d ∂L ∂L
· δ~q − δF =− δt 6= 0
dt ∂ ~q˙ ∂t
or
dH ∂L
=−
dt ∂t
This means that changing parameters of the system injects/extracts energy.

1.2 The meaning of H

H is nothing but the energy is many cases. Let’s see some examples

(i) If ~q are Cartesian coordinates,

1 ˙2
T = m~r and p~ = m~r˙
2
Then  
1 ˙2 1 ˙2
H = p~ · ~q˙ − L = m~r˙ · ~r˙ − m~r − V = m~r + V = E
2 2

1
 (ii) If ~q are generalized coordinates and T is quadratic in q̇i
X1
T = aij (~q)q̇i q̇j
i,j
2

Euler’s theorem on homogeneous functions says1

X ∂T
q̇i = 2T
i
∂ q̇i

Hence,
∂L ˙
H = p~ · ~q˙ − L = · ~q − L = 2T − (T − V ) = T + V = E
∂ ~q˙

2 Rigid bodies
A rigid body is a mechanical model for solid bodies with finite size2 .
• It is represented by a system of particles such that the distances between the particles do not vary.

|~ri − ~rj | = Cij = const.

• We will often take the a continuum limit in which the number of particles is infinite.

• We want to study rigid bodies using Lagrangian mechanics. This involves:

(i) Finding the #DoFs, and identifying the generalized coordinates

(ii) Finding the kinetic energy T

(iii) Constructing the Lagrangian L = T − V

1 f (x) is a homogeneous function of x of degree s if f (ax) = as f (x). In this case, x ∂f
∂x
= sf .
2 “Finite” usually means non-zero (not infinitesimal) and not infinitely large either: somewhere in between.

2
2.1 Body-fixed frame
To describe the motion of a rigid body, let us introduce a coordinate frame that is fixed to the rigid body
and moves with it.

• SI is an inertial frame (laboratory frame)

• SII is the coordinate system fixed to the rigid body: this is non-inertial.
• The origin O0 of SII is not necessarily the COM of the rigid body.

ρ
~ ~ +
= |{z}
R ~r
|{z} |{z}
position position position
relative to O of O 0 relative to O 0

• Studying the motion of a rigid body is equivalent to studying the motion of SII with respect to SI .

2.2 The number of degrees of freedom

How many parameters do we need to specify the configuration of SII relative to SI ?

(i) The position of O0 relative to O is three parameters:

(ii) Rotating the axes of SII relative to those of SI gives another three parameters (angles)

3
 Therefore we need
6 parameters = 3 translations + 3 rotations

2.3 Fundamental formula of rigid kinematics

The infinitesimal change of the position P
ρ ~ + ~r
~=R
in an infinitesimal time dt is given by
d~ ~ + d~r
ρ = dR

~ is a translation of O0 relative to O.
• dR
• d~r is a rotation by angle dφ around a certain instantaneous axis n̂ passing through O0

and we have
d~r = n̂dφ × ~r
Plugging this in gives
d~ ~ + n̂dφ × ~r
ρ = dR
Dividing both sides by dt:

ρ ~˙ + ω
~˙ = R ~ × ~r

This is the fundamental formula of rigid kinematics.

•R ~˙ is the velocity of O0 relative to O.

~ is the angular velocity of SII about its origin O0 :

•ω
dφ
ω ≡ n̂
dt

• We have decomposed the motion of P as a combination of translational and rotational motions.

• The above equation is a vector equation. Components of the equation may be obtained by projecting
the vectors on the axes of either SI or SII .
• The formula is true for all points of the rigid body. For point Pi

ρ ~˙ + ω
~˙ i = R ~ × ~ri

and ω
~ is the same for all i. Thus, we can talk about the angular velocity of the rigid body.

4
2.4 Kinetic energy
1X 1X ~˙ + ω
T = ~˙ i 2 =
mi ρ mi (R ~ × ~ri )2
2 i 2 i
1X ~˙ 2 + 1
X X
~˙ · (~
= mi R ω × ~ri )2 +
mi (~ mi R ω × ~ri )
2 i 2 i i

1X ~˙ 2 + 1
X
~˙ · (~
X
= mi R ω × ~ri )2 + R
mi (~ ω× mi~ri )
2 i 2 i i
| {z }
cross term

Two situations in which the cross term vanishes:

(A) If O0 is fixed, e.g. a physical pendulum with suspension point O0 . Then,

1X
T = ω × ~ri )2
mi (~
2 i

(B) If O0 is the center of mass. In this case,

P
i mi~ri = 0.

1X ~˙ 2 + 1
X
T = mi R ω × ~ri )2
mi (~
2 i 2 i

2.5 Components
So far the expressions for T involved vectors and were valid in any frame.
Let us now introduce components of the vectors. The components depend on the reference frame.
   
ri1 xi
~ri =  ri2  =  yi 
ri3 zi


ω1
ω
~ =  ω2 
ω3

i, j, . . . denote particle number and run over 1 . . . N ,

a, b, . . . denote component number and run over 1, 2, 3 or x, y, z.
At this point we do not specify which frame we are referring to.

5
 ω × ~ri )2 . Let us write this quantity in component form. Suppressing the particle number i
T involves (~
and using the Einstein summation convention (i.e. repeated indices are summed over)

ω × ~r)2 = (~
(~ ω × ~r)a (~
ω × ~r)a = abc ωb rc ade ωd re

~ × B)
where we have used (A ~ i = ijk Aj Bk .

What is abc ade ? Note that

abc ade = 1bc 1de + 2bc 2de + 3bc 3de

The first term is non-vanishing iff

{b, c} = {d, e} = {2, 3}
namely if

(i) (b, c) = (d, e) = (2, 3) this gives +1

(ii) (b, c) = (d, e) = (3, 2) this gives +1

(iii) (b, c) = (e, d) = (2, 3) this gives −1

(iv) (b, c) = (e, d) = (3, 2) this gives −1

and similarly for the 2nd and 3rd terms. These can be summarized in

abc ade = δbd δce − δbe δcd

Using this result,

ω × ~r)2 = ωb rc ωb rc − ωb rc ωc rb = ω
(~ ~ 2~r2 − (~
ω · ~r)2

So for the two cases in which the cross term vanished:

(A)
1X
T = ω 2~r2 − (~
mi (~ ω · ~r)2 )
2 i

This is the rotational energy around the fixed point O0 .

(B)
1 ~˙ 2 1 X
T = MR + ω 2~r2 − (~
mi (~ ω · ~r)2 )
2 2 i
This is the kinetic energy of the COM plus the rotational energy around the COM.

6
2.6 Inertia tensor
Let us focus on the rotational part of T :

X X
ω 2~ri2 − (~
ω · ~ri )2 ) = mi ωa ωb δab~ri 2 − ria rib


mi (~
i i

ωa has no particle index i. Thus, this can be written as

" #
X
2


= mi δab~ri − ria rib ωa ωb ≡ Iab ωa ωb = ω
~ · I~
ω
i
| {z }
Iab

I is the moment of inertia tensor.

yi2 + zi2 −xi yi −xi zi

 
X X
mi δab~ri 2 − ria rib =


Iab = mi  −xi yi x2i + zi2 −yi zi 
i i −xi zi −yi zi x2i + yi2 ab

The inertia tensor is

• symmetric: Iab = Iba

P
• additive: I{mi } = i Imi

2.6.1 Expressions for the kinetic energy

So again for the two cases in which the cross term vanished:

(A) If O0 is fixed:
1
T = ~ · IO0 ω
ω ~
2
(B) If O0 is the COM:
1 ~˙ 2 1
T = MR + ω ~ · ICOM ω
~
2 2

2.6.2 Continuous version of I

If we regard the rigid body as a continuum,

X Z
2
d3 x ρ(~x) δab ~x2 − xa xb



Iab = mi δab~ri − ria rib →
| {z }
i
dm:mass element

Here ρ(~x) is the mass density function for the rigid body.

7
2.7 What frame do we use?
The vector ω and the tensor I are defined independently coordinate frames and we can compute their
components in any frame. This is easier in certain frames than in others:
• in SI the ~ri are time-dependent
• in SII the ~ri are time-independent, so we will pick this one.
Henceforth, x, y, z will be taken to be in the body-fixed frame SII . The inertia tensor becomes a purely
geometric quantity inherent to the rigid body. It depends on the mass distribution of the rigid body.