You are on page 1of 9

Vector Spaces

Cross Product 3D Cartesian Coordinates:

r s =< ry sz rz sy , rz sx rx sz , rx sy ry sx >
BAC - CAB:

a (b c) = b(a c) c(a b)
Matrix Operations:

Matrix Multiplication:

a b d e g h

c A f D i G

B E H

C aA + bD + cG F = dA + eD + f G I gA + hD + iG

aB + bE + cH dB + eE + f H gB + hE + iH

aC + bF + cI dC + eC + f I gC + hF + iI

a b c 1 a1 + b2 + c3 d e f 2 = d1 + e2 + f 3 g h i 3 g1 + h2 + i3
Determinant:

a b c det d e f = aei + bf g + cdh ceg bdi af h g h i

Newtonian Mechanics
Newton's Second Law

F =m r
In Cylindrical/Polar Coordinates:

Fr F Fz

= m( r2 ) r = m(r + 2r) = m z

Momentum
Rocket Motion

mv = mvex + F ext
Angular Momentum and Torque

=rp L=

Energy
T =
Conservative Forces

p2 1 = mv 2 2m 2

1. Depend only on position 2. The work done is independent of the path

U (r) = W (ro r) F= U
Kinetic Energy in Polar Coordinates:

T =

1 m(r2 + r2 2 ) 2

Calculus of Variations
x 2

f [y(x), y (x), x]dx


x1
Is stationary along the path

y = y(x)

if and only if:

d f f =0 y dx y

Lagrangian Mechanics
L=T U
Steps to solving for equations of motion:

1. Write down kinetic and potential energies, nding 2. Choose generalized coordinates set

L=T U

in an inertial frame

q i , , qn

and convert original coordinates to this generalized

3. Write the Lagrangian in terms of the generalized coordinates 4. Write down all of the Lagrange equations:

L d L = qi dt qi
Finally, solve the equations for all

qi .
2

Generalized Momentum:

pi =
If

L qi pi
is constant.

L qi

= 0,

then

qi

is ignorable and the generalized momentum

Hamiltonian:

H=
If

pi qi L H
is the

L t = 0, then H is conserved; if the generalized coordinates are time-independent, energy of the system:

H=T +U

Central Force Motion


Center of Mass

1 m r M
Or, for the two-body problem:

m1 r1 + m2 r2 =R m1 + m2 P = MR
is conserved, ie.

We know that for two bodies, total momentum

is constant.

The Center of Mass frame is therefore an inertial frame.


Generalized Coordinates in Center of Mass Frame

r1 = R +
where

m2 r M

and

r2 = R + 1 2.

m1 r, M

is the vector adjoining the two particles

We introduce the reduced mass:

=
to arrive at our kinetic energy:

m1 m2 m1 + m2

T =

1 2 1 2 M R + r 2 2

The kinetic energy describes the energy of the center of mass motion, plus the motion of the particles relative to the center of mass. This leads us to the Lagrangian:

1 L = T U = M R2 + 2
3

) 1 2 U (r) , r 2

or:

L = Lcm + Lrel
But R is ignorable, meaning that the center of mass moves with constant velocity. To solve

Lrel ,

we note that this looks like a single particle with mass

and position r:

= U (r) r
reducing the problem to a 1-body problem. This means that all previous equations work, replacing

with

Angular Momentum and the Equivalent One-Dimensional Problem

= r

d Ue (r), dr

with

Ue (r) = U (r) + Ucf (r) = U (r) +


given

2r2

, the angular momentum of the relative coordinate, dened as:

= r r
This reduces the problem to a one-dimensional problem with potential

Ue .

Orbital Motion

Using energy: See example 8.2 In terms of

: u () = u()

2 u()2

F,
1 u() .

1 r. To use this transformed equation, rst solve it for


where

is a central force (eg. gravity), and

u=

u(),

then solve

r() =

For example, the solution in Kepler Orbits (planetary, comet motion):

r() =
as in this case,

c , 1 + cos

F =

Gm1 m2 r2

r2

= u2 .
2

Note that:

c= 2 ( 2 2

Gm1 m2
2

, and

E=
eccentricity

1) =

(Gm1 m2 )2 ( 2 2

1),

which gives

Orbit shape can be determined as follows:


energy orbit shape

=0 0< <1 =1 >1

E E E E

<0 <0 =0 >0

circle ellipse parabola hyperbola

Virial Theorem

< T >=
with a potential proportional to

n <U > 2 T
and

rn ,

given time averages of

(in rotating frames, one

period is sucient). For example, in gravitation:

< T >=

1 <U > 2

Mechanics in Noninertial Frames


No Rotation

Finertial = mA
To account for a noninertial reference frame, we must consider
Rotation

F + Finertial .

= u
We dene an angular velocity vector with direction along the axis of rotation, For any rigid body with angular velocity

u.

about an axis through the origin,

v =r
Taking

to be an inertial frame, and

dQ dt

So

to be a noninertial frame,

( =

So

dQ dt

) + Q
S

In other words, to calculate the velocity of a vector Q in an inertial frame, we nd its velocity in a convenient noninertial frame and correct for its rotation. With this, we nd Newton's second law in a noninertial frame:

m = F + Fcor + Fcf r
with the Coriolis and centrigual forces, respectively, dened as:

Fcor = 2m r
Centrifugal Force

and

Fcf = m( r)

The centrifugal force on earth necessitates a slight adjustment in our free-fall acceleration,

g:

g = g0 + m( R)
There is about a 0.3% dierence in

between the poles and equator.

Coriolis Force

The Coriolis force is always perpendicular to the velocity of the moving object (use the RHR).

Rigid Body Motion


Center of Mass Motion

L=L

(motion of CM)

+L

(motion relative to CM)

T=T

(motion of CM)

+T

(motion relative to CM)

Rotation About a Fixed Axis

(z) L =< Ixz , Iyz , Izz >

If a body is axially symmetric about a certain axis, and is rotating about this symmetry axis, then its angular momentum will be in the same direction. If the angular momentum is along the axis of rotation, this axis is called a principle axis. If

z=0

is a plane of reection symmetry, then

Ixz

and

Iyz

will be 0.

Inertia Tensor

Ixx I = Iyx Izx

Ixy Iyy Izy

2 2 Ixz m (y + z ) = m y x Iyz m z x Izz Lx L = Ly Lz

m x y 2 m (x2 + z ) m z y x = y z

m x z m y z , 2 m (x2 + y )

with

and

gives:

L = I
Note that the sums can be replaced by integrals for continuous mass distributions. If two of the coordinate planes the o-diagonal elements will be 0. For a diagonal matrix, ie.

x = 0, y = 0, z = 0

are planes of reection symmetry, then all

1 I= 0 0

0 2 0

0 0 , 3

L = I = < 1 x , 2 y , 3 z >
which means that if

points along a coordinate axis, then

must point along the same axis.

Kinetic Energy of a Rotating Body

T =
Finding Principle Axes

1 L 2

To nd principle axes, solve the eigenvalue equation:

(Ixx ) Ixy Ixz =0 Iyx (Iyy ) Iyz det Izx Izy (Izz )
Then, for each

i,

plug in

to the above matrix and multiply:

(Ixx ) Ixy Iyx (Iyy ) Izx Izy


This will yield 3 components for each These eigenvectors

Ixz x y = 0 Iyz (Izz ) z


eigenvectors in total.

and 3

To normalize these vectors, divide by their length.

1 , 2 , 3

are the principle axes of the system. The moment of inertia

about one of these axes is the corresponding eigenvalue


Euler's Equations

1 , 2 , 3 .

With

L =< 1 x , 2 y , 3 z > in ( dL dt )
S

the rotating body (noninertial) frame,

+L=
in a noninertial frame xed in the rotating body.

This allows us to determine the evolution of

The components of this equation are as follows (note that

is taken in the body frame):

1 1 (2 3 )2 3 2 2 (3 1 )3 1 3 3 (1 2 )1 2
These are most useful when the torque

= 1 = 2 = 3

=0

(see section 10.8 for examples).

Hamiltonian Mechanics
qi =
given the denition of

H pi

and

pi =

H , qi

as before.

Steps to solving motion:

1. Write down the Lagrangian for the system, in terms of generalized coordinates 2. Find the generalized momenta 3. Use

qi .

pi .

pi

to nd

qi ,

then plug this into the Hamiltonian.

4. Use Hamilton's Equations (above) to nd (see example 13.4)

qi

and

pi .

Ignorable Coordinates

l is constant.

Any coordinate ignorable in the Lagrangian is ignorable in the Hamiltonian. Independence of

qi pi

Because the coordinate is constant, ignorable coordinates in the Hamiltonian reduce the degrees of freedom by the appropriate amount. That is, the two-dimensional problem:

H = H(q1 , q2 , p1 , p2 )
with

q2

ignorable reduces to a one-dimensional problem:

H = H(q1 , p1 , k)
where

is just a constant relating to initial conditions.

Poisson Brackets

[f, g]
And, specically,

N f g f g qi pi pi qi i=1

[f, H] = f
Some mathematical properties:

[g, f ] = [f, g] [a1 f1 + a2 f2 , g] = a1 [f1 , g] + a2 [f2 , g] [uv, g] = u[v, g] + v[u, g]


8

[qi , qj ] = 0

and

[pi , pj ] = 0

and

[qi , pj ] = ij

(1)

We can use the Poisson brackets to evolve the system step by step:

f (q, p, t + t) = f (q, p, t) + t[f, H]


This allows us to nd any physical quantity (H). There are more possibilites for Poisson transformations, however. For example, we can transform the Hamiltonian itself:

at any time

given the inital state of the system

[H, p1 ] =
Sometimes we may nd

H q1 q1 ,
ie.

for which

is independent of

q1

ignorable. We know that

[H, p1 ] = 0
for these ignored variables, and therefore,

[p1 , H] = [H, p1 ] = 0
So,

p1

is 0, and

p1

is a constant of the motion. This is a simple example of the power of the

Poisson bracket.

Canonical Variables

Any set of q,p that satisfy (1) is called canonical. A canonical transformation is one that maps canonical coordinates to a new set of canonical coordinates. That is,

q, p Q, P
We note that the Hamiltonian itself is a canonical transformation. Denote the sh bracket as follows: Why is this important?

[[f, g]] =
where

N f g f g , Qi Pi Pi Qi i=1
the result of performing a Poisson transformation on

Qi

and

qi

(similarly for

Pi Pi ).

are dened as

Qi (q, p),

The transformation is invariant:

[[f, g]] = [f, g]


for any physical quantities equivalent. We can therefore choose any canonical coordinates we wish in order to simplify the problem. We can apply any canonical transformations to them and still use them in the Poission bracket. For example, we can take a rotated frame and use its vectors in the Poisson brackets just the same way as we would for the non-rotated frame. Another example, we can use the Hamiltonian in the Possion brackets to step through the motion of a system (x(x0 , p0 , t + t) brackets at every point in the motion.

and

g.

In other words, any set of canonical variables q,p which are transformed to Q,P are physically

= x(x0 , p0 , t) + t[x, H]).

At each point in time, the transformed coordinates will be canonical, so we can use the Poisson

You might also like