Documente Academic
Documente Profesional
Documente Cultură
Abstract
In my short essay on Kepler’s laws of planetary motion and Newton’s law of universal gravitation, the
trajectory of one massive object near another was shown to be a conic section. However, the frame of
reference was a coordinate system centered at one of the masses (typically the heavier). Unfortunately, since
even a heavier object will exhibit some acceleration toward the lighter one, such a frame cannot be regarded as
“inertial,” meaning that Newton’s laws are not entirely valid (without the introduction of “fictitious forces”).
The present treatment of eliminates this objection; however, relativistic effects are still not considered.
The force applied to each of the planets has, by Newton’s Law of Universal
Gravitation, magnitude
Gm1 m2
F = ,
r2
−→ −→
where r = | x1 − x2 | is the distance between the planets, and where G is
the universal gravitational constant, whose measured value in SI units is approx-
imately
G = 6.67 × 10−11 N · m2 /kg2 .
Since these forces are opposite each other, Newton’s Second Law immediately
implies that
d2 −→ d2 −→ d2 −→ −→
(m1 + m2 ) 2 R = m1 2 x1 +m2 2 x2 = 0 .
dt dt dt
This already says that the center of mass of the two-body system exhibits no
acceleration relative to the inertial frame of reference. Therefore the moving
center of gravity can itself be taken as an inertial frame of reference.
Now comes the math! Let’s gather together all the constants into a single
constant by defining
Gm32
K= ,
(m1 + m2 )2
leaving us with the task of solving the second-order differential equation
−→d2 −→ K
a = 2 R= 2. (1)
dt r
To deal with the above differential
−→ →
equation, we let r be the unit vector unit vector s in
−→ →
−→
Note: R=(rcos θ, rsin θ) d→
in the direction of R and let s be the direction of R
d −→ dt
unit vector in the direction of r. center-of-mass
−→ −→
dt
From r · r = 1, we get coordinate →
R
system mass m1
d −→ −→ θ
−→ d −→
0= r · r =2 r · r,
dt dt →
mass m2 unit vector r in
→
−→ −→
m2 → direction of R
which implies that r · s = 0.
- R
m1
−→
Next, from r = (cos θ, sin θ), get
Figure 3: Polar coordinates
d −→ dθ dθ −→
r = (− sin θ, cos θ) = s .
dt dt dt
−→ d −→ d −→ dr −→ d −→ dr −→ dθ −→
v= R= r· r = r +r r = r +r s . (2)
dt dt dt dt dt dt
The acceleration is the time derivative of Equation (2):
−→ d −→ d dr −→ dθ −→
a = v= r +r s
dt dt dt dt
2
d2 r −→ dr dθ −→ d2 θ −→ dθ −→
= r +2 s +r s −r r
dt2 dt dt dt2 dt
2
2 !
d2 θ −→
dr dθ −→ dr dθ
= −r r + 2 +r 2 s
dt2 dt dt dt dt
−→ K −→
However, recalling from Equation (1) that a = − 2 r , we extract two equations
r
of interest:
2
d2 r
dθ K
− r = − (3)
dt2 dt r2
and
dr dθ d2 θ
2 + r 2 = 0. (4)
dt dt dt
We can immediate extract useful information from Equation (4), namely that
d2 θ
d dθ dr dθ
r2 = 2r + r2 2 = 0.
dt dt dt dt dt
dθ
This means that the quantity L = r2 is a constant.
dt
Next, define the variable
L2
u= .
Kr
and that
dθ L K 2 u2
= 2= .
dt r L3
We have, using the Chain Rule, that
2
dr d L dθ L2 dθ du K du
= =− = − .
dt dθ Ku dt Ku2 dt dθ L dθ
Differentiate again and obtain
d2 r K d2 u dθ K 3 u2 d2 u
d dr dθ
= = − = −
dt2 dθ dt dt L dθ2 dt L4 dθ2
Substitute into Equation (3) and get
2 2 2 2
K 3 u2 d2 u L K u K 3 u2
− 4 − =− 4 .
L dθ2 Ku L3 L
K 3 u2
Dividing by the common factor of − 4 results in the inhomogeneous second-
L
order differential equation:
d2 u
+ u = 1.
dθ2
The general solution of this has the form
u = u(θ) = 1 + e cos(θ − θ0 ),
where e and θ0 are constants of integration which can be determined from the
initial conditions. In turn, we now know that the radial distance of Planet 1 from
the center of mass is given, in terms of θ, by
L2 L2
r= = , (5)
Ku K(1 + e cos(θ − θ0 ))
where the constants K and L are
Gm32 dθ
K= 2
, and L = r2 .
(m1 + m2 ) dt
The two-body problem will have in all four initial conditions: the two initial
positions of the planets and the two initial velocities. Correspondingly, there are
four constants of integration which occur: the initial position and velocity of the
center of mass relative to the original inertial frame, and the constants e and θ0
appearing in Equation (5). The motion of the center of mass is easily determined.
−→ −→
If x1 (0) and x2 (0) are the initial positions of Planets 1 and 2, respectively,
then the initial position of the center of mass is
−→ 1 −→ −→
R (0) = (m1 x1 (0) + m2 x2 (0)).
m1 + m2
−→ −→
Likewise, if v 1 (0) and v 2 (0) are Configuration at t=0
y
initial velocity of Planet 1
the initial velocities of Planets 1 and velocity of frame
(relative to
→
v moving frame)
2, then the combined system has total →
center-of-mass u1(0)
momentum (which is conserved) coordinate
−→ −→ −→ system mass m2 mass m1
p = m1 v 1 (0) + m2 v 2 (0),
→ x
u2(0)
and so the center of mass will move initial velocity of Planet 2
(relative to
with constant velocity moving frame)
−→ 1 −→
v= p .
m1 + m2
Figure 4: Initial conditions relative to moving frame
This means that relative to the moving
frame, the initial velocities of Planets 1 and 2 are
−→ −→ −→ −→ −→ −→
u1 (0) = v 1 − v and u2 (0) = v 2 − v ,
respectively. Finally, we choose the x-axis of our moving frame so as to pass
through the center of mass of Planet 1 at time t = 0. All of this is indicated in
Figure 4 above.
For the remainder of the section we shall be operating exclusively in the moving
inertial center-of-mass frame. Referring to Figure 4 we let r0 be the initial distance
−→
of Planet 1 from the center of mass. Next we let α be the angle that u1 makes
with the positive x-axis of the moving frame. Since at time t = 0 we have θ = 0,
Equation 5 at time t = 0 now reads
L2
r0 = . (6)
K(1 + e cos θ0 )
Equation (6) now can be written as
L2
e cos θ0 = − 1. (7)
Kr0
dr L2 e sin(θ − θ0 ) dθ
= ;
dt K[1 + e cos(θ − θ0 )]2 dt
at t = 0 this simplifies to
Kr02 e sin θ0 L Ke sin θ0
u0 cos α = − × = − ,
L2 r02 L
−→
where u0 is the magnitude of the vector u1 (0). Therefore,
Lu0 cos α
e sin θ0 = − (8)
K
Squaring and adding Equations (7) and (8) leads to
2 2
L L2 u20 cos2 α
e2 = −1 + (9)
Kr0 K2
Equation (9) can be written purely in terms of the initial data once we realize
that the constant L can be expressed as
dθ
L = r2 = r0 u0 sin α,
dt
giving rise to
2
r0 u20 sin2 α r02 u40 cos2 α
2
e = −1 + .
K K2
Since K is the constant
Gm21 m2
K= ,
(m1 + m2 )2
we see that the eccentricity of the orbit of Planet 1 (and hence of Planet 2)
about the moving center of mass can be expressed purely in terms of the initial
data r0 , u0 , the angle α, the masses m1 and m2 and the gravitational constant
G.
In the above, we have shown that each member planet of a two-body system
will orbit about the system’s center of mass in either an ellipse, a parabola, or
a hyperbola, depending on the eccentricity. A question that remains is whether
one planet’s orbit about the other will likewise be a conic section. If we can show
this, then Kepler’s First Law will be validated, namely that the planets in the
solar system orbit about the sun in elliptical orbits.
x2 x2
g(x) = 1- f(x) = - 1-
4 4
x
moving coordinate system centered at center of mass
Planet 1. We are to show that, in po-
L2
lar coordinates based at Planet 1, the r=
m2 K(1+ecos(θ−θ0))
orbit of Planet 2 is also a conic section
with the same eccentricity. If r is the
polar distance from the center of mass Figure 5: Coordinate System at Planet 1
to Planet 1, then in terms of the radial
angle θ, we have already shown that
L2
r= .
K(1 + e cos(θ − θ0 ))
Since the distance d between Planets 1 and 2 is given by
(m1 + m2 )r
d= ,
m2
it follows that the polar distance s from Planet 1 to Planet 2 must be (note
the minus sign!):
L2 (m1 + m2 )
s=− ,
Km2 (1 + e cos(θ − θ0 ))
which also describes a conic section with the same eccentricity e.