Chapter 1

Some Basic Concepts

1.1 Newtons law and inertial frames

The three laws of newton
First law: Every particle continues in its state of rest or uniform motion in a straight line
relative to an inertial reference frame, unless it is comepelled to change that state by
forces acting upon it
Second law: The time rate of change of linear momentum of a particle relative to an inertial
reference frame is proportional to the resultant of all forces acting upon that
particle and is collinear with and in the direction of the resultant force
Third law: If two particles exert forces on each other, these forces are equal in magnitude and
opposite in direction (action - reaction)
The first law defines an inertial reference frame. If one inertial reference frame is know a whole class
of inertial reference frames is known. These are non rotating frames which are in rest or in uniform
motion in a straight line with respect to the first inertial reference frame. This can be shown with the
Galilei transformations eq [1.1]. Assume that the first frame moves with a constant speed of V and that
the second frame moves with a constant velocity W with respect to the first frame. The time in frame
X’Y’Z’ differs with a constant T and for simplicity sakes it is assumed that at t0 the origins coincided.
The following holds: (also see fig 1.1)

r̄0 = r̄ − W̄ (t − t0 ) ; t0 = t + T (1.1)
dr̄ dr̄0
V̄ = ; V̄ 0 = 0 (1.2)
dt dt
differentiating eq [1.1] with respect to dt and substituting eq: [1.2]

dr̄0 dr̄ dt0

= − W̄ ; =1 (1.3)
dt dt dt
0
dr̄ dt
= V̄ − W̄ (1.4)
dt dt0
dr̄0
= V̄ 0 = V̄ − W̄ (1.5)
dt0
We defined that V̄ and W̄ were constant so also V̄ 0 has to be constant so X’Y’Z’ has to be a inertial
reference frame.
Newtons second law is:
d
F̄ = (mV̄ ) (1.6)
dt
From wikipedia: In mathematics and theoretical physics, an invariant is a property of a system which
remains unchanged under some transformation. It would be handy if this was also the case for Newtons
laws. We examine newtons second law in reference frame X’Y’Z’ and substitute eq: [1.5]

1
 d d d dm
F̄ = (mV¯0 ) = (m(V̄ − W̄ )) = F̄ = (mV̄ ) − W̄ (1.7)
dt dt dt dt
So Newtons second law is only invariant if dm
dt = 0 i.e. for a body of constant mass. For a rocket, which
has a varying mass m is considered to be the instantaneous mass and extra apparent forces needs to be
added to compensate. The solidification principle can be used to describe the rocket as a body with an
instantaneous mass m and the thrust as an external force.

1.2 Gravity force and potential

The gravity force of body m2 can be described as follows

m1 m2
F̄2 = m2 g2 = −G 3 r¯12 (1.8)
r12
m1
g2 = −G 3 r¯12 (1.9)
r12
ḡ2 is called the field strength and can simply be found by dividing both sides of eq [1.8] with m2 . The
field strength can also be found by taking the (negative) gradient of the potential field. (In astrophysics
it is assumed that the potential at infinity is zero so that U20 = 0). Furthermore we assume for the scaler
U2 eq [1.10]

m1 m1
U2 = −G + U20 = −G +0 (1.10)
r12 r12
   
¯ 2 = Gm1 ∇
ḡ2 = −∇U ¯ 1 = Gm1 d 1 , d 1 , d 1 = −Gm1
1 1 1
, ,
m1 m1
= −G 2 = −G 3 r̄12
r12 dx r12 dy r12 dz r12 X2 Y 2 Z2 r̄12 r12
(1.11)
The end of equation [1.11] is the same as equation [1.9] so apparently the potential at an arbitrary
distance r can be written as:
m1
U = −G (1.12)
r
eq [1.12] can be differentiated wrt the mass. To find the gravity field of a body this equation needs to
be integrated over the body. Only for spherical shells of constant mass density and spheres with a radial
symmetric mass density distribution there is a closed form analytical solution.
dm
dU = −G (1.13)
r
For a spherical shell the following holds (see also fig 1.3)

dm = (2πR sin θ)(Rdθ)tρ (1.14)

2
2πR tρ sin θdθ
dUP = −G (1.15)
r
M = 4πR2 tρ (1.16)
2 2 2
r = (R sin θ) + (l − R cos θ)
= R2 sin2 θ + l2 − 2Rl cos θ + R2 cos2 θ
= R2 (sin2 θ + cos2 θ) + l2 − 2Rl cos θ
= R2 + l2 − 2Rl cos θ (1.17)
2rdr = 2RL sin θdθ (1.18)
dr Rl sin θ Rl sin θ
= =√ (1.19)
dθ r R + l2 − 2Rl cos θ
2
√
R2 + l2 − 2Rl cos θ
dθ = dr (1.20)
Rl sin θ

2
substituting [1.14] into [1.13] gives [1.15]. Substituting [1.16] and [1.17] into [1.15] gives [1.21]. Substi-
tuting [1.17] into [1.18] gives [1.19].
Z π
1 sin θdθ
UP = − GM √ (1.21)
2 θ=0 R + l2 − 2Rl cos θ
2

Substituting [??] into [1.21] gives [1.22] for l< R (point inside sphere) and [1.23] for l > R (point outside
sphere).

Z r=l+R
1 GM 1 GM GM
Upinside = − dr = − (R + l − (R − l)) = − (1.22)
2 Rl r=R−l 2 Rl R
Z r=l−R
1 GM 1 GM GM
Upoutside = − dr = − (R + l − (l − R)) = − (1.23)
2 Rl r=R−l 2 Rl l

If a body has a radially symmetric density distribution it can be modeled as a series of spherical shells
with the same center. Giving each shell an index i with a mass Mi the following is valid
GX GMT
UPoutside = − Mi = − (1.24)
l i l

Looking back one can say that a force in an arbitrary direction, lets say x can be expressed as [1.25]

δUp
FPx = −mp (1.25)
δx
δ GM
FPxinside = mp =0 (1.26)
δx R
δ GM GMT mp
FPxoutside = mp =− (1.27)
δx l l2
So if the point lies inside the sphere the neto force is 0 and if it lies outside the sphere the force is directed
along l and can be calculated with [1.27].

It is of course also possible to write out [1.13] for arbitrary bodies. To come to a solution one can
do a series expension and truncated all higher order terms. Rearranging and simplifying gives [1.28]

GM 1G
U =− − 3 (A + B + C − 3D) (1.28)
Z l 2l
A= (y 2 + z 2 )dm (1.29)
Z
B= (x2 + z 2 )dm (1.30)
Z
C= (x2 + y 2 )dm (1.31)
Z
D= r2 sin2 θdm (1.32)

These terms might look familiar because A,B and C are the moment of inertia around the X,Y and Z
axis and D is the moment of inertia around the line OP (the line between the Origin of the reference
frame of the body and the particle P). For a sphere all the inertias are equal which gives equation [1.27]
again.

One can also simplify equation [1.28] for rotational symmetric bodies. For such bodies A=B and
D = A cos2 φ + C sin2 φ. This results in [1.34]

3
  
C −A
= J2 ≈ 1.086 · 10−3 (1.33)
M l2 earth
GM C −A GM
U =− [1 − (3 sin2 φ − 1)] = − [1 − J2 (3 sin2 φ − 1)] (1.34)
l M l2 l
This is solved by me, and not in the reader. Please confirm yourself
Per definition the force in the direction of l, so the find it, you only have to differentiate with respect to l

dU d 1
Fl = −m2 = m2 GM l−2 + J2 R2 (3 sin2 φ − 1)l−3 (1.35)
dl dl
 2 
GM 3 2 GM 2
= m2 − 2 − J2 R 4 (3 sin φ − 1) (1.36)
l 2 l
"  2 #
GM m2 3 R
= 1 + J2 (3 sin2 φ − 1) (1.37)
l2 2 l

hmm This is incorrect because the final answer should be:

"  2 #
GM m2 3 R 2
Fl = 1 + J2 (3 cos φ − 1) (1.38)
l2 2 l

1.3 rocket maneuvering

When using an impulsive shot the following is valid:

∆H̄ = r¯0 × V¯1 − r¯0 × V¯0 = r¯0 × ∆V̄ (1.39)

1 1 ¯
1
∆E = (V12 − V02 ) = (V0 + ∆v̄) · (V¯0 + ∆v̄) − V02 = (∆V )2 + V¯0 · ∆V̄ (1.40)
2 2 2
From these equation the following can be determined:

• For a given ∆V the maximum change in orbital angular momentum (H) is achieved if the impulsive
shot is executed when the spacecraft is farthest away from Earth and of ∆V̄ is perpendicular to
r¯0 .
• If the direction of the orbital angular momentum vector should bot be changed, ∆V̄ should be
directed in the initial orbital plane and vice versa

• The maximum change in (total) orbital energy is achieved if the impulsive shot is executed at the
point in the orbit where the velocity reaches a maximum value, and if ∆V̄ is directed along the
velocity vectorV¯0
Most rockets burns are not impulse shot so Tsiolkovski’s law has an extra term for the gravity losses
(second right hand term)
Z te
M0
∆V = ve ln − g sin γdt (1.41)
Me t0

4
Chapter 2

Many-body problem

X d2 r
Second law of Newton : F̄ = m · (2.1)
dt2
mi mj
Gravity law of Newton : Fij = G 3 r¯ij (2.2)
rij
2
d ri X mi mj
EoM of body i : mi 2 = G 3 r¯ij (2.3)
dt rij
j6=i

Potential field
 
X mi mj
g = −∇U = −∇  G  + ∇Ui0 (2.4)
rij
j6=i
X X X X X mj
−1
= −G mj ∇ rij =G mj r¯ij −2 = G 3 r¯
ij (2.5)
rij
j6=i j6=i j6=i j6=i j6=i

This forcefield is not conservative because when looking to a point fixed in an inertial frame the potential
field will differ because the bodies j move in time. The sum of the potential and kinetic and potential
energy of body mi is not constant. (Part its energy will flow into the bodies j and vice versa)

With the use of a auxiliary variable and differentiation the next equation can be found:
X1 1 X X mi mj
mi Vi2 − G =C (2.6)
i
2 2 i
rij
j6=i

Or in short (watch the sign change and definition of Ep )

Ek + EP = C (2.7)

When taking a closer look one can see that as two bodies approach each other very closely ,rij → ∞, the
potential energy goes to infinity and the only solution is that the speed of one of the bodies also goes to
infinity (Vi → ∞). Or in other words, at least one body reaches escape velocity (ri → ∞).

By using steiners rule for the moment of inertia, differentiating twice using an auxiliary variable and
substituting previous equations the following can be found:

dI X
=2 mi r¯i · V̄i (2.8)
dt i
d2 I
= 4Ek + 2EP (2.9)
dt2

5
 2
To have a stable system it is necessary that ddt2I < 0 or else the moment of inertia at the end will grow
to infity, meaning that at least one body is a infinity distance which is not of course not a stable system.
Substituting 2.7 in 2.9 gives:
d2 I
= 4Ek + 2(C − Ek ) = 2C + 2Ek (2.10)
dt2
So to have a stable system C < 0 but one also has to look further. To determine if a system is really
stable we integrate 2.9 over a long period of time, te , and take the average.

te te te
d2 I
Z Z Z
1 4 2
dt = Ek dt + EP dt (2.11)
te 0 dt2 te 0 te 0
  te
1 dI
= 4E¯k + 2E¯P (2.12)
te dt 0

substitute 2.8 gives:

2 X
[ mi r¯i · V̄i ]t0e = 4E¯k + 2E¯P (2.13)
te i
1 X
[ mi r¯i · V̄i ]t0e = 2E¯k + E¯P (2.14)
te i

Because in a stable system no collision or escapes happen the term between square brackets stays finite.
In the limit case (limte →0 ) the left hand side of the equation will become zero. Giving:

2E¯k + E¯P = 0 (2.15)

1
E¯k = − E¯P (2.16)
2
Also the following holds when substituting 2.7 in 2.15

2E¯k + (C − Ek ) = 0 (2.17)
Ek = −C (2.18)

6
Chapter 3

Three body problem

3.1 Equations of motion

When describing the motion of a body while there are only two other bodies the EoM reduces to:

d2 r¯i mj mk
2
=G r¯ij + G r¯ik (3.1)
dt rij rik
When this set is written as an set of first order differential equations we have three bodies with each six
differential equations (dx dy and dz with respect to two other bodies) giving us a total of 18 unknowns,
or in other words of the order 18. By using the ten integrals of motion, the invariable plane of Laplace as
reference plane and a angular coordinate to replace time, the order can reduced with a factor 12. This
still leaves an order six differential equation to be solved. An other approach is by using center-of-mass
integrals, rewriting the equations in such a way that they are with respect to the center of mass of the
system and with respect to the center of mass of body P1 and P2 combined, Jacobi reduced this set to
an order 12. (which can be solved with the above mentioned tricks) The Jacobi set is:

d2 R̄
 
r¯13 r¯23
= −GM α 3 + (1 − α) 3 (3.2)
dt2 r13 r23
2
  
d r¯12 r¯12 r¯13 r¯23
= −G (m1 + m2 ) 3 + m3 3 − r3 (3.3)
dt2 r12 r13 23
m1
α= (3.4)
m1 + m2
R is the vector between the CoM of body P1 and P2 combined and P3 . For the Lunar case were P1 is
the Earth, P2 the Moon and P3 is the sun. And for the planetary case were P1 is the Sun, P2 is the
Earth and P3 and planet, the following holds:

lunar case: α≈1 ; r¯13 ≈ r¯23 ≈ R̄ (3.5)

m3
planetary case: α≈1 ; 1 ; r¯13 ≈ R̄ (3.6)
m1 + m2
Both set of approximations give the same simplified results:

d2 R̄ R̄
2
= −GM 3 (3.7)
dt R
d2 r¯12 r¯12
= −G(m1 + m2 ) 3 (3.8)
dt2 r12
The first equation describes the motion of CoM of body P1 and P2 with respect to P3 and the second
equation describes the relative motion of body P1 with respect to P2 . In these special cases the motion
of the bodies may be approximated by a superposition of two two-body trajectories. (The earth moon
system circles around the sun, and the moon circles around the earth)

7
3.2 Circular restricted three-body problem
By assuming that the mass of body P1 and P2 are much larger than the mass of P3 and by assuming
that the two massive bodies move in circular orbits around the CoM of the system equation 3.1 reduces.
When also a new (pseudo-)inertial reference frame is used this gives: (see also fig 3.4)

d2 r̄ m1 m2
=G r¯1 + G r¯2 (3.9)
dt2 r1 r2
By rewriting this equation for the movement of body P3 in a rotating reference frame with the same
origin and its X-axis alligned with the line r12 , assuming that the velocity body P3 in this reference
frame is 0 gives the equation for the surface of Hill. In this equation x and y are the coordinates in the
rotating reference frame and r1 and r2 are the distances between body P1 and P3 , and between P2 and
P3 . µ is a mass ratio and C is an intergration constant which is determined by the position and velocity
of body P3 at time t =0.

2(1 − µ) 2µ
x2 + y 2 + + =C (3.10)
r1 r2
m2
µ= (3.11)
m1 + m2
By looking at cross section in the XY plane for z=0 simplifies this equation to:

2(1 − µ) 2µ
r2 + + =C (3.12)
r1 r2
If we look for solutions when C is very large there are three possibilities: r2 is very large (so body P3
very far away from the CoM of the whole system), r1 is very small (Body P3 is very close to Body P1 )
and r2 is very small (Body P3 is very close to Body P2 ). If C becomes smaller there will be more possible
solutions, It is possible for body P3 to be located further from P1 and P2 . So the areas will grow and
the places where to areas meet for the first time a Lagrange point is located. (see fig 3.6)

A Lagrange libration point is place where centripetal forces on body P3 have the same magnitude and
opposite direction as the combined gravitational force of body P1 and P2 of body P3 . (Remember the
whole reference frame is rotating, Body P1 and P2 are still rotating around the CoM of the whole system,
and with them body P3 . Only the relative position of body P3 does not change). The equilibrium in
point L1 , L2 and L3 are not stable, point L4 and L5 are stable for certain values of µ and eccentricity.
Sidenote, a lissajous orbit is a quasi-periodic orbital trajectory and does not have to be round or elliptical
(think figure of eight and even more complex figures). However for realistic cases they can be considered
as slowly-changing elliptical paths. It can be shown that for the collinear libration points (L1 , L2 and
L3 ) the major axis is aligned with the Y axis (perpendicular to the line between P1 and P2 ). The orbit
direction is clockwise looking top down. For the equilateral libration points(L4 and L5 ) it can be shown
that major axis is rotated 30 degrees with respect to the X axis. Halo orbits are special lissajous orbits
around collinear libration points which are very long elongated but have fixed geometry and size and are
closed loop. (They are ”real” orbits).

Halo orbits are very handy if you want the have a fixed position with respect to the other body. The
Earth-Sun L1 point is multiple times used to study the sun. The Sun-Earth L2 point can be used if you
always want to be in the dark (cold), to study deep space. Furthermore the gravity gradient is very low
(gravity field is weak) so gravity distorsions are lower and higer pointing accuracy can be obtained (deep
space telescopes)

8
Chapter 4

Relative motion in the many-body

problem

In real life there are no real two body problems. There are always other bodies like, the sun, other
planets and even galaxies which have an influence on the trajectory. These forces are small compared to
the forces between the two main bodies and can be seen as disturbance forces. These forces are small
because the distance is very large or because the mass is relatively low. To derive a usefull equation we
start again with

d2 ri X mi mj
mi 2
=G 3 r¯ij (4.1)
dt rij
j6=i

Lets assume that there is one main body k, a body i and a disturbing body j. First step is the write
the EoM of body i and k with respect to an inertial reference frame. By substrating the EoM of body
k from body i, the equation of motion for body i in a non rotating (non inertial) with as center body k
can be found. (The EoM of body i with respect to body k)

mi mk X mi mj
mi r¯¨i = G 3 r¯ik + G ¯
rij (4.2)
rik rij
j6=i6=k
m k m i
X mk mj
¯
mk r¯
¨k = G 3 r¯ki + G rkj (4.3)
rki rkj
j6=i6=k
mk X mj
r¯
¨i = G 3 r¯ik + G ¯
rij (4.4)
rik rij
j6=i6=k
mi X mj
¯
r¯
¨k = G 3 r¯ki + G rkj (4.5)
rki rkj
j6=i6=k

using the following vector rules

r¯ik = −r¯ki ; r¯ki = r¯i − r¯k ; r¯ij = r¯j − r¯i = r¯kj r¯ki (4.6)

It can be simplified to

mk X mj
¯ − G mi r¯ki − G
X mj
¯
r¯
¨ki = r¯
¨i − r¯
¨k = G 3 r¯ik + G rij 3 rkj
rik rij rki rkj
j6=i6=k j6=i6=k
mi + mk X r¯kj − r¯ki r¯kj
= −G 3 r¯ki + G mj 3 − (4.7)
rki rij rkj
j6=i6=k

The subscript k can now be dropped because everything is described with respect to k, and the above
mention reference frame is used.

9
 mi + mk X r¯j − r¯i r¯j
r¯
¨ki = −G r¯i + G mj − (4.8)
ri3 3
rij rj
j6=i6=k

Assuming a satellite orbiting the earth the equation can be simplified because ms  mE . The magni-
tude of the main force and the disturbing force can now be written as.

mE
am = G (4.9)
ri2
 
¯
 id r¯j 
ad = Gmd  3 −  (4.10)
rid rj
s    s
¯
id r¯j ¯
id r¯j ¯
id ¯
id ¯
id d̄ d̄ d̄
= Gmd 3 − r
rid
·
r 3 − r = Gmd
r 3 · r3 − 2 r3 · r3 r3 · r3
j id j id id id d d d
s s
2
rid r2 2rd rid cos β 1 1 2 cos β
= Gmd 6 + d6 − = Gmd 4 + r4 − r2 r2
rid rd rd3 rid
3 rid d d id

(4.11)

used geometry rules

ā · b̄ = |a||b| cos β ; ā · ā = ||a||2 = |a2 | ; c2 = a2 + b2 − 2ab cos α (4.12)

Lets introduce γ = ri /rd and α which is the angle between r¯i and r¯d (β is the angle between r¯d and r¯i d)

2
rid = rd2 + ri2 − 2ri rd cos α (4.13)
A rd − ri cos α
cos β = = (4.14)
S rid
ri
γ= (4.15)
rd
s
1 1 2(rd − ri cos α)
ad = Gmd 2 2 + 4−
(rd + ri − 2ri rd cos α)2 rd rd2 rid
3
s
md 1 2(1 − γ cos α)
=G 2 +1− 5 2
rd (1 + γ 2 − 2γ cos α)2 rd (rd + ri2 + rd ri cos α)3/2
s
md 1 2(1 − γ cos α)
=G 2 1+ − (4.16)
rd (1 − 2γ cos α + γ 2 )2 (1 − 2γ cos α + γ 2 )3/2

THERE IS AN ERROR SOMEWHERE rd5 SHOULD BE rd6

In this equation k is the main body with mass mk  mi . α is the angle between the vector rki
and rkd and β is the angle between the vector rkd and rdi . (see fig 4.2). With this equation it can be
shown that bodies outside the solar system have a negligible influence on the trajectories around the
sun, and that the planets have a negligible influence on satellites circling the earth.

When flying further away from a planet the sun will become more dominant in play of forces. At
some point it can not be regarded any more as a disturbing force. The region were it can be regarded
as an disturbing force is called the sphere of influence of a planet. Beyond this sphere the planet will
become a disturbing force. These sphere is are relatively small. An interplanetary trajectory can be
modeled in three separate parts, making the analysis easier.

10
Chapter 5

Two-body problem

Remember these formulas:

p
r= (5.1)
(1 + e cos θ)
H2
p= (5.2)
µ
H =ϕ̇r2 = θ̇r2 (5.3)
In the pericenter θ = 0 and in the apocenter θ = π. By differentiating 5.5 en substitute 5.5 back into its
derivative and substituting 5.3 a simple function for ṙ can be found
p p
2a = rpericenter + rapocenter = rθ=0 + rθ=π = +
(1 + e cos 0) (1 + e cos π)
p p p + ep + p − ep 2p
= + = 2 2
=
(1 + e) (1 − e) 1 + ep − ep − e 1 − e2
p = a(1 − e2 ) (5.4)
2
H
p a(1 − e2 ) µ
r= = = (5.5)
(1 + e cos θ) (1 + e cos θ) (1 + e cos θ)
dr H2 r2 µ µ
ṙ = = (−(1 + e cos θ)−2 )(−e sin θ)θ̇ = H 2 eθ̇ sin θ = r2 θ̇ 2 e sin θ = e sin θ (5.6)
dt µ µ
H H

Combining 5.5 and 5.3 gives:

H2
µ r2 θ̇ H
µ
r= = (5.7)
(1 + e cos θ) (1 + e cos θ)
µ
rθ̇ = (1 + e cos θ) (5.8)
H
One can also write the speed in a component which is perpendicular to the radius vector Vn and a
component which is perpendicular to the axis of symmetry of the conic section. Vl . (see fig 5.8)
ṙ ṙ ṙ ṙ
Vn = rθ̇ + = rθ̇ + = rθ̇ + = rθ̇ + (5.9)
tan( π2 − α) tan( π2 − (θ − π2 )) tan(θ − π) tanθ
µ µ cos θ µ µ µ µ
= (1 + e cos θ) − e sin θ = + e cos θ − e cos θ = (5.10)
H H sin θ H H H H
ṙ ṙ µ 1 µ
Vl = = = e sin θ =e (5.11)
cos(θ − π2 ) sin θ H sin θ H
Formula 5.6 and 5.8 can be used to make velocity hedographs. For elliptical (e<1), parabolic (e=1) and
hyperbolic (e>1) trajectories different speeds are calculated. The pericenter is located at the top, where
the tangential velocity is the largest, the apogee at the bottom. See also fig 5.7

11
 Ellipse parabola Hyperbola
θ ṙ r2 θ̇ ṙ r2 θ̇ ṙ r2 θ̇
µ µ µ
0 0 H (1 + e) 0 2H 0 H (1 + e)
π µ µ µ µ µ µ
2 eH H H H eH H
µ µ
π 0 H (1 − e) 0 0 0 H (1 − e) < 0
3π µ µ µ µ µ µ
2 −e H H −H H −e H H
The Theorem of Whittaker states that at the pericenter Vn and Vl have to same direction and can
therefore be added to find the maximum speed. In the apocenter Vn is directed in the the opposite way,
and the minimum speed can be obtained by subtracting Vn from Vl .

µ µ µ
Vp = Vl + Vn = e + = (1 + e) (5.12)
H H H
µ µ µ
Va = Vl − Vn = e − = (1 − e) (5.13)
H H H

5.1 Relativistic effects

With the use of the auxiliary variable u =1/r and some rearranging equation 5.14 can be writtin as eq
5.16
µ
r̈ − rϕ̇2 = 2 (5.14)
r
1
u= (5.15)
r
d2 u µ
+u= 2 (5.16)
dϕ2 H
According to general theory of relativity a good approximation for the relativistic motion of a body can
be given by:

d2 u µ µ
2
+ u = 2 + 3 2 u2 (5.17)
dϕ H c
In these equation µ is the gravitation parameter of the large body, c is the speed of light, u = 1/r and H
is the classical angular momentum. Comparing the two equations it becomes clear that the most right
part of the right hand side of the equations is due to relativity. The right hand side can be rewritten
1
H = ϕ̇r2 = ϕ̇ (5.18)
u2 !
H 2 u2 ϕ̇2 (ϕ̇r)2 Vϕ2
     
µ µ µ µ µ µ
+ 3 2 u2 = 2 1+3 = 1+3 = 1+3 = 1+3 (5.19)
H2 c H c2 H2 u2 c2 H2 c2 H2 c2

The relativistic part depends on the ration between the velocity of the body and the speed of light. In
celestical mechanics the speed of the body is much smaller than the speed of light (Vϕ  c). Therefore
the ration will be small and therefore the right term will be much smaller than the left term. A first
order approximation is:

µ2
 
µ 1 2 1 2
u= [1 + e cos (ϕ − ω)] + α 1 + e + eϕ sin (ϕ − ω) − e cos 2(ϕ − ω) (5.20)
H2 H4 2 6
µ
α=3 2 (5.21)
c
The relativistic part of this equation consist of three parts:
• The constant part (1 + 1/2e2 ) which increase the value of u with the constant value αµ2 (1 +
e2 /2)/H 2 . Time has no influence on this part.
• eϕ sin ϕ − ω is a fluctuating term, of which the amplitude continuously increases with increasing
values of ϕ. Because ϕ is the total angle traveled, ϕ and this term will become larger in time.

12
 • 1/6e2 cos 2(ϕ − ω) is a pure oscillation with a constant amplitude.
So in the long run only the second term has influence and the equation can be simplified:
µ
u= [1 + e cos(ϕ − ω) + βeϕ sin ϕ − ω] (5.22)
H2
µ mu2
β=α 2 =3 2 2 (5.23)
H c H
In celestial mechanics the value of β is very small. It is so small that even for large values of ϕ the
combination of the two stays small. Therefore the following is valid:

cos βϕ ≈ 1 : sin βϕ ≈ βϕ (5.24)

With this known equation 5.22 can be even further simplyfied:

cos (x − y) = cos x cos y + sin x sin y (5.25)

µ
u = 2 [1 + e · 1 · cos(ϕ − ω) + e · ϕβ · sin (ϕ − ω)]
H
µ
= 2 [1 + e cos β$ cos (ϕ − ω) + e sin ϕβ sin (ϕ − ω)]
H
µ
= 2 [1 + e cos(ϕ − ω − βϕ)] (5.26)
H
In the cosine the instantaneous argument of the pericenter is given by ω +ωβ. Because this term depends
on ϕ the location of the pericenter will change over time. The amount of change after one full rotation,
∆ω, can be found by taking 2π for ϕ (one full rotation). This gives:

µ mu2
∆ω = β2π = α · 2π = 6π (5.27)
H2 c2 H 2
When looking to this equation the only term which differs for planets in our solar system is H. Assuming
that the orbits for the planets are more or less circular (P ≈ r) one can say that H 2 ≈ µr which shows
that the relativistic effects is the largest when the planet is closest to the sun, in other words Mercuries.

5.2 Solar radiation pressure

Photons emitted by the sun reflect on satellites, losing kinetic energy/ apply a force on the satellite.
This can be modeled as follows:

µ F
r̈ − rϕ̇2 = − + sin δ (5.28)
r2 m
d 2 F
(r ϕ̇ = r cos δ (5.29)
dt m
The solar radiation force can be written as
WS
F = CR (5.30)
c
CR is the reflectivity, c is the speed of light, W is the power density of the light falling on the object and
S is the effective surface. Because the surface area of a sphere is 4πr2 the power density dimmished with
a factor r2 . When the power density at the sun is know (WS ) the local power density can be found by
multiplying it with RS /r where RS is the radius of the sun and r is the distance from the center of the
sun to the object. Assuming the object is a perfect sphere, the effective area is a circle with a surface of
πR2 and a mass of 4/3πR3 ρ. This gives:

F RS 2 1 2 1 3 CR WS RS2 1 α
= C R WS ( ) πR 3
= = 2 (5.31)
m r c 4/3πR ρ 4 cρR r2 r

13
Because the body has a radial velocity component with respect to the sun, the frequency of the light is
shifted because of the Doppler effect. The power density is linear with the frequency (a lower frequency
means that the photons have less energy, so produce a smaller force) Therefore:

ṙ
v 0 = v(1 − ) (5.32)
c
0 ṙ
W = W (1 − ) (5.33)
c
Furthermore because also light needs time to travel the distance between the sun and the object, the
object already has rotated and it appears that the light comes more from the front.

r
∆t = (5.34)
c
ϕ̇r
γ = ϕ̇∆t = (5.35)
c
A bit of geometry and because γ is very small (small angle approach)

π π
sin δ = sin( + γ) = cos γ ≈ 1 ; cos δ = cos( + γ) = − sin γ ≈= −γ (5.36)
2 2
Substituting these answers back into 5.28 gives:

 
µ α ṙ
r̈ − rϕ̇2 = − + 1 − (5.37)
r2 r2 c
 
d 2 α ṙ rϕ̇
(r ϕ̇ = −r 2 1 − (5.38)
dt r c c

Rewriting and neglecting second order terms (c2 )

µ−α αṙ
r̈ − rϕ̇2 = − − 2 (5.39)
r2 cr
d 2 α$
(r ϕ̇ = − (5.40)
dt c
A couple of things can be noticed here. All particle will be slowed down because of equation 5.40. Due
to this drag force the object will spiral down in the direction of the sun. Equation 5.39 can be even
further simplified by neglected the second term.
 
µ α µ
r̈ − rϕ̇2 = − 2 1 − = − 2 (1 − β). (5.41)
r µ r
For the right values of α/µ (β > 1) particles will not spiral inwards but will be blown out of the solar
system. When taking a closer look to definition of α one can see that small, low density, high reflectivity
particles will have larger forces applied to them. This is why our sun is bright and not hazy because all
dust is blown away.

14
Chapter 6

Parabolic and hyperbolic orbits

The motion of a satellite in an elliptical orbit can be described with:

r
θ 1+e E
r = a(1 − e cos E) ; tan = tan ; E − e sin E = n(t − τ ) = M (6.1)
2 1−e 2
For a hyperbolic orbit:

r
θ 1+e F
r = a(1 − e cosh F ) ; tan = tanh ; e sinh F − F = n̄(t − τ ) = M̄ (6.2)
2 1−e 2
In these equations r is the radius, a is the semi-major axis, e is the eccentricity, θ is the true anomaly,
E is the eccentric anomaly, F is the hyperbolic anomaly, n is the mean angular motion, n̄ is an angular
velocity, t is the time, τ is the time of (last) pericenter passage, M is the mean anomaly and M̄ is also
a kind of mean anomaly. see fig 6.7 and 8.1

T µ T µ
n= = 3 ; n̄ = = (6.3)
2π a 2π −a3
To demonstrate how these equations work, the calculation of the orbital period T and the position (θ
and r) are calculated for a satellite orbiting the Earth in an orbit with e=0.05 and a perigee altitude of
hp = 500 km after 60 minutes. Furthermore given is that µ = 398600.4 km3 /s2 and R = 6378.14 km.

15
 a(1 − e2 )
r= (6.4)
1 + e cos θ
r(1 + e cos θ) (500 + 6378.14)(1 + 0.05 cos 0)
a= 2
= = 7204.036908 Km (6.5)
1−e 1 + 0.052
s r
a3 72043
T = 2π = 2π = 6085 s (6.6)
µ 398600.4
r r
µ 398600.4
n= = = 0.0010325355 s−1 (6.7)
a3 72043
t = 60 · 60 = 3600 < 6085 → τ = 0 (6.8)
M = n(t − τ ) = 0.0010325355(3600 − 0) = 3.717127 (6.9)
Ek=0 = M = 3.68935344 (6.10)
Ek+1 = M + e sin Ek (6.11)
E1 = 3.717127 + 0.05 sin 3.717127 = 3.689913439 (6.12)
E2 = 3.717127 + 0.05 sin 3.689913439 = 3.691064876 (6.13)
E3 = 3.717127 + 0.05 sin 3.691064876 = 3.691015761 (6.14)
E4 = 3.717127 + 0.05 sin 3.691015761 = 3.691017855 (6.15)
r = a(1 − e cos E) = 7204.036908(1 − 0.05 cos 3.691017855) = 7511.225Km (6.16)
r !
1 + 0.05 3.691017855
θ = 2 · arctan tan = −2.62 (6.17)
1 − 0.05 2
π + 2.64
θ= · 180 = 330◦ (6.18)
π
As can be seen in the above example, the eccentric anomaly is found with an iterative process (you van
also plot two lines with your GRM and find the intersection). For elliptical trajectories is possible to find
an analytical approximation by using a series expension. For the hyperbolic case this can not be done
because these series do not converge because 1 > e > ∞ and because the function sinhF and M̄ − F are
non-periodic.

Hyperbolic problems can be solved numerical or graphical. A graphical method is to plot the func-
tion sinhF and the linear function (F + M̄ )/e. The intersection of the two is the solution. The linear
function can be plotted by drawing a line through (−M̄ ,0) and (−M̄ + e),1), or use a GRM.

16