Orbital Mechanics

Lecture 6

Orbital Position as a Function

of Time
•Time Since Periapsis
•Circular Orbits (e = 0)
•Elliptical Orbits (e < 1)
•Parabolic Trajectories (e = 1)
•Hyperbolic Trajectories (e > 1)
•Universal Variables

1
 Time since periapsis
• The orbit formula, r = (h2/μ)/(1 + ecos θ), gives the position of body m2 in its
orbit around m1 as a function of the true anomaly
• For many practical reasons we need to be able to determine the position of m2
as a function of time
• For elliptical orbits, we have a formula for the period T:
• But, we cannot yet calculate the time required to fly
between any two true anomalies
• In this lecture we will come up with the formulas that allow us to do that
calculation
• The one equation we have which relates true anomaly directly to time is Eqn
2.47, ℎ = 𝑟 2 𝜃, which can be written

Substituting r = (h2/μ)/(1 + ecos θ), we find, after separating variables,

Integrating both sides of this equation yields

tp is the time at periapsis passage

2
where by definition θ = 0
 Time since periapsis
• tp is the sixth constant of the motion
• The origin of time is arbitrary
• It is convenient to measure time from periapsis passage, so, set tp = 0
• In that case the time versus true anomaly integral is

• The integral on the right may be found in any standard math handbook

3
 Circular orbits (e = 0)
𝜃
• If e = 0 the integral in Eqn 3.2 is simply 0
𝑑𝜗, which means
• Recall that for a circle (Eqn 2.62), r = h2/μ
• Therefore h3 = r3/2μ3/2, so that
• Finally, substituting the formula for the period T of a circular orbit,
• 𝑇 = 2𝜋𝑟 3/2 / 𝜇, yields

• The reason that t is directly proportional to θ in a circular orbit is simply that

the angular velocity 2π/T is constant
• Therefore the time Δt to fly through a true anomaly of Δθ is (Δθ/2π)T
• Because the circle is symmetric about any
diameter, the apse line—and therefore
the periapsis—can be chosen arbitrarily

4
 Elliptical orbits (e < 1)
• Set a = 1 and b = e in Eqn 3.3 to obtain

• Therefore, Eqn 3.2, , in this case becomes

• Let

• Me is called the mean anomaly

• The subscript e reminds us this is mean anomaly for the ellipse and not for
parabolas and hyperbolas, which have their own “mean anomaly” formulas
• Then,
5
 Elliptical orbits (e < 1)
• Mean anomaly versus true anomaly for ellipses of various eccentricities is
plotted in the following figure

• Observe that for all values of the eccentricity e, Me is a monotonically

increasing function of the true anomaly θ 6
 Elliptical orbits (e < 1)
• The formula for the period T of an elliptical orbit, we have
μ2 (1 − e2)3/2/h3 = 2π/T
so that the mean anomaly in Eqn 3.7 can be written much more simply as

• The angular velocity of the position vector of an elliptical orbit is not constant
• But since 2π radians are swept out per period T, the ratio 2π/T is the average
angular velocity
• This is given the symbol n and called the mean motion

• In terms of the mean motion, mean anomaly can be written simpler still as

• The mean anomaly is the azimuth position (in radians) of a fictitious body
moving around the ellipse at the constant angular speed n
• For a circular orbit, the mean anomaly Me and the true anomaly θ are 7
identical
 Elliptical orbits (e < 1)
• It is convenient to simplify Eqn 3.6 by introducing an auxiliary angle E called
the eccentric anomaly
• This is done by circumscribing the
ellipse with a concentric auxiliary
circle having a radius equal to the
semimajor axis a of the ellipse
• Let S be that point on the ellipse
whose true anomaly is θ
• Through point S we pass a
perpendicular to the apse line,
intersecting the auxiliary circle at
point Q and the apse line at point V
• The angle between the apse line and
the radius drawn from the center of the circle to Q on its circumference is the
eccentric anomaly E
• Observe that E lags θ from periapsis P to apoapsis A (0 ≤θ < 180°) whereas it
leads θ from A to P (180° ≤ θ < 360°)
8
 Elliptical orbits (e < 1)
• To find E as a function of θ, we first observe that,
in terms of the eccentric anomaly,
𝑂𝑉 = 𝑎 cos 𝐸
whereas in terms of the true anomaly,
𝑂𝑉 = 𝑎𝑒 + 𝑟 cos 𝜃, Thus,

• Using Eqn 2.72,

r = a(1 − e2)/(1 + ecos θ),
we can write this as

• Simplifying the right-hand side, we get

• Solving this for cos θ we obtain the inverse relation

• Substituting Eqn 3.10a into the trigonometric
9
identity sin2E + cos2E = 1 and solving for sin E yields
 Elliptical orbits (e < 1)

• Eqn 3.10a would be fine for

obtaining E from θ, except
that, given a value of cos E
between –1 and 1, there are
two values of E between 0 and 360°
• The same comments hold for Eqn 3.11
• To resolve this quadrant ambiguity,
we use the following
trigonometric identity
From Equation 3.10a
and
Therefore,
where the last step required
applying the trig identity in
Eqn 3.12 to the term (1 − cos θ)/(1 + cos θ) 10
 Elliptical orbits (e < 1)
• Finally, therefore, we obtain
or

• For any value of tan(E/2), there

is only one value of E between
0 and 360°
• There is no quadrant ambiguity
• Substituting Eqns 3.11 and 3.13b
into Eqn 3.6 yields Kepler’s equation

• This monotonically increasing relationship

between mean anomaly and eccentric
anomaly is plotted for several values of
eccentricity in the figure on the right
11
 Elliptical orbits (e < 1)
• Given the true anomaly θ, we calculate the eccentric anomaly E using Eqn 3.13
• Substituting E into Kepler’s formula, Equation 3.14, yields the mean anomaly
directly
• From the mean anomaly and the period T we find the time (since periapsis)
from Eqn 3.8, Finally, therefore, we obtain

• On the other hand, if we are given the time, then Eqn 3.15 yields the mean
anomaly Me
• Substituting Me into Kepler’s equation, we get the following expression for the
eccentric anomaly:
• We cannot solve this transcendental equation directly for E
• A rough value of E might be read from the graph of Me vs. E
• However, an accurate solution requires an iterative, “trial and error”
procedure
• Newton’s method, or one of its variants, is one of the more common and
efficient ways of finding the root of a well-behaved function
12
 Elliptical orbits (e < 1)
• To find a root of the equation f(x) = 0 Newton’s method
(see figure), we estimate it to be xi and
evaluate the function f(x) and its first
derivative f'(x) at that point
• We then extend the tangent to the
curve at f(xi) until it intersects the
x-axis at xi+1, which becomes our
updated estimate of the root
• The intercept xi+1 is found by setting
the slope of the tangent line equal to
the slope of the curve at xi, that is,

from which we obtain

• The process is repeated, using xi+1 to estimate xi+2, and so on, until the root has
been found to the desired level of precision 13
 Elliptical orbits (e < 1)
• To apply Newton’s method to the solution of Kepler’s equation, we form
the function
and seek the value of eccentric anomaly that makes f(E) = 0
Since
• For this problem Eqn 3.16 becomes
Algorithm 3.1
Solve Kepler’s equation for the eccentric anomaly E given the eccentricity e and
the mean anomaly Me
1. Choose an initial estimate of the root E as follows. If Me < π, then E = Me +
e/2. If Me > π, then E = Me − e/2. Note that angles E and Me are in radians.
2. At any given step, having obtained Ei from the previous step, calculate
𝑓 𝐸𝑖 = 𝐸𝑖 − 𝑒 sin 𝐸𝑖 − 𝑀𝑒 and 𝑓 ′ 𝐸𝑖 = 1 − 𝑒 cos 𝐸𝑖
3. Calculate ratio𝑖 = 𝑓 𝐸𝑖 /𝑓 ′ 𝐸𝑖
4. If |ratioi| exceeds the chosen tolerance (e.g., 10-8), then calculate an updated
value of E: Ei+1 = Ei − ratioi
Return to step 2.
5. If |ratioi| is less than the tolerance, then accept Ei as the solution to 14
within the chosen accuracy.
 Elliptical orbits (e < 1)
Example 1:
A geocentric elliptical orbit has a perigee radius of 9600
km and an apogee radius of 21,000 km. Calculate the
time to fly from perigee P to a true anomaly of 120°.
Solution:
Before anything else, let us find the primary orbital parameters
e and h.
The eccentricity is readily obtained from the perigee and apogee radii by means
of Eqn 2.84,

We find the angular momentum using the orbit equation, evaluated at perigee:

With h and e, the period of the orbit is obtained from Eqn 2.82,

15
 Elliptical orbits (e < 1)
Solution (cont’d):
Eqn 3.11a yields the eccentric anomaly from the true
anomaly,

Then Kepler’s equation, Eqn 3.14, is used to find the mean anomaly,

Finally, the time follows from Eqn 3.12,

16
 Elliptical orbits (e < 1)
Example 2:
In the previous example, find the true anomaly at three
hours after perigee passage.
Solution:
Since the time (10,800 seconds) is greater than one-half
the period, the true anomaly must be greater than 180°.
First, we use Eqn 3.12 to calculate the mean anomaly for t = 10,800 s.

Kepler ’s equation, E − esin(E) = Me (with all angles in radians) is then employed

to find the eccentric anomaly.
This transcendental equation will be solved using Algorithm 3.1 with an error
tolerance of 10-6.
Since Me > π, a good starting value for the iteration is E0 = Me − e/2 = 3.4166.
Executing the algorithm yields the following steps:

17
 Elliptical orbits (e < 1)
Solution (cont’d):
Step 1:
E0 = 3.4166  f(E0) = −0.085124 and f'(E0) = 1.3585
−0.085124
ratio = = −0.062658, ratio > 10−6 , so repeat
1.3585
Step 2:
E1 = 3.4166 − (−0.062658) = 3.4793  f(E1) = −0.0002134 and f'(E1) = 1.3515
−0.0002134
ratio = = −1.5778 × 10−4 , ratio > 10−6 , so repeat
1.3515
Step 3:
E2 = 3.4793 − (−1.5778 × 10−4 ) = 3.4794  f(E2) = −1.5366 × 10−9
and f'(E2) = 1.3515
−1.5366×10−9
ratio = = −1.137 × 10−9 , ratio < 10−6 , so accept E = 3.4794 as the
1.3515
solution
With this value of the E, the true anomaly is found from Eqn 3.13a:

18
 Elliptical orbits (e < 1)
Example 3:
Let a satellite be in a 500 km by 5000 km orbit
with its apse line parallel to the line
from the earth to the sun, as shown.
Find the time that the satellite is in the
earth’s shadow if: (a) the apogee is towards the sun;
(b) the perigee is towards the sun.
Solution:
We start by using the given data to find the primary orbital parameters, e and h.
The eccentricity is obtained from Eqn 2.84

The orbit equation can then be used to find the angular momentum

The semimajor axis may be found from Eqn 2.71,

19
 Elliptical orbits (e < 1)
Solution (cont’d):
The period of the orbit follows from Eqn 2.83

(a) If the apogee is towards the sun,

then the satellite is in earth’s shadow between points a
and b on its orbit.
These are two of the four points of intersection of the orbit with the lines that are
parallel to the earth-sun line and lie at a distance RE from the center of the earth.
The true anomaly of b is therefore given by sin θ = RE/r , where r is the radial
position of the satellite.
It follows that the radius of b is:
From Eqn 2.72 we also have
Substituting (e) into (f), collecting terms and simplifying yields an equation in θ,

20
 Elliptical orbits (e < 1)
Solution (cont’d):
Substituting (a) and (c) together with
RE = 6378 km into (g) yields

This equation is of the form

It has two roots, which are given by

For apogee towards the sun, the flight from perigee to point b will be in shadow.
To find the time of flight from perigee to point b, we first compute the eccentric
anomaly of b using Equation 3.13b:

21
 Elliptical orbits (e < 1)
Solution (cont’d):
From this we find the mean anomaly using
Kepler’s equation,

Finally , Eqn (3.5) yields the time at b,

The total time in shadow, from a to b, during

which the satellite passes through perigee, is
(b) If the perigee is towards the sun, then the satellite is in shadow near apogee,
from point c (θc = 143.36°) to d on the orbit.
Following the same procedure as above, we obtain:

The total time in shadow, from c to d, is:

22
 Elliptical orbits (e < 1)
Series Solutions to Kepler’s Equation
• We have observed that there is no closed-form solution for the eccentric
anomaly E in Kepler’s equation, E − esin(E) = Me
• However, there exist infinite series solutions
• One of these, due to Lagrange (Battin, 1999), is a power series in the
eccentricity e,

where the coefficient an is given by the somewhat intimidating expression

• Here, floor(x) means x rounded to the next lowest integer

[e.g., floor(0.5) = 0, floor(π) = 3]
• If e is sufficiently small, then the Lagrange series converges
• Unfortunately, if e exceeds 0.6627434193, the series diverges, which means
taking more and more terms yields worse and worse results for some values
of M – Laplace Limit 23
 Elliptical orbits (e < 1)
Series Solutions to Kepler’s Equation
• In practice, we must truncate the Lagrange series to a finite number of terms
N, so that

• For example, setting N = 3 and calculating each an by means of Eqn (3.19)

• For small values of the eccentricity e, this

yields good agreement with the exact solution
of Kepler’s equation
• However, as we approach the Laplace limit,
the accuracy degrades unless more terms
of the series are included
• Eqn (3.21) (N = 3) yields a solution that
oscillates around the exact solution but is
fairly close to it everywhere
• Setting N = 10 produces a curve that, at the
24
given scale, is indistinguishable from the exact solution
 Elliptical orbits (e < 1)
Series Solutions to Kepler’s Equation
• On the other hand, for an eccentricity of 0.90, far above the Laplace limit, Eqn
(3.21) is a poor approximation to the exact solution, and using N = 10 makes
matters even worse

25
 Elliptical orbits (e < 1)
Series Solutions to Kepler’s Equation
• Another infinite series for E (Battin, 1999) is given by

where the coefficients Jn are Bessel functions of the first kind, defined by

• It turns out that, unlike the Lagrange series,

the Bessel function series solution converges
for all values of the eccentricity less than 1

• Truncated Bessel series solutions:

26
 Elliptical orbits (e < 1)
• We can combine Eqn (3.10) and Eqn (2.72) as follows to obtain the orbit
equation for the ellipse in terms of the eccentric anomaly:

• From this it is easy to see that

• Earlier, we defined the true-anomaly-averaged radius 𝑟𝜃 of an elliptical orbit

• Alternatively, the time-averaged radius 𝑟𝑡 of an elliptical orbit is defined as

• According to Eqns (3.12) and (3.14),

• Upon using this relationship to change the variable of integration from t to E

and substituting Eqn (3.25), Eqn (3.26) becomes
27
 Elliptical orbits (e < 1)

• Comparing this result with Eqn (2.87) reveals, as we should have expected (?),
that 𝑟𝑡 > 𝑟𝜃
• In fact, combining Eqn (2.87) and Eqn (3.27) yields

28
 Parabolic Trajectories (e = 1)
• For the parabola, Eqn (3.2) becomes

• Set a = 1 and b = e in Eqn (3.4) yields

• Therefore , Eqn (3.29) may be written as

where
• Mp is dimensionless, and it may be thought of as the “mean anomaly” for the
parabola (parabolic mean anomaly)
• Eqn (3.30), which plays the role of
Kepler’s equation for parabolic
trajectories, is also known as
Barker’s equation

29
 Parabolic Trajectories (e = 1)
• There is no “eccentric anomaly” for the parabola
• Given the true anomaly θ, we find the time directly from Eqns 3.30 and 3.31
• If time is the given variable, then we must solve the cubic equation

which has but one real root, namely,

Example :
A geocentric parabola has a perigee velocity of 10 km/s. How far is the satellite
from the center of the earth six hours after perigee passage?
Solution:
The first step is to find the orbital parameters e and h.
We know that e = 1 for a parabola.
To get the angular momentum, we can use the given perigee speed and Eqn 2.90
(the energy equation) to find the perigee radius, 30
 Parabolic Trajectories (e = 1)
Solution (cont’d):

The angular momentum is

We can now calculate the parabolic mean anomaly by means of Eqn (3.31),

Therefore, 3Mp = 20.321 rad, which, when substituted into Eqn (3.32), yields the
true anomaly

Finally, we substitute the true anomaly into the orbit equation to

find the radius,

31
 Hyperbolic Trajectories (e > 1)
• Setting a = 1 and b = e in Eqn (3.5) yields

• Therefore, for the hyperbola, Eqn (3.1) becomes

• Multiplying both sides by (e2 – 1)3/2, we get

where

• Mh is the hyperbolic mean anomaly

32
 Hyperbolic Trajectories (e > 1)
• Eqn (3.33) is plotted in the figure below:

• Recall that θ cannot exceed θ∞

33
 Hyperbolic Trajectories (e > 1)
• We can simplify Eqn 3.33 by introducing an auxiliary angle analogous to the
eccentric anomaly E for the ellipse
• Consider a point on a hyperbola whose polar coordinates are r and θ
• Let x be the horizontal distance of the point from
the center C of the hyperbola, and let y be its
distance above the apse line
• The ratio y/b defines the hyperbolic sine of the
dimensionless variable F that we will use as the
hyperbolic eccentric anomaly
• That is, we define F to be such that:

• In view of the equation of a hyperbola

• it is consistent with the definition of sinh F to

define the hyperbolic cosine as
34
 Hyperbolic Trajectories (e > 1)
• It should be recalled that sinh x = (ex – e−x)/2 and cosh x = (ex + e−x)/2 and,
therefore, that cosh2 x − sinh2 x = 1
• From the figure we see that y = r sin θ
Substituting this into Eqn 3.35, along with
r = a(e2 – 1)/(1 + ecos θ) (Eqn 2.104)
and
(Eqn 2.106)
we get

This can be used to solve for F in terms of the true

anomaly,

35
 Hyperbolic Trajectories (e > 1)
• Using the formula sinh−1 𝑥 = ln 𝑥 + 𝑥 2 + 1 , we can, after simplifying the
algebra, write Eqn 3.38 as

• Substituting the trigonometric identities,

and doing some more algebra yields

• Fortunately, but not too obviously, the numerator and the denominator in the
brackets have a common factor, so that this expression for the hyperbolic
eccentric anomaly reduces to

36
 Hyperbolic Trajectories (e > 1)
• Substituting Eqns (3.37) and (3.39) into Eqn (3.33)

yields Kepler’s equation for the hyperbola

37
 Hyperbolic Trajectories (e > 1)
• If we substitute the expression for sinh F, Eqn (3.37),
into the hyperbolic trig identity cosh2 𝐹 − sinh2 𝐹 = 1
A few steps of
algebra lead to

Then,

Solving this for cos θ,

we obtain the inverse relation,
• The hyperbolic tangent is found in terms of the hyperbolic sine and cosine by
the formula

• We can use the hyperbolic trig identity,

• Substituting Eqns (3.37) and (3.41a) into this formula and simplifying yields

38
 Hyperbolic Trajectories (e > 1)
• Eqn (3.42) holds for ordinary trig functions, too; that is,

• Therefore, Eqn (3.43) can be written

• This is a somewhat simpler alternative to Eqn (3.39) for computing eccentric

anomaly from true anomaly, and it is a whole lot simpler to invert:

• If time is the given quantity, then Eqn (3.40)—a transcendental equation—

must be solved for F by an iterative procedure
• To apply Newton’s procedure to the solution of Kepler’s equation for the
hyperbola, we form the function
and seek the value of F that makes f(F) = 0,
Eqn (3.16) becomes 39
All are in radians
 Hyperbolic Trajectories (e > 1)
Algorithm 3.2
Solve Kepler’s equation for the hyperbola for the hyperbolic eccentric anomaly F
given the eccentricity e and the hyperbolic mean anomaly Mh
1. Choose an initial estimate of the root F:
a. For hand computations read a rough value of F0 (no more than two
significant figures) from the graph of Mh vs. F (Figure 3.17) in order to
keep the number of iterations to a minimum
b. In computer software let F0 = Mh, an inelegant choice which may result in
many iterations but will nevertheless rapidly converge on today’s high
speed desktop and laptop computers
2. At any given step, having obtained Fi from the previous step, calculate
𝑓 𝐹𝑖 = 𝑒 sinh 𝐹𝑖 − 𝐹𝑖 − 𝑀ℎ and 𝑓 ′ 𝐹𝑖 = 𝑒 cosh 𝐹𝑖 − 1
3. Calculate ratio𝑖 = 𝑓 𝐹𝑖 /𝑓 ′ 𝐹𝑖
4. If |ratioi| exceeds the chosen tolerance (e.g., 10−8), then calculate an updated
value of F: Fi+1 = Fi − ratioi
Return to step 2.
5. If |ratioi| is less than the tolerance, then accept Fi as the solution to 40
within the chosen accuracy.
 Hyperbolic Trajectories (e > 1)
Example :
A geocentric trajectory has a perigee velocity of 15 km/s and a perigee altitude of
300 km. (a) Find the radius and the time when the true anomaly is 100°; (b) find
the position and speed three hours later.
Solution:
We first calculate the primary orbital parameters e and h.
The angular momentum is calculated from Eqn 2.31 and the given perigee data:

The eccentricity is found by evaluating the orbit equation,

r = (h2/μ)[1/(1 + ecos θ)], at perigee:

(a) Since e > 1 the trajectory is a hyperbola

Note that the true anomaly of the asymptote of the hyperbola is, according to
Eqn 2.97:

41
 Hyperbolic Trajectories (e > 1)
Solution (cont’d):
Evaluating the orbit equation at θ = 100° yields:

To find the time since perigee passage at θ = 100°, we first use Eqn (3.44a) to
calculate the hyperbolic eccentric anomaly

Kepler’s equation for the hyperbola then yields the mean anomaly,

The time since perigee passage is found by means of Eqn (3.34),

42
 Hyperbolic Trajectories (e > 1)
Solution (cont’d):
(b) Three hours later the time since perigee passage is:

The corresponding mean anomaly, from Eqn (3.34), is

We will use Algorithm 3.2 with an error tolerance of 10−6 to find the hyperbolic
eccentric anomaly F
Referring to Figure 3.17, we see
that for Mh = 40.69 and e =
2.7696, F lies between 3 and 4
Let us arbitrarily choose F0 = 3
as our initial estimate of F
Executing the algorithm yields
the following steps:

43
 Hyperbolic Trajectories (e > 1)
Solution (cont’d):
Step 1:
F0 = 3  f(F0) = −15.944494 and f'(F0) = 26.883397
−15.944494
ratio = = −0.59309818
26.883397
F1 = 3 − (−0.59309818) = 3.5930982
ratio > 10−6 , so repeat
Step 2:
f(F1) = 6.0114484 and f'(F1) = 49.370747
ratio = 0.12176134
F2 = 3.5930982 − (0.12176134) = 3.4713368
ratio > 10−6 , so repeat
Step 3:
f(F2) = 0.35812370 and f'(F2) = 43.605527
ratio = 8.2128052 × 10−3
F3 = 3.4713368 − (8.2128052 × 10−3 ) = 3.4631240
44
ratio > 10−6 , so repeat
 Hyperbolic Trajectories (e > 1)
Solution (cont’d):
Step 4:
f(F3) =1.4973128 × 10−3 and f'(F3) = 43.241398
ratio = 3.4626836 × 10−5
F4 = 3.4631240 − (3.4626836 × 10−5 ) = 3.4630894
ratio > 10−6 , so repeat
Step 5:
f(F4) =2.6470781 × 10−3 and f'(F4) = 43.239869
ratio = 6.1218459 × 10−10
F5 = 3.4630894 − (6.1218459 × 10−10 ) = 3.4630894
ratio < 10−6 , so accept F = 3.4631 as the solution
We substitute this value of F into Eqn (3.44b) to find the true anomaly

45
 Hyperbolic Trajectories (e > 1)
Solution (cont’d):
With the true anomaly, the orbital equation yields the radial coordinate at the
final time

The velocity components are obtained from Eqns (2.31) and (2.49)

Therefore, the speed of the spacecraft is

Note that the hyperbolic excess speed for this orbit is

46
 Hyperbolic Trajectories (e > 1)
Solution (cont’d):

The results of this analysis

are shown in the figure

47
 Hyperbolic Trajectories (e > 1)
• When determining orbital position as a function of time with the aid of
Kepler’s equation, it is convenient to have position r as a function of eccentric
anomaly F
• The orbit equation in terms of hyperbolic eccentric anomaly is obtained by
substituting Eqn (3.41b) into Eqn (2.104)

• This reduces to

48