Overview
Orbital mechanics is Newtonian gravity applied to trajectories: six numbers (the Keplerian elements) describe any bound orbit, and everything else — transfers, rendezvous, perturbations — is built on them. This skill covers the element set, the vis-viva equation, Hohmann transfers, propagation with perturbations (J2, drag, third body), and the numerical habits that keep simulations honest.
When to use
- Designing a transfer: LEO to GEO, interplanetary porkchop plots, lunar trajectories
- Propagating a satellite orbit with realistic perturbations
- Interpreting TLEs and conjunction/rendezvous geometry
- Estimating delta-v budgets for a mission concept
- Debugging an orbit propagator: energy drift, wrong frames, unit errors
Core concepts
-
Keplerian elements: a (semi-major axis), e (eccentricity), i (inclination), Ω (RAAN), ω (argument of periapsis), ν or M (anomaly) — size, shape, and orientation; the anomaly locates the body along the orbit.
-
Vis-viva: v² = μ(2/r − 1/a) — the workhorse equation linking speed, position, and orbit size; escape velocity falls out at a → ∞.
-
Conservation: energy and angular momentum are conserved in the two-body problem — use them as invariants to validate every numerical propagation.
-
Hohmann transfer: the minimum-energy two-impulse transfer between coplanar circular orbits; Δv split between departure and arrival burns — optimal only for the coplanar circular case.
-
Perturbations: J2 (Earth's oblateness — nodal precession, the dominant LEO perturbation), atmospheric drag (altitude- and solar-cycle-dependent), third-body (Moon/Sun), solar radiation pressure — each dominant in a different regime.
-
Frames: ECI (inertial, for propagation) vs ECEF (rotating, for ground tracks) — mixing them is the classic propagator bug; be explicit at every interface.
-
Patched conics: interplanetary trajectories as linked two-body arcs (departure hyperbola, heliocentric ellipse, arrival hyperbola) — the screening tool behind porkchop plots; good to ~1% for preliminary design.
-
Lagrange points: equilibrium points of the circular restricted three-body problem — halos around Sun–Earth L1/L2 (SOHO, JWST) and Earth–Moon gateways; station-keeping costs are small but non-zero.
-
Gravity assists: stealing momentum from a planet's orbital motion — the trajectory design trick behind Voyager, Cassini, and Parker Solar Probe; v∞ in = v∞ out in the planet frame, direction changed.
Practical workflow
1. Set up the two-body baseline
import numpy as np
mu_earth = 398600.4418 # km^3/s^2
# Vis-viva: circular LEO at 400 km
r = 6378.0 + 400.0
v_circ = np.sqrt(mu_earth / r) # ~7.67 km/s
v_esc = np.sqrt(2 * mu_earth / r) # ~10.85 km/s
- Work in consistent units (km, s, km³/s² for Earth); convert TLE-era data carefully.
- Compute the baseline orbit (period from Kepler's third law: T = 2π√(a³/μ)) before adding perturbations.
2. Design maneuvers
- Hohmann: Δv₁ at departure (raise apoapsis), Δv₂ at arrival (circularize) — compute both, sum for the budget, add margin (3–10%) and plane-change costs (expensive: Δv = 2v sin(Δi/2)).
- Plane changes: do them at the lowest velocity (highest altitude) or combine with the arrival burn — never in LEO if avoidable.
- Interplanetary: patched conics for screening (departure v∞ from porkchop plots), then high-fidelity n-body for the real trajectory.
3. Propagate with perturbations
- Start with two-body + J2 (analytic secular rates for Ω, ω give quick lifetime/node-drift estimates).
- Add drag below ~600 km with an atmospheric model (NRLMSISE-00/JB2008) and realistic area-to-mass and solar-cycle assumptions — drag is the lifetime limiter.
- Use a proper integrator (DOP853 or IAS15-class) with tight tolerance; verify energy/angular-momentum conservation in the unperturbed case to machine precision before trusting perturbed runs.
4. Validate relentlessly
- Check invariants: two-body energy and angular momentum must be conserved to integrator tolerance.
- Test against known cases: GEO period = sidereal day, ISS-like decay rates, GPS semi-synchronous period.
- Propagate backward as well as forward — asymmetric errors reveal bugs.
5. Design a rendezvous
- Match planes first (or accept the plane-change cost), then phase: the chaser's orbit period sets the catch-up rate — phasing orbits trade time for delta-v.
- Compute the Lambert solution for the transfer arc between the two positions at the chosen times of flight.
- Budget station-keeping and proximity-operations fuel separately — rendezvous is not over at intercept; docking/berthing has its own delta-v and plume-impingement constraints.
Common pitfalls
- Frame confusion: ECI vs ECEF mixups produce ground tracks off by Earth's rotation — label every vector's frame.
- TLE misuse: TLEs are mean elements for SGP4 only — don't osculate them into a custom propagator without conversion.
- Ignoring drag uncertainty: atmospheric density varies 10× over the solar cycle — lifetime estimates need solar-activity scenarios, not a single number.
- Impulsive-burn fantasy: real burns are finite; long low-thrust spirals need entirely different math (Edelbaum, or direct optimization).
- J2 blindness: nodal precession (~5°/day at ISS altitude/inclination) dominates LEO geometry over weeks — sun-synchronous orbits exploit it deliberately.
- Unit errors: km vs m in μ is the oldest bug in astrodynamics — assert units at function boundaries.
- Two-body thinking in three-body regimes: near the Moon or at Lagrange points, the third body's gravity is not a perturbation — use CR3BP dynamics from the start.
- Impulsive-burn fantasy for low thrust: ion/electric propulsion spirals need months and continuous-thrust optimization (Edelbaum or direct methods) — Hohmann math does not apply.