Poincaré Section (Surface of Section)

Author: Tianjiang Shuo
Website: https://cislunarspace.cn

Definition

A Poincaré section (also: surface of section, SOS) is a lower-dimensional submanifold $\Sigma$ of phase space used to reduce a continuous flow to a discrete sequence of points. Classical definition: for an autonomous flow $\dot{\mathbf{x}}=\mathbf{f}(\mathbf{x})$ , $\mathbf{x}\in\mathbb{R}^N$ , choose an $(N-1)$ -dimensional hypersurface $\Sigma$ and record every crossing of $\Sigma$ in a fixed direction (one-sided section); the set of crossings is the Poincaré section plot. When the system has an energy integral (e.g. the Jacobi constant $C$ in CR3BP), the flow is confined to an energy surface, and combining that with the section reduces dimension by two: the planar CR3BP collapses to a 2D map, the spatial problem to a 4D map (Poincaré 1892; Hénon 1969; Haapala & Howell 2014).

The section is only the cut; the first-return map $P:\Sigma\to\Sigma$ induced on it is the Poincaré map. The two are sometimes conflated in the literature, but separating the section (a geometric object) from the map (a discrete dynamical system) keeps the concepts clear.

Three families of section construction

1. Configuration-hyperplane sections

The most common choice: $\Sigma$ is the plane on which one coordinate equals a constant. Typical choices in the CR3BP synodic frame:

$y=0$ section: records the state when the orbit crosses the primary–secondary line, usually one-sided via $\dot{y}>0$ or $\dot{y}<0$ to avoid double counting; well suited to displaying the $x$ – $\dot{x}$ phase plane.
$x=1-\mu$ section (the $x$ location of $P_2$ ): the de facto standard for $L_1$ / $L_2$ manifold patching and near-libration-point transfers in the Earth–Moon and Sun–Earth systems (Gómez et al. 2001; Haapala & Howell 2014). Allows 2D or 4D plots of $(y,z)$ -related variables.
$x=x_{L_i}$ section: taken at a libration point to study transit/non-transit trajectories through the gateways (Koon et al. 2000).

2. Event-defined sections

Defined by an event condition rather than a fixed coordinate plane. The most important is the periapse surface of section (Villac & Scheeres 2004; Paskowitz & Scheeres 2006):

\Sigma = \{\mathbf{x}\mid \dot{\rho}=0,\ \ddot{\rho}>0\},\qquad \rho=\sqrt{(x-x_{P_i})^2+y^2+z^2}

i.e. the distance from the spacecraft to a chosen primary $P_i$ is at a local minimum ( $\dot{\rho}=0$ ) with outward radial acceleration ( $\ddot{\rho}>0$ , ensuring periapse rather than apoapse). Physical advantage: the velocity at the crossing is tangential, providing the minimum-energy baseline for impulsive transfers and lunar-gravity-assist window analysis. The apoapse section takes $\ddot{\rho}<0$ . Haapala & Howell (2014) extended this to the spatial problem to classify transit trajectories and identify long-term lunar capture orbit families.

Common naming variants in Earth–Moon, Sun–Earth and planetary-moon engineering: perigee section (relative to Earth), perilune section (relative to the Moon), periapse/apoapse section (relative to an arbitrary primary), apogee section. These are the same construction with a different central body.

3. Pseudo-arclength hyperplane sections

To avoid numerical ill-conditioning when the flow runs nearly parallel to $\Sigma$ , a pseudo-arclength hyperplane may be used:

\Sigma = \{\mathbf{x}\mid (\mathbf{x}-\mathbf{x}_0)\cdot\mathbf{n}=0\}

The normal $\mathbf{n}$ is not fixed to a coordinate axis but adapted along a reference trajectory so that each crossing is well conditioned, reducing interpolation error; frequently used for higher-dimensional manifold section representations.

One-sided vs. two-sided; direction

Crossings are usually recorded in a fixed direction (e.g. $\dot{x}>0$ ), giving a one-sided section and avoiding duplicate symmetric points from a single periodic orbit. On a one-sided section, a periodic orbit corresponds to isolated discrete points; a quasi-periodic orbit to closed curves (torus trace); a chaotic orbit to a dense scatter-filled region (Arnol'd 1989; Wiggins 2003).

Numerical implementation

Integrator accuracy: long integrations need a high-order Runge–Kutta scheme (e.g. DOP853 with $10^{-14}$ relative/absolute tolerance) or a symplectic integrator, to prevent energy drift from contaminating the section.
Event detection: use a root finder (e.g. Brent) to pin the crossing time precisely, then interpolate the crossing state; coarse-step scanning introduces systematic error.
Grid scan: seed a uniform grid on $\Sigma$ , integrate each point, and classify by crossing behavior; a standard method for systematic identification of ballistic capture solutions and transfer seeds.
Energy stacking: overlay contours of varying Jacobi constant on a single section (e.g. $\theta_3=\pi/2$ ) to compose a map of all orbit families on one figure (Qiao et al. 2025).

Applications

Libration-point orbit family identification: on a $\theta_2=0$ or $\theta_3=\pi/2$ section, the Lyapunov, vertical Lyapunov, Lissajous, quasi-halo and halo (northern/southern) families show distinct geometric signatures; the section makes the bifurcation of halo orbits from Lyapunov orbits and the north/south halo symmetry directly legible (Qiao et al. 2025).
Manifold patching and transfer seeding: on the $x=1-\mu$ section, intersections of the $L_1$ unstable manifold and the $L_2$ stable manifold identify candidate heteroclinic/homoclinic connections (Gómez et al. 2001; Haapala & Howell 2014).
Capture and escape analysis: on a periapse section, the boundary between transit and non-transit points is carved out by the invariant manifold tubes asymptotic to Lyapunov orbits (Conley 1968; Koon et al. 2000).
Orbit identification and cataloguing: projecting an observed state onto a section map identifies the orbit family of an unknown spacecraft, like looking up a dictionary (Qiao et al. 2025).

References

Poincaré H. Les méthodes nouvelles de la mécanique céleste. Gauthier-Villars, 1892.
Hénon M. Numerical exploration of the restricted problem, V: Hill's case. Astronomy & Astrophysics, 1969, 1: 223–267.
Conley C C. Low energy transit orbits in the restricted three-body problem. SIAM J. Applied Math., 1968, 16(4): 732–746.
Koon W S, Lo M W, Marsden J E, Ross S D. Heteroclinic connections between periodic orbits and resonance transitions in celestial mechanics. Chaos, 2000, 10(2): 427–469.
Villac B F, Scheeres D J. On the concept of periapsis in Hill's problem. Dynamics & Control of Systems, 2004.
Paskowitz M E, Scheeres D J. Geometry of quasiperiodic orbits in the Hill problem. Celestial Mechanics and Dynamical Astronomy, 2006.
Gómez G, Llibre J, Martínez R, Simó C. Dynamics and Mission Design near Libration Points. Vol. II. World Scientific, 2001.
Haapala A F, Howell K C. Representations of higher-dimensional Poincaré maps with applications to spacecraft trajectory design. Acta Astronautica, 2014, 96: 23–46.
Qiao C, Long X, Yang L, et al. Orbital parameter characterization and objects cataloging for Earth-Moon collinear libration points. Chinese Journal of Aeronautics, 2025.