Cauchy-Green Tensor Method

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

Definition

The Cauchy-Green tensor method uses the deformation gradient of a dynamical system's flow map — i.e. the state transition matrix (STM) $\Phi(t_f,t_0)$ — to build the finite-time Cauchy-Green strain tensor (CGST)

C(t_0,t_f) = \Phi(t_f,t_0)^{\!\top}\Phi(t_f,t_0).

$C$ is symmetric positive definite; its largest eigenvalue $\lambda_1$ gives, for integration time $\Delta t=t_f-t_0$ , the squared factor by which an initial small perturbation along the most-stretched direction is amplified. The finite-time Lyapunov exponent (FTLE) is

\sigma(t_0,t_f) = \frac{1}{\Delta t}\ln\!\sqrt{\lambda_1(C)}.

Evaluating $\sigma$ on a grid of initial conditions yields the FTLE field; ridges of high value in this field are Lagrangian coherent structures (LCS). LCS mark the most significant finite-time barriers separating regions of distinct flow behaviour — transit, capture, and escape trajectories are partitioned by LCS (Haller 2001; Shadden et al. 2005).

Distinction from Monodromy Matrices and Poincaré Sections

The boundary between CGST/FTLE/LCS and traditional CR3BP geometrical tools needs to be drawn carefully:

Monodromy matrix (monodromy matrix): the STM integrated over one full period of a periodic orbit; its eigenvalues serve local stability analysis of periodic orbits. CGST is built from the STM along any trajectory over finite time, periodicity not required.
Stable/unstable invariant manifolds (invariant manifold): spanned by the real eigenvectors of a periodic orbit's monodromy matrix; strictly autonomous geometric objects. LCS are "finite-time approximations" of transport barriers — when LCS are generated by integrating along a periodic orbit long enough, they align with the stable/unstable directions of its invariant manifolds; in non-autonomous or transient flows, LCS provide the working surrogate.
Poincaré section (Poincaré section): a dimension-reduction visualisation tool; FTLE fields are routinely built on two-dimensional Poincaré sections (section + integration time $\Delta t$ ), projecting the stretching of a 4D flow onto a 2D map.

In short: monodromy gives "is the orbit stable?", invariant manifolds give "the geometric corridors near a periodic orbit", and CGST/FTLE give "the finite-time transport barriers in any flow". The three are complementary.

Computation Pipeline

The standard recipe for an FTLE map on a CR3BP section $\Sigma$ (Canales & Howell 2024):

Lay an initial-condition grid $\{\vec q_i\}$ on $\Sigma$ (e.g. a $y\dot y$ section through $L_1$ or $L_2$ ).
Complete each $\vec q_i$ to a full state at the chosen energy $C_J$ and propagate forward (or backward) for time $\Delta t$ , integrating the STM alongside.
At $t_f$ form $\Phi(t_f,t_0)$ , build $C=\Phi^\top\Phi$ , and extract the largest eigenvalue $\lambda_1$ .
Compute $\sigma_i=(1/\Delta t)\ln\sqrt{\lambda_1}$ .
Render $\sigma_i$ on the grid as a colour map or contour; the ridges are LCS.

Backward integration yields "repelling" LCS (approximating future stable manifolds); forward integration yields "attracting" LCS (approximating future unstable manifolds). Their intersection points often coincide with the "neck" geometry at $L_1/L_2$ neighbourhoods (see libration points).

Practical Notes

Choice of $\Delta t$ : too short, the LCS does not form; too long, trajectories fall into a chaotic sea and FTLE values homogenise. Canales & Howell (2024) use roughly one host-planet orbital period near Ganymede.
Resolution and cost: a typical FTLE map needs $10^5$ – $10^6$ trajectory integrations; recent work accelerates this with GPUs or high-order methods such as Jet Transport (Pérez-Palau et al. 2015).
Energy-threshold identification: inside LCS-bounded regions, trajectories can be classified as captured, impacting, or transiting — providing a decision map for Earth-Moon transfer and capture design.
Symmetry acceleration: the time-reflection symmetry of the CR3BP makes forward and backward LCS mirror images, halving the computation (Canales & Howell 2024).

Applications

Cislunar transport analysis: LCS identify low-energy transit corridors near $L_1/L_2$ , informing invariant-manifold patching strategies; Short & Howell (2014) used them for ARTEMIS stationkeeping assessment.
Jovian and Saturnian moon tours: Canales & Howell (2024) build an FTLE atlas around Ganymede and Europa to characterise the gateway geometry for endgame design.
Asteroid neighbourhoods: FTLE maps around rubble-pile bodies reveal stable rings, impact zones, and escape corridors, coupled with attitude dynamics.
Debris cloud evolution: long-term debris evolution analyses use FTLE maps to delineate capture belts and re-impact belts.

References

Haller, G. (2001). Distinguished material surfaces and coherent structures in three-dimensional fluid flows. Physica D, 149(4), 248–277.
Shadden, S. C., Lekien, F., & Marsden, J. E. (2005). Definition and properties of Lagrangian coherent structures from finite-time Lyapunov exponents in two-dimensional aperiodic flows. Physica D, 212(3–4), 271–304.
Gawlik, E. S., Marsden, J. E., Du Toit, P. C., & Campagnolo, S. (2009). Lagrangian coherent structures in the restricted three-body problem. Celestial Mechanics and Dynamical Astronomy, 103(3), 227–249.
Short, C., & Howell, K. C. (2014). Lagrangian coherent structures in various maps for Earth–Moon systems. Acta Astronautica, 94(1), 592–607.
Pérez-Palau, D., Barrabés, E., & Gomez, G. (2015). Dynamical indicators in the restricted three-body problem. Celestial Mechanics and Dynamical Astronomy, 122(4), 319–341.
Canales, D., & Howell, K. C. (2024). Understanding flow around planetary moons via finite-time Lyapunov exponent maps. Celestial Mechanics and Dynamical Astronomy, 136(2), 11.