1. Introduction
N-body simulations in gravitational fields remains a central issue in celestial mechanics and stellar dynamics. While a limited number of exact integrable cases are known for stationary potentials, they do not adequately describe the rich dynamical behavior observed in realistic stellar systems, such as rotating galaxies, close binaries, or dense star clusters. Particle trajectories in such systems often display a mixed character, with regular motion confined to certain phase-space regions and chaotic behavior elsewhere. A particular interest lies in so-called local integrals of motion, which are conserved not globally throughout phase space, but only on specific invariant submanifolds. The concept was introduced in the astrophysical context by (Reference Antonov1) for non-rotating systems, where the integrals typically appear as quadratic functions of velocities. This concept contrasts with the earlier global approaches of Lynden-Bell (1962, 2016) and provides a means to analyze systems that are not fully integrable in the classical sense. Despite their limited domains of validity, local integrals offer several significant advantages: they reduce the effective dimensionality of the system, define barriers to phase-space transport, and provide insight into the long-term behavior of orbits. They are particularly relevant for describing confined motion in self-gravitating systems and for understanding the structure and evolution of stellar subsystems within larger aggregates. The paper gives an investigation of the existence and structure of local integrals of motion, both linear and quadratic in the velocity components, in stationary potentials observed in rotating frames. Two- and three-dimensional models are analyzed, and the classes of potentials admitting such integrals are derived. We show that these classes are more general than those associated with global integrability. In three-dimensional settings, we further introduce the concept of multi-valued velocity fields–extending traditional box orbits - which allow a more detailed classification of orbit families in non-integrable systems.
2. General Formulas
2.1. Linear Local Integrals in Planar Rotating Systems
We consider the planar motion of a particle in a stationary gravitational potential in a uniformly rotating frame with angular velocity
$\Omega=$
const. A class of potentials of the form
admits a local integral linear in velocities, expressed as
where u and v are the velocity components in the rotating frame, and
$\varphi$
is a scalar function of spatial coordinates. A specific example of such a system is constructed, and the domain of admissible motion is explicitly determined. The particle trajectories exhibit coiled structures, and a comparison is made with more realistic galactic potentials (Reference Antonov and Shamshiev2; Reference Shamshiev9).
2.2. Quadratic Local Integrals in Rotating Systems
We investigate local integrals that are quadratic in the velocity components and are valid on invariant surfaces in phase space, even when the potential lacks global symmetries. Such integrals take the general form
We identify classes of two-dimensional rotating potentials admitting such integrals. Among them are the following specific forms:
These potentials define phase-space surfaces on which the velocity field can be directly inferred for specific initial conditions. The complete classification of such cases and corresponding integrals is discussed in (Reference Shamshiev13).
2.3. Three-Dimensional Potentials with Multi-Valued Velocity Fields
In three dimensions, certain stationary potentials permit local integrals that define multivalued velocity fields. In particular, velocity vectors of the form
$$\left.\begin{array}{c}u=\frac{\partial \varPsi}{\partial x} \pm \frac{x}{r} \sqrt{\varphi(r)+2 h}+\Omega y \\[3pt]v=\frac{\partial \varPsi}{\partial y} \pm \frac{y}{r} \sqrt{\varphi(r)+2 h}-\Omega x \\[3pt]w=\frac{\partial \varPsi}{\partial z} \pm \frac{z}{r} \sqrt{\varphi(r)+2 h}\end{array}\right\}$$
lead to the corresponding class of potentials: In the non-rotating case (
$\Omega=0$
):
as found in (Reference Antonov and Shamshiev5);
In the rotating frame (
$\Omega=$
const):
as derived in (Reference Shamshiev8). These multi-valued structures generalize conventional box orbits and represent new classes of organized motion in three-dimensional systems.
3. Cases of Integrability
3.1. Planar Motion in a Rotating Potential
Analytically Integrable Cases. We consider special cases of two-dimensional stationary gravitational fields rotating with constant angular velocity
$\Omega$
, for which the existence of a local integral permits full analytical integration of particle trajectories (Reference Antonov and Shamshiev3). Although local integrals generally reduce, but do not eliminate, the complexity of motion, certain specific forms of the potential allow explicit expressions for the trajectories. One such class of potentials is given by:
where
$S(r), \chi$
, and
$\sigma$
are functions or parameters describing the potential structure. In these cases, the equations of motion can be fully integrated, and the orbital paths determined in closed analytical form. While some solutions correspond to trivial configurations, the framework also yields nontrivial, previously unidentified integrable cases, extending the class of analytically tractable models in stellar dynamics.
3.2. Extended Integrability in Three-Dimensional Systems
Building on the notion of local integrals, we explore further conditions under which full integrability can be achieved in three-dimensional rotating systems. A key result, introduced by Antonov and Shamshiev (1992), involves the identification of a local integral that specifies the velocity field for selected initial conditions, thereby reducing the phasespace dimensionality (Reference Antonov and Shamshiev5; Reference Shamshiev11)
We examine cases where this reduction leads to explicit relationships among the spatial coordinates and time along a trajectory, i.e.,
via direct integration of the equations of motion. A classification of such potentials is provided, and one representative case is analyzed in detail. These examples highlight the potential of local integrals to reveal hidden symmetries and partially integrable structures even in the absence of global conservation laws (Reference Shamshiev11; Reference Shamshiev13).
4. Conclusions
-
• In this work, we have investigated the existence of local integrals of motionconserved quantities defined on invariant submanifolds of phase space-in stationary gravitational potentials relevant to stellar dynamics. Both rotating and non-rotating systems were considered.
-
• We derived broad families of two- and three-dimensional potentials that admit local integrals linear or quadratic in velocity. These integrals effectively reduce the dimensionality of the equations of motion and act as constraints on the dynamics, limiting phase-space mixing and facilitating the existence of long-lived, isolated subsystems. In certain cases, they also enable complete analytical integration of the particle trajectories.
-
• In three-dimensional systems, we introduced and analyzed multi-valued velocity fields, which generalize the classical concept of box orbits and provide new insights into the orbital structure in rotating frames. Our analysis demonstrates that the class of potentials admitting local integrals is substantially broader than that of fully integrable systems with global integrals.
-
• These findings extend the theoretical framework for partially integrable gravitational systems and offer practical tools for the modeling of dense stellar systems, rotating galaxies, and gravitationally confined subsystems. The methods developed here may also have potential applications in adjacent fields, including plasma dynamics and Hamiltonian systems theory.
Acknowledgments
The author expresses special gratitude to the organizers of IAU Symposium No. 398 Professors Hyung Mok Lee and Rainer Spurzem - for the excellent organization of the event, as well as to its participants for the valuable discussion of the presented work, which significantly contributed to the improvement of its content.
Funding
The travel was partially supported by a grant from the International Astronomical Union (IAU). We gratefully acknowledge the support of Dr. Laura Ferrarese, Assistant General Secretary of the IAU.