The elimination of the critical terms of a first order Uranus-Neptune theory through Hori's method

An outline for the elimination of the critical terms of a first order Uranus-Neptune theory is presented with a stress on the application of Hori's procedure to the problem.     

An elimination of the critical terms of a first-order Uranus-Neptune theory by Hori's method

We eliminate the 1:2 critical terms — after a previous elimination of the short period terms — in the Hamiltonian of a first order U-N theory. We take into account terms of degree 0, 1, 2, 3, 4 in the eccentricity-inclination. We apply for this elimination the Hori-Lie technique through the Poincaré canonical variables and the Jacobi coordinates. The purely principal first order secular U-N Hamiltonian admits a complete solution. We obtained the U-N equations of motion generated by the principal first order long period U-N Hamiltonian which will be solved later. This part III is closely related to the two previous papers (Kamel, 1982, 1983).     

Third-order Uranus-Neptune theory by Hori's method

We construct a U-N secular canonical planetary theory of the third order with respect to planetary masses. The Hori-Lie procedure is adopted to solve the problem. Expansions have been carried out by hand, neglecting powers higher than the second with respect to the eccentricity-inclination. We take into account the principal as well as the indirect part of the planetary disturbing function. The theory is expressed in terms of the Poincaré canonical variables, referring to the Jacobi-Radau set of origins. We assume that the 1:2 U-N critical terms and its multiples are the only periodic terms.     

A second-order Uranus-Neptune theory by Hori's method

We shall establish a second order - with respect to a small parameter which is of the order of planetary masses - Uranus-Neptune canonical planetary theory. The construction will be through the Hori-Lie perturbation theory. We perform the elliptic expansions by hand, taking into account powers 0, 1, 2 of the eccentricity-inclination. Only the principal part of the planetary Hamiltonian will be taken into consideration. Our theory will be expressed in terms of the canonical variables of Henri Poincaré, referring the planetary coordinates to the Jacobi-Radau system of origin. Only U- N critical terms will be assumed as the periodic terms.     

The elimination of short-periodic terms in a Uranus-Neptune general planetary theory by von Zeipel's method

We eliminate by the method of von Zeipel the short-period terms in a first order-with respect to planetary masses—general planetary Uranus-Neptune theory. We exclude in the expansion terms of eccentricities and sines of inclinations higher than the third power.Our variables are the Poincaré canonical variables. We use the Jacobi-Radau set of origins, and we refer the planes of the osculating ellipses to a common fixed plane, the longitudes to a common origin. The short-periodic terms arising from the indirect and principal parts of the disturbing functions, are eliminated separately. The Fourier series of the principal part of the disturbing function, is reduced to the sum of only the first three terms.     

On the elimination of short-period terms in second-order general planetary theory investigated by Hori's method

1. The short-period terms of a second-order general planetary theory are removed through the Hori's method based on a development of the HamiltonianF in a Lie series which involves a determining functionS not depending upon mixed canonical variables as in the Von Zeipel's method but upon all the canonical variables resulting from the elimination of the short period terms ofF. Canonical variables adopted are the slow Delaunay variables. Eccentricitiese _j and sines γ_j of the semi inclinations are respectively replaced by the Jacques Henrard variablesE _j ,J _j which lead to formulas remarkably simple.F is reduced to the sumF ₀+F ₁ of its terms of degrees 0,1 in small parameter ε of the order of the masses. Only one disturbing planet is considered.F ₁ is not calculated beyond its terms of degree 3 inE _j ,E _j ,J _j , the determining functionS ₂ of degree 2 in ε not being therefore calculated beyond its terms of degree 2 inE′ _j ,E _j ,J _j and the expressions of slow Delaunay canonical variables of the disturbed planetP ₁ and the disturbing planetP ₂ in terms of the new slow Delaunay canonical variables ofP ₁ andP ₂ which result from the elimination of the short period terms ofF ₁ being therefore reduced to their terms of degree <1 in theE′ _j ,E′ _j ,J′ _j . Calculation of the principal partF _1m ofF ₁ is carried out through Laplace coefficients and operatorD=α(d/dα) applied to Laplace coefficients, α ratio of the semi major axis ofP ₁ andP ₂. Eccentricitye ₂ of the disturbed planetP ₂ is assumed to be zero, such an assumption not restricting our aim which is to investigate the mechanism of the elimination of short period terms in a second order general planetary theory carried out through the Hori's method, not to perform the elimination of those terms for a complete second order general planetary theory. Expressions of the slow Delaunay canonical variables in terms of the new ones resulting from the elimination of the short period terms ofF ₁ are written down only for the disturbed planetP ₁.
2. Small divisors in 1/E′ ₁ and 1/E′ ₁ ² appear in the longitude ?₁ of perihelia ofP ₁. No small divisors appear in the other five slow Delaunay variables ofP ₁. The only Jacques Henrard variables which appear in the longitude Ω₁ of the ascending node ofP ₁ are the J _j′ j=1, 2 and no Jacques Henrard variables appear in the slow Delaunay canonical variablesX ₁,Y ₁,Z ₁, λ₁. The solving of the ten canonical equations ofP ₁ andP ₂ in the slow Delaunay canonical variablesX′ _j ,Y′ ₁,Z′ _j ,λ′ _j ,ω′ _j ,Ω′ _j resulting from the elimination of the short period terms ofF ₁ reduces to that of four canonical equations inZ′ _j ,©′ _j and to six quadratures three of them expressing theX′ _j ,Y′ ₁ are constants and the three others expressingλ′ _j ,?′ _j as functions of timet. Solving of the four canonical equations inZ′ _j ,Ω′ _j reduces to that of a first order non linear differential equation and to two quadratures. Sinceγ′ ₁ is then constant, so is the Jacques Henrard variableE′ ₁. If the eccentricitye ₂ ofP ₂ is no more assumed to be zero, additive small divisors inE′ ₂/E′ ² ₁ appear in longitude ?′₁ of perihelia ofP ₁ and the solving of the twelve canonical equations ofP ₁ andP ₂ inX′ _j ,Y′ _j ,Z′ _j ,λ′ _j ,?′ _j ,Ω′ _j is reduced to that of eight canonical equations inY′ _j ,?′ _j ,Z′ _j ,Ω′ _j and to four quadratures expressingX′ _j are constants andλ′ _j as functions oft. Those eight canonical equations split into two systems of four canonical equations, one of them inY′ _j ,?′ _j and the other one inZ′ _j ,Ω′ _j . Each of those two systems is identical to the system inZ′ _j ,Ω′ _j corresponding toe ₂=0 and its solving reduces to that of a first order non linear differential equation and to two quadratures identical to those of the casee ₂=0.
3. Expressions ofX ₁,Y ₁,Z ₁,λ ₁,? ₁,Ω ₁ as functions ofX′ _j ,Y′ ₁,Z′ _j ,λ′ _j ,?′ ₁,Ω′ _j ;j=1, 2 are sums of sines and cosines of the multiples ofλ′ _j ,?′ ₁,Ω′ _j for the terms arising from the indirect partF _1j ofF ₁, Fourier series in those sines and cosines or products of two such Fourier series for the terms arising from the principal partF _1m ofF ₁, coefficients of those sums and Fourier series having one of the eight forms: $$A,{\text{ }}\frac{B}{{E'}},{\text{ }}\frac{C}{{E'^2 }},{\text{ }}D\frac{{j'^{2_1 } }}{{E'^{2_1 } }},{\text{ }}E\frac{{j'^{2_2 } }}{{E'^{2_1 } }},{\text{ }}F\frac{{j'^{_1 } j'^2 }}{{E'^{2_1 } }},{\text{ }}G\frac{{j'^2 }}{{j'^{_1 } }},{\text{ }}H\frac{{j'^{22} }}{{j'^{2_1 } }}{\text{.}}$$ A,..., H being constants which depend upon ratio α. Numerical calculation of the constantsA,..., H arising from the terms ofF _1j is easily carried out; that of theA,..., H arising from the terms ofF _1m require more manipulations, Fourier series in sines and cosines of the multiples ofλ′ _j ,?′ _j ,Ω _ij and products of two such Fourier series having then to be reduced to sums of a finite number of terms and treated through the methods of harmonic analysis. Divisors inp+qα^3/2;p, q relative integers, or products of such divisors appear inA,..., H.
4. the method extends to the case whenF ₁ is calculated beyond its terms of degree 3 in the Jacques Henrard variables.F ₁ being calculated up to its terms of degree 8 in the Jacques Henrard variables which is the precision required to eliminate the short period terms of a complete second order general planetary theory,S ₂ has to be calculated up to its terms of degree 7 and the expression of the slow Delaunay canonical variables ofP ₁ andP ₂ in terms of the slow Delaunay canonical variables ofP ₁ andP ₂ resulting from the elimination of the short period terms ofF ₁ have, therefore, to be calculated up to their terms of degree 5 in the Jacques Henrard variables.

     

The elimination of the 1:2 critical terms of a first order theory of Uranus perturbed by Neptune through Von Zeipel method

In this paper we eliminate in a first order U-N theory the 1 : 2 critical terms up to the third degree with respect to eccentricity — inclination in both parts, main and indirect of the U-N planetary Hamiltonian. We operate the Von Zeipel technique. We adopt, in this theory, the Jacobi-Radau coordinates, and the Poincaré canonical variables. We neglect powers higher than the third in the eccentricity-inclination. This paper is related to the two previous articles (Kamel, 1982; 1983).     

On the elimination of the critical term in a general first-order Uranus-Neptune theory

A solution of the Uranus-Neptune planetary canonical equations of motion through the Von Zeipel technique is presented. A unique determinging function which depends upon mixed canonical variables, reduces the 12 critical terms of the Hamiltonian to the set of its secular terms. The Poincaré canonical variables are used. We refer to a common fixed plane, and apply the Jacobi-Radau set of origins. In our expansion we neglected terms of power higher than the fourth with respect to the eccentricities and sines of the inclinations.     

Elimination of the critical terms in a first-order general planetary theory

We construct a first order canonical general planetary theory, assuming the solar system to be composed of 8 planets excluding Pluto, referring to common fixed plane and applying the Jacobi-Radau set of origins. We eliminated by von Zeipel's method the 2:5 and 1:2 critical terms of Jupiter-Saturn and Uranus-Neptune inequalities. Our variables are those of Poincaré, and we expanded up to power three in the eccentricities and sines of the inclinations.     

Construction of a general planetary theory of the first order

All the necessary formulae for constructing a general solution for the motion of a planet, in rectangular coordinates, at the first order of the disturbing masses, in purely literal form in eccentricities and inclinations, are given. The authors present the transformation formulae in the two-body problem which give the correspondence between the constants of integration introduced in the theory and the classical keplerian elements. The practical elaboration of the algorithm and some partial results for the couple of planets Jupiter and Saturn are described.     

Application of Hori's technique in general planetary theory

In this part we present the complete solution of the planetary canonical equations of motion by the method of G. Hori through successive changes of canonical variables using the Lie series. Thus, we can eliminate the long or critical terms of the planetary perturbing function, in our general planetary theory. In our formulas, we neglect perturbation terms of order higher than the third with respect to planetary masses.     

The generalized canonical form of Hori's method for non-canonical systems

Any dynamical system can be put in generalized canonical form through the introduction of a set of auxiliary ‘conjugate’ variables or momenta and solved by perturbation theory based on Lie series. The application of Hori's method for generalized canonical system leads to a new canonical transformation — the Mathieu transformation — defined by the solution of the Hori auxiliary system. This new transformation simplifies the algorithm since the inversion of the solution of the Hori auxiliary system is no longer necessary. In this paper, we wish to show some peculiarities of this technique.     

Expansion of the first order planetary disturbing function by Smart's method

In this part we expand the indirect part of the planetary perturbing function by Smart's method, via Taylor's theorem. We neglect, in our expansion, terms of degree higher than the fourth with regard to the eccentricities and tangents of the inclinations.     

Solution of canonical equations which result from elimination of short-periodic terms in second-order planetary theory

We solve the first order non-linear differential equation and we calculate the two quadratures to which are reduced the canonical differential equations resulting from the elimination of the short period terms in a second order planetary theory carried out through Hori's method and slow Delaunay canonical variables when powers of eccentricities and the sines of semi-inclinations which are >3 are neglected and the eccentricity of the disturbing planet is identically equal to zero. The procedure can be extended to the case when the eccentricity of the disturbing planet is not identically equal to zero. In this latter general case, we calculatedthe two quadratures expressing angular slow Delaunay canonical variable ₁ of the disturbed planet and angular slow Delaunay canonical variable ₂ of the disturbing planet in terms of timet.     

The elimination of the parallax in satellite theory

When the perturbation affecting a Keplerian motion is proportional tor ^–n (n3), a canonical transformation of Lie type will convert the system into one in which the perturbation is proportional tor ^–2. Because it removes parallactic factors, the transformation is called the elimination of the parallax.In the main problem for the theory of artificial satellites, the elimination of the parallax has been conducted by computer to order 4. The first order in the reduced system may now be integrated in closed form, thereby revealing the fundamental property of the first-order intermediary orbits in line with Newton's Propositio XLIV.Extension beyond order 1 leads to identify a new class of intermediaries for the main problem in nodal coordinates, namely the radial intermediaries.The technique of smoothing a perturbation prior to normalizing the perturbed Keplerian system, of which the elimination of the parallax is an instance, is applied to derive the intermediaries in nodal coordinates proposed by Sterne, Garfinkel, Cid-Palacios and Aksnes, and to find the canonical diffeomorphisms which relate them to one another and to the radial intermediaries.     

The construction of a third order secular analytical J-S-U-N theory by Hori-Lie technique

We construct the outline of a third order secular theory for the four major planets. We apply the Hori-Lie technique to solve the problem. We take into consideration both parts of the perturbing function. Our canonical variables are those of H. Poincaré. Our periodic terms are the only 2:5 and 1:2 critical terms of J-S and U-N respectively. Terms of degree higher than the second in the Poincaré canonical variables H, K, P, Q are neglected.     

The construction of a third order secular analytical J-S-U-N theory by Hori-Lie technique

The first and second order generating functions and the first second and third order new J-S-U-N Hamiltonians are calculated by the Hori-Lie procedure.     

The elimination of short and intermediate period terms from the problem of a high altitude earth satellite

The Delaunay-Similar elements of Scheifele are applied to the problem of an Earth satellite that is perturbed by the Sun, Moon andJ ₂. All three effects are assumed to be the same order of magnitude. Since the external body terms depend explicitly on time, the time element appears as an additional angle variable. Also, the eccentric anomaly is used as a noncanonical auxiliary variable. A solution to the first Von Zeipel equation allows simultaneous elimination of short and intermediate period terms. The canonical transformation to mean elements is defined by a generating function that is a series involving Bessel coefficients.     

Quantitative perturbation theory by successive elimination of harmonics

We revisit some results of perturbation theories by a method of successive elimination of harmonics inspired by some ideas of Delaunay. On the one hand, we give a connection between the KAM theorem and the Nekhoroshev theorem. On the other hand, we support in a quantitative fashion a semi-numerical method of analysis of a perturbed system recently introduced by one of the authors.     

GEOS-II and 13th order terms of the geopotential

The resonance of GEOS-II (1968-002A) with 13th-order terms of the geopotential is analyzed. The odd-degree geopotential coefficients (13, 13), (15, 13), and (17, 13) given by Yionoulis most accurately model the resonance effects on GEOS-II of any of the published sets of 13th-order coefficients. However, this set is not adequate for precision orbit determination; additional even-degree coefficients are required.Values ofC _14,13(=0.57×10^–21) andS _14,13(=6.5×10^–21) to be used with the odd-degree set of Yionoulis were obtained from an analysis of the observed along-track position variation of GEOS-II. These coefficients, when used with those of Yionoulis, yield greatly improved fits to the data and orbital prediction capability. However, further refinement is possible because the small effects of the remaining even-degree resonant terms were not modeled.The composite coefficientsC _13,13(=1.7×10^–20) andS _13,13(=+2.7×10^–20) were obtained under the assumption that the (13, 13) spherical harmonic of the geopotential is responsible for all of the observed along-track variation of GEOS-II due to resonance. The good agreement of these deliberately composite values with some published values ofC _13,13 andS _13,13 suggests that some of the published values may also be composite to some extent.These coefficients are hereinafter referred to as the APL coefficients.