Analytical Mechanics This page intentionally left blank
Download 10.87 Mb. Pdf ko'rish
|
g w j P 2 Fig. 4.6 Applying equations (4.153) we then find ω ± = Ω 2 ⎧ ⎪ ⎨ ⎪ ⎩ 1 2 1 +
Ω 1 Ω 2 2 + 1 µ ± 1 2 ⎡ ⎣ 1 + Ω 1 Ω 2 2 − 1 µ 2 + 4 µ ⎤ ⎦ 1/2 ⎫ ⎪ ⎬ ⎪ ⎭ 1/2 . What happens in the limiting cases Ω 1 / Ω 2 1 and Ω 1 / Ω 2 1? Problem 6 Consider the holonomic system with smooth fixed constraints, and with l degrees of freedom, associated with the kinetic energy T =
1 2 ˙q T S · q, where S is a constant symmetric positive definite matrix. Let q = A(s)Q be a group of linear transformations with A(s) an l ×l matrix, such that A(0) = 1. (i) How must A(s) be chosen for T to be invariant? (ii) If there are no active forces, which is the corresponding first integral (Noether’s theorem)? (iii) If there exists a potential U(q), what conditions on its structure must be imposed to ensure that the transformation is admissible for the Lagrangian?
178 The dynamics of discrete systems. Lagrangian formalism 4.14 Solution
(i) A T SA = S. This transformation leaves the Riemannian metric associated with the kinetic energy invariant. Verify that the matrices with this property form a group and study the special case l = 2. (Hint: Choose the coordinates so that S is diagonal and prove that det(A) = 1, A 11 = A 22 . Then A(s) can be looked for in the form A(s) =
cos s c sin s
−c −1 sin s cos s , obtaining c = (S 22 /S 11 ) 1/2
.) (ii) Since p = S ˙q and dA ds
s =0 = ˙ A(0)q, the first integral is given by I = ˙q T S ˙
A(0)q. (iii) It must be that U(A(s)Q) = U(Q). In the particular case that S = k1, k > 0, then A(s) is a group In this case ˙ A(0) = Ω
the first integral takes the form I = i>j
Ω ij (p i q j − p j q i ).
5 MOTION IN A CENTRAL FIELD 5.1
Orbits in a central field Consider a point particle of mass m and denote by r the position vector in the space R 3
F(r) = f (r) r r , r =
|r| = / 0,
(5.1) where f : (0, + ∞) → R is a regular function, is conservative (Example 2.2) with potential energy V (r) = − f(r) dr. The moment of the field (5.1) with respect to the centre is zero, yielding conservation of the angular momentum L. The motion takes place in the plane passing through the origin and orthogonal to L, namely the plane identified by the initial position vector r 0 and the initial velocity vector v 0 (note that in the case L = 0, the vectors r 0 and v
0 are
necessarily parallel and the motion takes place along a line). We now introduce in the orbit plane (which we assume to be the (x, y) plane, as shown in Fig. 5.1) the polar coordinates x = r cos ϕ, y = r sin ϕ. (5.2)
The angular momentum of the system, L, can then be identified with the component L z :
z = m(x ˙
y − y ˙x) = mr 2 ˙
(5.3) and the conservation of L yields that L z is constant along the motion. The conservation of L z also yields Kepler’s second law, about the area swept by the vector r = r(t) in the time interval (0, t): S(t) =
1 2 ϕ (t) ϕ (0) r 2 (ϕ) dϕ = 1 2 t 0 r 2 ˙ ϕ dτ =
L z t 2m . (5.4) T heorem 5.1 (Kepler’s second law) The areal velocity ˙ S(t) =
L z 2m (5.5) is a constant, and its value is also known as the area constant. 180 Motion in a central field 5.1
0
r 0
w 0
L O x Fig. 5.1
Introduce the radial unit vector e r = (cos ϕ, sin ϕ) and the orthogonal unit vector e ϕ = ( − sin ϕ, cos ϕ). The equation of motion m¨ r = f (r) r r = −V (r) r r (5.6) can then be written componentwise as follows: 1 r
dt (r 2 ˙ ϕ) = 2 ˙r ˙ ϕ + r ¨ ϕ = 0,
m¨ r − mr ˙ϕ 2 = − dV dr , (5.7) and the first equation simply expresses the conservation of L z .
efinition 5.1 The function V e (r) = V (r) + L 2 z 2mr
2 (5.8)
is called the effective potential energy. Using V
e (r) in (5.6), and considering equation (5.3), the equation governing the radial motion becomes m¨ r = − dV e dr (r).
(5.9) The total energy E also takes a simple form, given by E = 1
m ˙r 2 + V (r) = 1 2 m ˙r 2 + V
e (r),
(5.10) 5.1 Motion in a central field 181 showing that the problem is equivalent to the one-dimensional motion of a point particle of mass m under the action of a force field with potential energy equal to the effective potential V e . Note that shifting the term mr ˙ ϕ 2 to the right-hand side of equation (5.7) is equivalent to writing the equation of motion in the non-inertial reference system with an axis coinciding with the direction of the radius r. The effective potential energy is the potential energy computed by such an observer. Remark 5.1 It is possible to reach the same conclusion through the use of the Lagrangian formalism. Indeed, the Lagrangian of a point particle of mass m under the action of a central field can be written as L = m
( ˙ x 2 + ˙ y 2 + ˙ z 2 ) − V ( x
2 + y
2 + z
2 ), (5.11) and is clearly invariant under the action of rotations around the origin. It follows from Noether’s theorem (4.4) that the angular momentum L is conserved. If the motion is in the (x, y) plane and ˙ z ≡ 0, and after introducing polar coordinates (5.2) the Lagrangian becomes L =
m 2 ( ˙r 2 + r
2 ˙ ϕ 2 ) − V (r). (5.12) The coordinate ϕ is cyclic, and hence L z = ∂L/∂ ˙
ϕ is constant, and the motion is reduced to one-dimensional motion with energy (5.10). If L z
It is a solution of the equation m¨ r = f (r) which we discussed in Section 3.1. Otherwise the polar angle ϕ is a monotonic function of time (increasing if L z > 0 and decreasing L z
the trajectory can be parametrised as a function of the angle ϕ; we then write dr dt = ˙ ϕ dr dϕ = L z mr 2 dr dϕ · (5.13) It follows from the fact that energy is conserved that the equation for the function r = r(ϕ) describing the orbit is dr dϕ = ± mr 2 L z 2 m (E − V e (r)). (5.14) This equation is called the first form of the orbit equation. The sign in (5.14) is determined by the initial conditions and equation (5.14) can be integrated by separation of variables: ϕ − ϕ
0 = ± r r 0 L z m m 2 dρ ρ 2 E − V e (ρ) , (5.15)
182 Motion in a central field 5.1 where r
0 = r(ϕ
0 ). We find then ϕ = ϕ 0 + ϕ(r), and inverting this expression we obtain r = r(ϕ). Remark 5.2 It is possible to have circular motion; by Theorem 5.1 such motion must be uniform, in correspondence with the values of r which annihilate the right-hand side of (5.9), and hence of the stationary points of V e (r). If r = r c is one such value, equation (5.10) shows that the energy corresponding to the circular motion is E
c = V
e (r c ). We shall return to this in Section 5.3. Example 5.1: the harmonic potential Let V (r) =
1 2 mω 2 r 2 (5.16) (motion in an elastic field). The effective potential corresponding to it is given by V e (r) = L 2 z 2mr
2 + 1 2 mω 2 r 2 . (5.17) It is easily verified (Fig. 5.2) that V e (r)
≥ E c = V e (r c ), where (Remark 5.2) r c = |L z | mω , E c = ω |L z |. (5.18) V e (r) E c r m r c r M r Fig. 5.2
5.1 Motion in a central field 183 For every fixed value of E > E c , the equation V e (r) = E has two roots: r m = E mω 2 1 − 1 −
E 2 c E 2 , r M = E mω 2 1 + 1 − E 2 c E 2 . (5.19) From (5.15) we derive (note that r m /r
= r c /r M ) ϕ − ϕ 0 = r (ϕ)
r 0 dr r 2 2mE L 2 z − 1 r 2 − r 2 r 4 c , (5.20) from which, setting w = 1/r 2 , ϕ − ϕ
0 = 1/r 2 0 1/r(ϕ) 2 dw 2 m 2 E 2 L 4 z − 1 r 4 c − w − mE L 2 z 2 , (5.21)
and by means of the substitution w − mE L 2 z = m 2 E 2 L 4 z − 1 r 4 c cos ψ = mE L 2 z 1 − E 2 c E 2 cos ψ, we find that the integration yields ψ/2. Choosing the polar axis in such a way that r = r m for ϕ = 0, we finally obtain 1 r(ϕ)
2 = mE L 2 z 1 + 1 − E 2 c E 2 cos 2ϕ . (5.22) Equation (5.22) describes an ellipse centred at the origin, whose semi-axes are given by (5.19). Note that the orbit is a circle if E = E c , yielding r = r c . Another form of the orbit equation can be obtained by the substitution of u = 1/r into the equation of motion (5.9). Since d dt = ˙ ϕ d dϕ , (5.23) we obtain, as in (5.13), ¨ r = ˙ ϕ d dϕ ˙ ϕ d dϕ r =
L 2 z u 2 m 2 d dϕ u 2 du dϕ dr du = − L 2 z u 2 m 2 d 2 u dϕ 2 · (5.24) 184 Motion in a central field 5.1 On the other hand − ∂ ∂r V e (r) = u 2 d du V e 1 u , (5.25) and substituting (5.24) and (5.25) into (5.9) we obtain the equation d 2 u dϕ 2 = − m L 2 z d du V e 1 u , (5.26)
called second form of the orbit equation. Using the variable u the energy can be written in the form E = 1
L 2 z du dϕ 2 + V e 1 u . (5.27) Example 5.2 Consider the motion of a point particle of mass m = 1 in the central field V (r) = −k
/2r 2 , where k is a real constant. Setting u = 1/r, the effective potential is given by V e (1/u) = 1 2 (L 2 z − k 2 )/2u
2 ; substituting the latter into (5.26) yields the equation d 2 u dϕ 2 + 1 − k 2 L 2 z u = 0. (5.28) If we set ω 2 = 1
− k 2 /L 2 z , the solution of (5.28) corresponding to the data u (0) = −r (0)/r(0) 2 is given by u(ϕ) = ⎧ ⎪ ⎪ ⎪ ⎨ ⎪ ⎪ ⎪ ⎩ u(0) cos ωϕ + u (0) ω
if k 2
2 z
u(0) + u (0)ϕ, if k
2 = L
2 z , u(0) cosh ωϕ + u (0)
ω sinh ωϕ,
if k 2 > L 2 z . If k 2 > L 2 z and the energy E = 1 2 (L 2 z /2)[(u (0)) 2 −ω 2 (u(0)) 2 ] is negative, the orbit is bounded (i.e. u(ϕ) does not vanish) and it describes a spiral turning towards the centre of the field if u (0) > 0 (the so-called Cotes spiral ; see Danby 1988). We now return to the general case and fix a non-zero value of L z ; the orbit belongs to A E,L z = {(r, ϕ)|V e (r)
≤ E}, (5.29)
consisting of one or more regions bounded by circles. In each region the radius r lies between a minimum value r m (pericentre) and a maximum r M (apocentre, see Fig. 5.3), where r m and r M are two consecutive roots of V e (r) = E (except 5.2 Motion in a central field 185
F F r m Fig. 5.3
in the case r m = 0 or r M = +
∞). If the point is initially positioned in a region in which
0 ≤ r
m ≤ r
M < + ∞ the motion is bounded. If r m < r M (otherwise the motion is circular), from equation (5.3) it follows that the polar angle ϕ varies monotonically, while r oscillates periodically between r m and r
M . In general the orbit is not closed. Indeed, from equation (5.15) it follows that the angle Φ between a pericentre and an apocentre is given by the integral Φ = r M r m L z mr 2 dr 2 m [E − V e (r)] (5.30) (the integral converges provided that r m and r
M are simple roots of the equation V e
Φ . Hence the necessary and sufficient condition that the orbit is closed is that there exist two integers n 1 and n 2 such that Φ = 2π
n 1 n 2 , (5.31) i.e. that the ratio Φ /2π is rational. If, on the other hand, Φ /2π is not rational, one can prove that the orbit is dense in the annulus r m
M .
Kepler’s problem In this section we study the motion under the action of the Newtonian potential V (r) = −
r , k > 0. (5.32) 186 Motion in a central field 5.2 The effective potential corresponding to (5.32) is V e (r) = L 2 z 2mr 2 − k r = L 2 z u 2 2m − ku, (5.33)
where we set u = 1/r. Substituting equation (5.33) into the orbit equation (5.26) we find
d 2 u dϕ 2 = −u + km L 2 z . (5.34) The solution of the latter is the sum of the integral of the associated homogeneous equation, which we write in the form u = (e/p) cos(ϕ − ϕ
0 ), and of a particular solution of the non-homogeneous equation u = 1/p, corresponding to the unique circular orbit admissible for the Newtonian potential, of radius r c
L 2 z km (5.35)
and corresponding to energy E c = − k 2 m 2L 2 z · (5.36) The parametric equation of the orbit is given by u = 1
(1 + e cos(ϕ − ϕ
0 )),
and Kepler’s first law follows: r =
p 1 + e cos(ϕ − ϕ 0
, (5.37)
where e ≥ 0 is the eccentricity of the orbit. Hence the orbit is a conical section, with one focus at the origin: if 0 ≤ e < 1 the orbit is an ellipse, if e = 1 it is a parabola and if e > 1 it is a hyperbola. The eccentricity is determined by e =
1 + 2L 2 z E k 2 m = 1 + E |E c | , E ≥ E
c . (5.38) In the elliptic case (E < 0) the two semi-axes a and b are given by a =
1 2 (r m + r
M ) =
p 1 − e 2 = k 2 |E|
, b = a
Download 10.87 Mb. Do'stlaringiz bilan baham: |
ma'muriyatiga murojaat qiling