1: \documentstyle[prl,aps]{revtex}
2: \twocolumn
3: \begin{document}
4: \input{epsf}
5: \draft
6: \twocolumn[\hsize\textwidth\columnwidth\hsize\csname@twocolumnfalse\endcsname
7: \title{Parametric autoresonance}
8: \author{Evgeniy Khain and Baruch Meerson }
9: \address{Racah Institute of Physics, Hebrew
10: University of Jerusalem, Jerusalem 91904, Israel} \maketitle
11:
12: \begin{abstract}
13: We investigate parametric
14: autoresonance: a persisting phase locking which
15: occurs
16: when the driving frequency of a parametrically excited nonlinear
17: oscillator
18: slowly varies with
19: time. In this regime,
20: the resonant excitation is continuous and unarrested
21: by the oscillator nonlinearity. The
22: system
23: has three
24: characteristic time scales, the fastest one corresponding to
25: the natural frequency of
26: the oscillator. We perform averaging over the
27: fastest time scale and analyze the reduced set of equations
28: analytically and numerically. Analytical results are obtained
29: by exploiting the scale separation between the two remaining time
30: scales which enables one to use the
31: adiabatic invariant of the perturbed nonlinear motion.
32: \end{abstract}
33: \pacs{PACS numbers: 05.45.-a} \vskip1pc] \narrowtext \pagebreak
34:
35: \section{INTRODUCTION}
36:
37: This work addresses a combined action of two
38: mechanisms of resonant excitation of (classical) nonlinear oscillating
39: systems. The first is {\it parametric resonance}. The
40: second is {\it autoresonance}.
41:
42: There are numerous oscillatory systems which interaction with the
43: external world amounts only to a periodic time dependence of their
44: parameters. The corresponding resonance is called {\it parametric}
45: \cite{Landau,Bogolubov}. A textbook
46: example is a simple pendulum with a vertically
47: oscillating point of suspension \cite{Landau}. The main resonance
48: occurs when the excitation frequency
49: $\omega$ is nearly twice the natural frequency of the oscillator
50: $\omega_{0}$ \cite{Landau,Bogolubov}. Applications of this basic
51: phenomenon in physics and technology are ubiquitous.
52:
53: {\it Autoresonance} occurs in nonlinear oscillators
54: driven by a small {\it external} force, almost periodic in time.
55: If the force is
56: {\it exactly}
57: periodic, and in resonance with the natural frequency of the oscillator,
58: the resonance region of the phase plane
59: has a finite (and relatively small) width \cite{Sagdeev,Lichtenberg}.
60: If instead the driving
61: frequency is slowly varying in time (in the right direction
62: determined by the nonlinearity sign), the oscillator can stay
63: phase-locked despite the nonlinearity. This leads to a continuous
64: resonant excitation. Autoresonance has found
65: many applications. It was extensively
66: studied
67: in the context of relativistic particle acceleration: in
68: the 40-ies
69: by McMillan \cite{McMillan}, Veksler \cite{Veksler} and
70: Bohm and Foldy \cite{Bohm1,Bohm2}, and more recently
71: \cite{Golovanivsky2,Meerson2,Meerson10,Friedland1}.
72: Additional applications include a quasiclassical scheme
73: of excitation of atoms \cite{Meerson1}
74: and molecules \cite{molecules}, excitation of nonlinear
75: waves \cite{Deutsch,Friedland2},
76: solitons \cite{Aranson,Friedland3},
77: vortices \cite{Friedland4,Friedland5} and other collective
78: modes \cite{Fajans1}
79: in fluids and
80: plasmas,
81: an autoresonant mechanism of transition to chaos in Hamiltonian systems
82: \cite{Yariv1,Cohen}, etc.
83:
84: Until now autoresonance was considered only in systems
85: executing {\it externally} driven oscillations. In this work we
86: investigate autoresonance in a {\it parametrically} driven
87: oscillator.
88:
89: Our presentation will be as follows. In Section 2 we briefly review
90: the parametric resonance in non-linear oscillating systems.
91: Section 3 deals, analytically and numerically,
92: with parametric autoresonance. The conclusions are presented
93: in Section 4. Some details of derivation are
94: given in Appendices A and B.
95:
96:
97: \section{PARAMETRIC RESONANCE WITH A CONSTANT DRIVING FREQUENCY}
98:
99: The parametric resonance
100: in a weakly nonlinear oscillator with finite dissipation and
101: detuning is describable by the
102: following equation of motion \cite{Bogolubov,Struble2,Morozov}:
103: \begin{equation}
104: \ddot{x}+2\gamma\dot{x}+
105: \left[1+\epsilon \cos \left((2+\delta)t\right)\right]x- \beta
106: x^3=0. \label{Bbegin1}
107: \end{equation}
108: where the units of time are chosen in such a way that the scaled
109: natural frequency of the oscillator in the small-amplitude limit
110: is equal to 1. In Eq. (\ref{Bbegin1}) $ \epsilon$ is the amplitude
111: of the driving force, which is assumed to be small: $
112: 0<\epsilon\ll1$, $ \delta \ll 1$ is the detuning parameter, $
113: \gamma$ is the (scaled) damping coefficient $ ( 0<\gamma\ll1 ) $
114: and $\beta$ is the nonlinearity coefficient. For concreteness we
115: assume $\beta>0$ (for a pendulum $\beta =1/6$).
116:
117: Working in
118: the limit of weak nonlinearity, dissipation and driving, we can
119: employ the method of
120: averaging \cite{Bogolubov,Sagdeev,Rabinovich,Drazin},
121: valid for most of the initial conditions \cite{Sagdeev,Lichtenberg}.
122: The unperturbed oscillation period is the fast time.
123: Putting $ x=a(t)\cos\theta(t)$ and $
124: \dot{x}=-a(t)\sin\theta(t)$ and performing averaging over the fast
125: time, we
126: arrive at the averaged equations
127: \begin{eqnarray}
128: \dot{a}&=&-\gamma a+\frac{\epsilon a}{4}\sin 2\psi,\nonumber
129: \\
130: \dot{\psi}&=& -\frac{\delta}{2}-\frac{3\beta a^2}{8}
131: +\frac{\epsilon}{4}\cos 2\psi, \label{B6}
132: \end{eqnarray}
133: where a new phase $\psi=\theta-[(2+\delta)/2] t$ has been
134: introduced. The averaged
135: system (\ref{B6}) is an autonomous dynamical
136: system with two degree of freedom
137: and therefore integrable.
138: In the conservative case $\gamma=0$ Eqs. (\ref{B6})
139: become:
140: \begin{eqnarray}
141: \dot{a}&=&\frac{\epsilon a}{4}\sin 2\psi,\nonumber
142: \\
143: \dot{\psi}&=&-\frac{\delta}{2}-\frac{3\beta
144: a^2}{8}+\frac{\epsilon}{4}\cos 2\psi.\label{B61}
145: \end{eqnarray}
146: As $\sin 2\psi$ and $\cos 2\psi$ are periodic functions of $\psi $
147: with a period $\pi$, it is sufficient to consider the interval
148: $-\pi/2<\psi\leq\pi/2 $. For small enough detuning, $
149: \delta<\epsilon/2 $, there is an elliptic fixed point with a
150: non-zero
151: amplitude:$$a_{*}=\pm\left[\frac{2\epsilon}{3\beta}\left(1-\frac{2\delta}{\epsilon}\right)\right]^{1/2};
152: \indent\psi_{*}=0.$$ We need to calculate the period of motion in
153: the phase plane along a closed orbit around this fixed point (such
154: an orbit is shown in Fig.~\ref{close}).
155:
156: \begin{figure}[h]
157: \vspace{-0.3 cm} \center{\epsfxsize=7.5 cm %\epsfysize=8 cm
158: \epsffile{close1_obr.eps}} \caption{Parametric resonance with a
159: constant driving frequency. Shown is a typical closed orbit in the
160: phase plane with period $T$ ($\epsilon=0.04, \delta=-0.04,
161: \beta=1/6$ and $\gamma=0.$). For a time-dependent driving
162: frequency $\nu(t)$, the autoresonance will occur if the
163: characteristic time for variation of $\nu(t)$ is much greater than
164: $T$, see criterion~(\ref{Ccrit}). \label{close}}
165: \end{figure}
166:
167: This calculation was performed
168: by Struble \cite{Struble2}.
169: For a zero detuning,
170: $\delta=0$,
171: Hamilton's function (we will call it the Hamiltonian) of
172: the system~(\ref{B61})
173: is the following:
174:
175: \begin{equation}
176: H(I,\psi)=\frac{\epsilon I}{4}\cos 2\psi-\frac{3\beta
177: I^2}{8}=H_{0} = const., \label{help1}
178: \end{equation}
179: where we have introduced the action variable
180: $I=a^2/2$. Solving Eq.~(\ref{help1})
181: for $I$ and substituting the result into the
182: Hamilton's equation for $\dot{\psi}$ we obtain:
183: \begin{equation}
184: \dot{\psi}=\mp\frac{\epsilon}{4}\left(\cos^2 2\psi-\frac{24\beta
185: H_{0}}{\epsilon^2}\right)^{1/2}, \label{help2}
186: \end{equation}
187: where the minus (plus) sign corresponds to the upper (lower) part of the closed
188: orbit. The
189: period of the amplitude and phase oscillations is therefore
190: \begin{equation}
191: T=\frac{8}{\epsilon}\int_{-\overline{\psi}}^{\overline{\psi}}\frac{d\psi}{\left(\cos^2
192: 2\psi-\frac{24\beta H_{0}}{\epsilon^2}\right)^{1/2}},
193: \label{help3}
194: \end{equation}
195: where $-\overline{\psi}$ and $\overline{\psi}$ are the roots of
196: the equation $ \cos^2 2\psi=24\beta H_{0}/\epsilon^2.$ Calculating the
197: integral, we
198: obtain
199: \begin{equation}
200: T=\frac{8}{\epsilon}K(m), \label{help4}
201: \end{equation}
202: where $K(m)$ is the complete
203: elliptic integral of the first kind \cite{Abramowitz},
204: and $m=1-24\beta H_{0}/\epsilon^2$. This
205: result will be used in Section 3 to establish a necessary
206: condition for the parametric autoresonance to
207: occur.
208:
209: \section{PARAMETRIC RESONANCE WITH A TIME-DEPENDENT DRIVING
210: FREQUENCY: PARAMETRIC AUTORESONANCE}
211:
212: Now let the driving frequency
213: vary with time. This time dependence introduces
214: an additional (third) time scale into the problem.
215: The governing equation becomes
216: \begin{equation}
217: \ddot{x}+ 2\gamma\dot{x} + (1+\epsilon\cos\phi)x-\beta x^3=0,
218: \label{B6.5}
219: \end{equation}
220: where $\dot{\phi}=\nu(t)$. We will assume $\nu(t)$ to be a
221: {\it slowly} decreasing function which initial value
222: is $\nu(t=0)=2+\delta.$ Using the scale separation, we obtain
223: the averaged equations. The averaging
224: procedure of Section 2 can be repeated by
225: replacing $(2+\delta)t$ by $\phi$ in all equations.
226: There is one new point that
227: should be treated more accurately.
228: The averaging procedure is applicable (again, for most
229: of the initial conditions) if there is a separation
230: of time
231: scales. It requires, in particular,
232: a strong inequality $2\dot{\theta}+\nu (t)\gg 2\dot{\theta}-\nu(t)$.
233: This inequality can limit the time of validity of the method of
234: averaging. Let us assume, for concreteness, a linear frequency ``chirp'':
235: \begin{equation}
236: \nu(t)=2+\delta-2\mu t, \label{B7}
237: \end{equation}
238: where $\mu\ll 1$ is the
239: chirp rate. In this case the averaging procedure is
240: valid as long as $\mu t\ll 1$.
241:
242: Introducing a new phase $\psi=\theta-\phi/2 $,
243: we obtain a reduced set of equations (compare to
244: Eqs.~(\ref{B6})):
245: \begin{eqnarray}
246: \dot{a}&=&-\gamma a+\frac{\epsilon a}{4}\sin 2\psi,\nonumber
247: \\
248: \dot{\psi}&=&-\frac{\delta}{2}+\mu t-\frac{3\beta
249: a^2}{8}+\frac{\epsilon}{4}\cos 2\psi. \label{B8}
250: \end{eqnarray}
251: The first of Eqs. (\ref{B8})
252: is typical for {\it parametric} resonance: to get excitation one should
253: start from a non-zero oscillation amplitude. As we will see,
254: the $\mu t$ term in the second of Eqs. (\ref{B8}) (when small enough
255: and
256: of the right sign) provides a continuous phase locking, similar to
257: the externally driven
258: autoresonance.
259:
260: Consider a numerical example. Fig.~\ref{figexample} shows
261: the time dependence $a (t)$ found by solving
262: Eqs. ~(\ref{B8}) numerically. One
263: can see that the system remains
264: phase locked which allows the amplitude of oscillations
265: to increase, on the average, with time in spite of the nonlinearity.
266: The time-dependence of the amplitude includes
267: a slow trend and relatively
268: fast, decaying oscillations. These are the two time
269: scales
270: remaining after the averaging over the fastest time scale.
271:
272: \begin{figure}[h]
273: \vspace{-0.4 cm} \center{\epsfxsize=7.5 cm %\epsfysize=8 cm
274: \epsffile{example_obr.eps}}
275:
276: \caption{An example of parametric autoresonance. Shown is the
277: oscillation amplitude versus time, computed numerically from the averaged
278: equations~(\ref{B8}). The system remains phase-locked
279: which allows the amplitude to increase, on the average,
280: with time. The parameters are
281: $\mu=6.5\cdot 10^{-5}, \epsilon=0.04, \delta=-0.01,
282: \beta=1/6$ and $\gamma=0.001$. \label{figexample}}
283: \end{figure}
284:
285: Similar to the externally-driven autoresonance, a persistent
286: growth of the oscillation amplitude requires the characteristic
287: time of variation of $\nu(t)$ to be much greater than the
288: ``nonlinear'' period $T$ [see Eq.~(\ref{help4})] of oscillations
289: of the amplitude:
290: \begin{equation}
291: \left|\frac{\nu(t)}{\dot{\nu}(t)}\right|\gg T. \label{Ccrit}
292: \end{equation}
293:
294: Like its externally-driven analog, the
295: parametric autoresonance is insensitive to the exact form of
296: $\nu(t)$. For a given set of
297: parameters, the optimal chirping rate can be found: too low a
298: chirping rate means an inefficient excitation, while too high a rate
299: leads to phase unlocking and termination of the excitation.
300:
301: In the remainder of the paper we will develop an analytical theory
302: of the parametric autoresonance. The first objective of this
303: theory is a description of
304: the slow trend in the amplitude (and phase) dynamics.
305: When the driving frequency $\nu$ is constant, there is an elliptic
306: fixed point $a_{*}$ (see Section 2). When $\nu$ varies with time,
307: the fixed point ceases to exist. However, for a {\it slowly}-varying
308: $\nu(t)$ one can define a ``quasi-fixed'' point $a_{*}(t)$ which
309: is a slowly varying function of time. It is this quasi-fixed point
310: that represents the slow trend seen in Fig. \ref{figexample} and
311: corresponds to an ``ideal'' phase-locking regime. The
312: fast, decaying oscillations seen in Fig. \ref{figexample}
313: correspond to oscillations around the quasi-fixed point in the
314: phase plane [this phase plane is actually projection of the
315: extended phase space ($a,\psi, t$) on the ($a,\psi$)-plane].
316:
317: In the main part of this Section we neglect the dissipation
318: and use a Hamiltonian
319: formalism. First we will consider excitation in
320: the vicinity of the quasi-fixed point. Then
321: excitation from arbitrary initial conditions will be investigated.
322: Finally, the role of dissipation will be briefly analyzed.
323:
324: For a time-dependent $\nu(t)$, the
325: Hamiltonian becomes [compare to Eq. (\ref{help1})]:
326: \begin{equation}
327: H(I,\psi,t)=\frac{\epsilon I}{4}\left(\alpha(t)+\cos
328: 2\psi\right)-\frac{3\beta I^2}{8}, \label{C1}
329: \end{equation}
330: where $\alpha(t)=(4/\epsilon)(1-\nu(t)/2).$ The
331: Hamilton's equations are:
332: \begin{eqnarray}
333: \dot{I}&=&\frac{\epsilon I}{2}\sin 2\psi,\nonumber
334: \\
335: \dot{\psi}&=&\frac{\epsilon}{4}\left(\alpha+\cos
336: 2\psi\right)-\frac{3\beta I}{4}. \label{C2}
337: \end{eqnarray}
338: Let us find the quasi-fixed point of~(\ref{C2}), i.e.
339: the special autoresonance trajectory $I_{*}(t),\,\psi_{*}(t)$
340: corresponding to
341: the ``ideal'' phase locking (a pure trend without oscillations).
342:
343: Assuming a slow time dependence, we put $\dot{\psi_{*}} = 0$,
344: that is
345: \begin{equation}
346: \frac{\epsilon}{4}\left(\alpha+\cos 2\psi_{*}\right)-\frac{3\beta
347: I_{*}}{4}=0. \label{C2-1}
348: \end{equation}
349: Differentiating it with respect to time and using Eqs.
350: (\ref{C2}), we obtain an algebraic equation for $\psi_* (t)$:
351: \begin{equation}
352: 2\alpha (t) \sin 2\psi_{*}+\sin 4\psi_{*} = \frac{16\mu}{\epsilon^2} .
353: \label{C2-2}
354: \end{equation}
355: At this point we should demand that $\dot{\psi_*} (t)$,
356: evaluated
357: on the solution of Eq. (\ref{C2-2}),
358: is indeed negligible compared to the rest of terms in the
359: equation (\ref{C2}) for $\dot{\psi} (t)$. It is easy to see that
360: this requires $16 \mu/\epsilon^2 \ll 1$.
361: In this case the sines in Eq. (\ref{C2-2}) can be
362: replaced by their arguments, and we obtain
363: the following simple expressions for the
364: quasi-fixed point:
365: \begin{eqnarray}
366: I_{*}&\simeq&\frac{\epsilon}{3\beta}\left(\alpha+1\right)\,,\nonumber
367: \\
368: \psi_{*}&\simeq&\frac{k}{\alpha+1}\,, \label{C3}
369: \end{eqnarray}
370: where $k=4\mu/\epsilon^2$.
371:
372: \subsection{Excitation in the vicinity of the quasi-fixed point}
373:
374: Let us make the
375: canonical transformation from
376: variables $I$ and $\psi$ to $ \delta I=I-I_{*}$ and
377: $\delta\psi=\psi-\psi_{*}.$ Assuming $\delta I$ and $\delta\psi$
378: to be small and keeping terms up to
379: the second order in $ \delta I$ and $\delta\psi$,
380: we obtain the new Hamiltonian:
381: \begin{eqnarray}
382: H(\delta I,\delta\psi,\alpha(t))&=&
383: %\frac{\epsilon^2}{24\beta}(\alpha+1)^2
384: -\frac{\epsilon k}{\alpha+1}\delta I \delta\psi- \nonumber
385: \\
386: &-&\frac{3\beta}{8}(\delta I)^2-
387: \frac{\epsilon^2}{6\beta}(\alpha+1)(\delta\psi)^2\,. \label{C4}
388: \end{eqnarray}
389: Here and in the following small terms of order of $k^2$ are neglected.
390: Let us start with the calculation of the local maxima of $\delta I
391: (t)$
392: and $\delta\psi (t)$, which will be called
393: $\delta I_{max}(t)$ and $\delta\psi_{max}(t)$, respectively.
394: As $\alpha(t)$ is a
395: slow function of time [so that
396: the strong inequality (\ref{Ccrit}) is satisfied], we can
397: exploit the approximate
398: constancy of the adiabatic invariant \cite{Landau,Goldstein}:
399: \begin{equation}
400: J=\frac{1}{2\pi}\oint\delta I d(\delta\psi) \simeq const. \label{C5}
401: \end{equation}
402: $|J|$ is the
403: area of the ellipse defined by Eq.~(\ref{C4}) with the
404: time-dependencies ``frozen''. Therefore,
405: \begin{equation}
406: J=\frac{2}{\epsilon}\frac{H}{(\alpha+1)^{1/2}}
407: %\left(\frac{\epsilon^2}{24\beta}(\alpha+1)^2-H\right)
408: \simeq const. \label{C6}
409: \end{equation}
410: This expression can be rewritten in terms of $\delta
411: I$ and $\delta\psi$:
412: \begin{eqnarray}
413: |J| &=&\frac{2k}{(\alpha+1)^{3/2}}\delta I
414: \delta\psi+\frac{3\beta}{4\epsilon}\frac{1}{(\alpha+1)^{1/2}}(\delta
415: I)^2\nonumber
416: \\
417: &+&\frac{\epsilon}{3\beta}(\alpha+1)^{1/2}(\delta\psi)^2.
418: \label{C7}
419: \end{eqnarray}
420: If $k=4 \mu/\epsilon^2 \ll 1$, the term with $\delta I
421: \delta\psi$ in (\ref{C7}) can be neglected (in this approximation
422: one has $\psi_* = 0$).
423: Then $J$ becomes a sum of two non-negative terms, one of
424: them having the maximum value when the other one vanishes. Therefore,
425: \begin{equation}
426: \delta I_{max}(t)=2\left(\frac{\epsilon
427: J}{3\beta}\right)^{1/2}\left(\alpha+1\right)^{1/4}, \label{C8}
428: \end{equation}
429: and
430: \begin{equation}
431: \delta\psi_{max}(t)=\left(\frac{3\beta
432: J}{\epsilon}\right)^{1/2}\frac{1}{(\alpha+1)^{1/4}}. \label{C9}
433: \end{equation}
434: Now we calculate the period of oscillations of the action and
435: phase. Using the well-known relation \cite{Landau}
436: $T=2\pi(\partial J/\partial H)$, we obtain from Eq. (\ref{C6}):
437: \begin{equation}
438: T=\frac{4\pi}{\epsilon}\frac{1}{(\alpha+1)^{1/2}}. \label{C10}
439: \end{equation}
440: The period of oscillations versus time is shown in
441: Fig.~\ref{fiq-C3}. The theoretical curve [Eq.~(\ref{C10})] shows
442: an excellent agreement with the numerical solution.
443:
444: Now we obtain the complete solution $\delta I(t)$ and
445: $\delta\psi(t)$. The Hamilton's equations corresponding to the
446: Hamiltonian (\ref{C4}) are:
447:
448: \begin{eqnarray}
449: \dot{\delta I}&=&
450: \frac{\epsilon^2}{3\beta}\left(\alpha+1\right)\delta\psi+\frac{\epsilon
451: k}{\alpha+1}\delta I,\nonumber
452: \\
453: \dot{\delta\psi}&=&-\frac{3\beta}{4}\delta I-\frac{\epsilon
454: k}{\alpha+1}\delta\psi. \label{C11}
455: \end{eqnarray}
456:
457: Differentiating the second equation with respect to time and
458: substituting the first one, we obtain a linear differential
459: equation for $\delta\psi(t)$:
460: \begin{equation}
461: \ddot{\delta\psi}+\omega^2(t)\delta\psi=0, \label{C12}
462: \end{equation}
463: where $\omega(t)=(\epsilon/2)(\alpha(t)+1)^{1/2}$. For the
464: linear $\nu(t)$ dependence (Eq.~(\ref{B7})) we
465: have $\alpha(t)=4\mu t/\epsilon-2\delta/\epsilon$, therefore for
466: $k\ll 1$ the criterion $\dot{\omega}/\omega^2\ll 1$ is satisfied,
467: and Eq. (\ref{C12}) can be solved by the WKB method (see, {\it
468: e.g.} \cite{Lichtenberg}).
469:
470: \begin{figure}[h]
471: \vspace{0 cm} \rightline{\epsfxsize=7.5 cm %\epsfysize=8 cm
472: \epsffile{period_fixed_obr.eps}}
473:
474: \caption{Excitation in the vicinity of the quasi-fixed point:
475: the time-dependence of the period $T$ of the action and phase
476: oscillations. The solid line is the theoretical curve, Eq.~(\ref{C10}),
477: the asterisks are points obtained numerically. The parameters are
478: $\mu=6.5\cdot 10^{-5}, \epsilon=0.04, \delta=-0.01$ and
479: $\beta=1/6$. \label{fiq-C3}}
480: \end{figure}
481:
482: The WKB solution takes the form (details are given in Appendix A):
483: \begin{eqnarray}
484: \delta\psi(t)&=&\left(\frac{3\beta
485: J}{\epsilon}\right)^{1/2}\frac{1}{(\alpha+1)^{1/4}}\nonumber
486: \\
487: & \times & \cos\left(q_{0}+\frac{(\alpha+1)^{3/2}}{3k}\right),
488: \label{C15}
489: \end{eqnarray}
490: where the phase $q_{0}$ is determined by the initial conditions.
491: The full solution for the phase is $\psi=\delta\psi+\psi_{*}$ and Fig.
492: ~\ref{fig-c2} compares it with a numerical solution of Eq.
493: (\ref{C2}). Also shown are the minimum and maximum
494: phase deviations predicted by Eqs. (\ref{C9}) and (\ref{C3}). One
495: can see that the agreement is excellent.
496:
497: The solution for $\delta I (t)$ can be obtained by substituting
498: Eq.~(\ref{C15}) into the second equation of the
499: system~(\ref{C11}). In the same order of accuracy (see Appendix A)
500: \begin{equation}
501: \delta I(t)=2\left(\frac{\epsilon
502: J}{3\beta}\right)^{1/2}\left(\alpha+1\right)^{1/4}\sin\left(q_{0}+\frac{(\alpha+1)^{3/2}}{3k}\right)\,.
503: \label{C19}
504: \end{equation}
505:
506: Fig.~\ref{fig-c1} shows the dependence of the action variable with
507: the trend $I_*(t)$ subtracted, $\delta I (t)$, on time predicted
508: by Eq.~(\ref{C19}), and found from the numerical solution. It also
509: shows the minimum and maximum action deviations (\ref{C8}). Again,
510: a very good agreement is obtained.
511:
512: \begin{figure}[h]
513: \vspace{0 cm} \leftline{\epsfxsize=7.5 cm %\epsfysize=8 cm
514: \epsffile{phase2_obr.eps}} \caption{Parametric autoresonance
515: excitation in the vicinity of the quasi-fixed point. Shown is the
516: phase $\psi (t)$ found analytically [Eqs. (\ref{C3}) and (\ref{C15})]
517: and by solving Eq. (\ref{C2}) numerically. The
518: analytical and numerical curves are indistinguishable. Also shown
519: are the minimum and maximum phase deviations predicted by Eq.
520: (\ref{C9}) and (\ref{C3}). The parameters are the same as in Fig.
521: \ref{fiq-C3}. \label{fig-c2}}
522: \end{figure}
523:
524: \begin{figure}[h]
525: \vspace{0 cm} \leftline{\epsfxsize=7.5 cm %\epsfysize=8 cm
526: \epsffile{action2_obr.eps}} \caption{Parametric autoresonance
527: excitation in the vicinity of
528: the quasi-fixed point. Shown is the action variable
529: $\delta I (t)$ from Eq.~(\ref{C19}) and from the numerical solution.
530: Also shown are the minimum and maximum action
531: deviations predicted by Eq. (\ref{C8}). The parameters are the same as
532: in Fig. \ref{fiq-C3}.
533: \label{fig-c1}}
534: \end{figure}
535:
536:
537: \subsection{Excitation from arbitrary initial conditions}
538:
539: In this Subsection we go beyond
540: the close vicinity of the quasi-fixed
541: point
542: and
543: calculate the maximum
544: deviations of the action $I$ and phase $\psi$
545: for arbitrary initial conditions. Again, these calculations are made
546: possible by employing the adiabatic invariant for the general case.
547: Correspondingly,
548: the period of the action and phase oscillations will
549: be also calculated.
550:
551: Let us first express the maximum and minimum {\it action} deviations
552: in terms of the Hamiltonian $H$ and driving frequency
553: $\nu(t)$. Solving Eq.~(\ref{C1}) as a quadratic equation for $I$,
554: we obtain: $$ I_{1,2}=\frac{\epsilon}{3\beta}\left(\alpha+\cos
555: 2\psi\right)\pm\left[ \frac{\epsilon^2}{9\beta^2}\left(\alpha+\cos
556: 2\psi\right)^2-\frac{8H}{3\beta}\right]^{1/2}.$$ The time derivative
557: of
558: $I$ vanishes when $I=I_{max}$
559: or $I=I_{min}$. Therefore, from
560: the first equation of the system~(\ref{C2}) $\psi=0$ so that
561: \begin{equation}
562: I_{max,min}=\frac{\epsilon}{3\beta}\left(\alpha+1\right)\pm\left[
563: \frac{\epsilon^2}{9\beta^2}\left(\alpha+1\right)^2-\frac{8H_{up,down}}{3\beta}\right]^{1/2},
564: \label{C20}
565: \end{equation}
566: where $H_{up, down}=H(I_{max, min},\psi=0)$.
567:
568: Now we express the maximum and minimum {\it phase} deviations
569: through the Hamiltonian $H$ and driving frequency $\nu(t)$. The
570: time derivative $\dot{\psi}$ vanishes if $\psi=\psi_{max}$ or
571: $\psi=\psi_{min}$, then the second equation of the system
572: ~(\ref{C2}) yields $I=(\epsilon/3\beta)(\alpha+\cos 2\psi)$. In
573: this case the Hamiltonian~(\ref{C1}) becomes
574: $H_{right,left}=(\epsilon^2/24\beta)(\alpha+\cos
575: 2\psi_{max,min})^2$. Finally, the expression for $\psi_{max,min}$
576: is
577: \begin{equation}
578: \psi_{max,min}=\pm\frac{1}{2}\arccos\left[\left(\frac{24\beta
579: H_{right,left}}{\epsilon^2}\right)^{1/2}-\alpha\right]\,. \label{C21}
580: \end{equation}
581:
582: Fig.~\ref{fiq-C6} shows a part of a typical autoresonant orbit in
583: the phase plane. For $\nu(t)=const.$ this orbit is determined by
584: the equation $H(I,\psi,\nu)=const.$, and it is closed. As in our
585: case $\nu(t)$ changes with time, the trajectory is not closed. To
586: calculate the maximum and minimum deviations of action and phase
587: we should know the values of the Hamiltonian at 4 points of the
588: orbit that we will call ``up", ``down", ``left", and ``right" in
589: the following.
590:
591: \begin{figure}[h]
592: \vspace{-0.1 cm} \rightline{\epsfxsize=7.5 cm %\epsfysize=8 cm
593: \epsffile{phase_plane_obr.eps}} \caption{A part of the
594: autoresonant orbit in the phase plane. Knowing the Hamiltonian at
595: the 4 points, we can calculate the maximum and minimum deviations
596: of the action and phase. The parameters are the same as in Fig.
597: \ref{fiq-C3}. \label{fiq-C6}}
598: \end{figure}
599:
600: Knowing the values of the Hamiltonian at these 4 points, we
601: calculate $I_{max,min}$ from Eq.~(\ref{C20}) and $\psi_{max,min}$
602: from Eq.~(\ref{C21}). Figs.~(\ref{fiq-C7}) and~(\ref{fiq-C8}) show
603: these deviations for action and phase correspondingly, and the
604: values of $I$ and $\psi$, found from numerical solution. The
605: theoretical and numerical results show an excellent
606: agreement.
607:
608: \begin{figure}[h]
609: \vspace{-0.1 cm} \leftline{\epsfxsize=7.5 cm %\epsfysize=8 cm
610: \epsffile{action_arb_ogib_obr.eps}} \caption{The maximum and
611: minimum deviations of the action, calculated from Eq.~(\ref{C20})
612: (thick line) and from numerical solution (thin line). The
613: parameters are the same as in Fig. \ref{fiq-C3}. \label{fiq-C7}}
614: \end{figure}
615:
616: \begin{figure}[h]
617: \vspace{-0.5 cm} \leftline{\epsfxsize=7.5 cm %\epsfysize=8 cm
618: \epsffile{phase_arb_ogib_obr.eps}} \caption{The maximum and
619: minimum deviations of the phase, calculated from Eq.~(\ref{C21})
620: (thick line) and from numerical solution (thin line). The
621: parameters are the same as in Fig. \ref{fiq-C3}. \label{fiq-C8}}
622: \end{figure}
623:
624: Now we are prepared to calculate the adiabatic invariant
625: $J(H,\nu(t))$. Its (approximate) constancy in time allows one, in
626: principle, to find the Hamiltonian $H(t)$ at any time $t$, in
627: particular at the points of the maximum and minimum action and
628: phase deviations (see Fig.~\ref{fiq-C6}).
629:
630: It is convenient to rewrite the adiabatic invariant in the following
631: form:
632: \begin{equation}
633: J=\frac{1}{2\pi}\oint\psi dI. \label{C22}
634: \end{equation}
635:
636: Using Eq.~(\ref{C1}), we can find $\psi=\psi(H,I,\alpha(t))$:
637: \begin{equation}
638: \psi=\pm \frac{1}{2}\arccos\left(\frac{8H+3\beta I}{2\epsilon I}
639: -\alpha\right)\,, \label{C23}
640: \end{equation}
641: so that Eq. (\ref{C22}) becomes:
642: \begin{equation}
643: J=\frac{1}{2\pi}\int^{I_{max}}_{I_{min}}\arccos\left(\frac{8H+3\beta
644: I}{2\epsilon I} -\alpha\right) dI, \label{C24}
645: \end{equation}
646: where $I_{max}$ and $I_{min}$ are given by Eq.~(\ref{C20}). Notice
647: that $H(t)$ and $\alpha(t)$ should be treated as constants under
648: the integral (\ref{C24}), see Refs.
649: \cite{Landau,Sagdeev,Goldstein}. This integral can be expressed in
650: terms of elliptic integrals (see Appendix B for details). For
651: definiteness, we used the values of $H(t)$ and $\alpha(t)$ in the
652: ``up'' points, see Fig.~\ref{fiq-C6}. We checked numerically that
653: the adiabatic invariant $J(H(t),\alpha(t))$ is constant in our example
654: within 0.12 per cent.
655:
656: Now we calculate the period of action and phase oscillations. From
657: the first equation of system ~(\ref{C2}) we have:
658: \begin{equation}
659: T=2\int^{I_{max}}_{I_{min}}\frac{dI}{(\epsilon I/2)\sin 2\psi},
660: \label{C26}
661: \end{equation}
662: where $I_{max}$ and $I_{min}$ are given by Eq.~(\ref{C20}), while
663: $\psi=\psi(I)$ is defined by~(\ref{C23}).
664:
665: Using Eq.~(\ref{C1}), we obtain after some algebra:
666: \begin{equation}
667: T=\frac{8}{3\beta}\int^{I_{max}}_{I_{min}}\frac{dI}{G(I)^{1/2}},
668: \label{C27}
669: \end{equation}
670: where $G(I)$ is given in Appendix B, Eq. (\ref{C-G}). Again, we
671: treat $H(t)$ and $\alpha(t)$ as constants under the
672: integral~(\ref{C27}), and take their values in the ``right''
673: points, see Fig.~\ref{fiq-C6}. The final result is:
674:
675: \begin{equation}
676: T=C_{2}K(C_{3}), \label{period}
677: \end{equation}
678: where $C_{2}=4(2/3\beta H\epsilon^2)^{1/4}$ and
679: $$C_{3}=\frac{1}{2}-\frac{C_{2}^2}{16}\left[\frac{3\beta
680: H}{2}+\frac{\epsilon^2}{16}\left(1-\alpha^2\right)\right]\,.$$
681:
682: \begin{figure}[h]
683: \vspace{-0.0 cm} \rightline{\epsfxsize=7.5 cm %\epsfysize=8 cm
684: \epsffile{period_arb_obr.eps}} \caption{The period $T$ of the phase (action)
685: oscillations obtained from Eq.~(\ref{period}) (solid line), and
686: from numerical solution (asterisks).
687: The parameters are the same as
688: in Fig. \ref{fiq-C3}.
689: \label{fiq-C10}}
690: \end{figure}
691:
692:
693: Figure \ref{fiq-C10} shows the period $T$ of the phase and action
694: oscillations versus time obtained analytically and from numerical solution.
695: This completes our consideration of the parametric autoresonance without
696: dissipation.
697:
698: \subsection{Role of dissipation}
699:
700: Now we very briefly consider the role of dissipation in the
701: parametric autoresonance. Consider the
702: averaged equations~(\ref{B8}) and assume that the detuning is zero.
703: The non-trivial
704: quasi-fixed point exists when the dissipation is not too strong:
705: $\gamma < \epsilon/4$, and it is given by
706: \begin{eqnarray}
707: a_{*}&=&\left(\frac{2\epsilon}{3\beta}\right)^{1/2}\left[\alpha (t) +
708: \left(1-\frac{16\gamma^2}{\epsilon^2}\right)^{1/2}\right]^{1/2},\nonumber
709: \\
710: \psi_{*}&=&\frac{1}{2}\arcsin\left(\frac{4\gamma}{\epsilon}+
711: \frac{2k}{\alpha (t) +(1-16\gamma^2/\epsilon^2)^{1/2}}\right)\,.
712: \label{C29}
713: \end{eqnarray}
714: Again, we assume $k\ll 1$. This quasi-fixed point describes the
715: slow trend in the dissipative
716: case.
717: As we see numerically, fast oscillations around
718: the trend, $\delta a=a-a_{*}$ and $\delta\psi=\psi-\psi_{*}$
719: decay with time. Therefore,
720: one can expect that the $a(t)$ will approach, at
721: sufficiently large times, the trend $a_{*}(t)$.
722: Fig.~\ref{fiq-C13} shows the time dependence of the
723: amplitude, found by solving numerically the system of averaged
724: equations~(\ref{B8}), and the amplitude trend
725: from~(\ref{C29}). We can see that indeed
726: the amplitude $a(t)$ approaches the trend $a_{*}(t)$ at large times.
727:
728: \begin{figure}[h]
729: \vspace{0 cm} \leftline{\epsfxsize=7.5 cm %\epsfysize=8 cm
730: \epsffile{ampl_diss_obr.eps}}
731:
732: \caption{Parametric autoresonance with
733: dissipation: the time dependence of the amplitude
734: of oscillations, obtained from numerical solution of
735: Eqs. (\ref{B8}), and the amplitude trend $a_*(t)$, predicted by
736: Eq. (\ref{C29}). The parameters are $\mu=6.5\cdot 10^{-5},
737: \epsilon=0.04, \delta=0, \gamma=0.002$ and $\beta=1/6$.
738: \label{fiq-C13}}
739: \end{figure}
740:
741: Therefore, a small amount of dissipation enhances the stability of
742: the parametric autoresonance excitation scheme. A similar result
743: for the externally-driven autoresonance was previously known \cite{Yariv2}.
744:
745: \section{CONCLUSIONS}
746:
747: We have investigated, analytically and numerically, a combined
748: action of two mechanisms of resonant excitation of
749: nonlinear oscillating systems: parametric resonance and
750: autoresonance. We have shown that parametric autoresonance represents a
751: robust and efficient method of excitation of nonlinear oscillating
752: systems. Parametric autoresonance can be extended for the
753: excitation of nonlinear {\it waves}. We expect that
754: parametric autoresonance will find applications in different
755: fields of physics.
756:
757: \section*{ACKNOWLEDGEMENTS}
758: This research was supported by the Israel Science Foundation,
759: founded
760: by the Israel Academy of Sciences and Humanities.
761: \appendix
762:
763: \section{CALCULATION OF PHASE AND ACTION DEVIATIONS BY THE WKB-METHOD}
764:
765: Changing the variables from time $t$ to $\alpha$, we can rewrite
766: Eq.~(\ref{C12}) in the following form:
767: \begin{equation}
768: \delta\psi^{\prime\prime}+
769: \left(\frac{\alpha(t)+1}{4k^2}\right)\delta\psi=0\,, \label{AP31}
770: \end{equation}
771: where $^{\prime\prime}$ denotes the second derivative with respect to $\alpha$.
772: Solving this equation by the
773: WKB-method \cite{Lichtenberg}, we obtain
774: for $\delta\psi$:
775: \begin{equation}
776: \delta\psi(t)=\frac{(2kC)^{1/2}}{(\alpha+1)^{1/4}}\cos\left(\Omega_{0}+\frac{(\alpha(t)+1)^{3/2}-1}{3k}\right),
777: \label{AP36}
778: \end{equation}
779: where $\Omega_{0}$ and $C$ are constants to be found later.
780: Now we obtain the solution for $\delta I$.
781: Substituting~(\ref{AP36}) into the second equation of the
782: system~(\ref{C11}), we obtain in the same order of accuracy:
783: \begin{eqnarray}
784: \delta
785: I(t)&=&\frac{2\epsilon}{3\beta}\left(2kC\right)^{1/2}\left(\alpha+1\right)^{1/4}\nonumber
786: \\
787: & \times &
788: \sin\left(\Omega_{0}+\frac{(\alpha(t)+1)^{3/2}-1}{3k}\right).
789: \label{AP37}
790: \end{eqnarray}
791: The constant $C$ can be expressed through the adiabatic invariant
792: $J$, given by~(\ref{C7}). From Eqs.~(\ref{AP36}) and~(\ref{AP37})
793: we have: $$
794: 2kC=\left(\frac{3\beta}{2\epsilon}\right)^2\frac{1}{(\alpha+1)^{1/2}}\left(\delta
795: I\right)^2+\left(\alpha+1\right)^{1/2}(\delta\psi)^2.$$ Comparing
796: it with~(\ref{C7}) we find: $ C\simeq 3\beta J/2k\epsilon.$
797: Substituting this value into Eqs.~(\ref{AP36}) and~(\ref{AP37}) we
798: obtain the final expressions~(\ref{C15}) and (\ref{C19}) for
799: $\delta\psi(t)$ and $\delta I(t)$.
800:
801: \section{CALCULATION OF THE ADIABATIC INVARIANT}
802:
803: After integration by parts and some algebra, using
804: Eqs.~(\ref{C1}) and ~(\ref{C20}), we obtain the following
805: expression for the
806: adiabatic invariant:
807: \begin{equation}
808: J=\frac{1}{2\pi}\int^{I_{max}}_{I_{min}}\left(\frac{I^2-
809: \frac{8H}{3\beta}}{G(I)^{1/2}}\right)
810: dI, \label{C25}
811: \end{equation}
812: where
813: \begin{equation}
814: G(I)=\left(I_{max}-I\right)\left(I-I_{min}\right)\left[\left(I+\frac{\epsilon(1-\alpha)}{3\beta}\right)^2-\frac{16D}{9\beta^2}\right],
815: \label{C-G}
816: \end{equation}
817: and we assume $D=(\epsilon^2/16)(1-\alpha)^2-3\beta H/2<0.$
818: Calculation of this integral employs several changes of
819: variable shown in the best way by Fikhtengolts
820: \cite{Fikhtengolts}. Using the reduction formulas
821: \cite{Abramowitz}, we arrive at:
822: \begin{eqnarray}
823: J=C_{1}\left[\frac{1+mm'}{(1-m)^2(1+m')}\Pi\left(\frac{m}{m-1}\backslash
824: k^2\right)\right.\nonumber
825: \\
826: \left.-\frac{1}{1-m}K\left(k^2\right)+\frac{m+m'}{(1-m)(1+m')}E\left(k^2\right)\right],
827: \label{invariant}
828: \end{eqnarray}
829: where
830: $$m=\frac{(\epsilon/3\beta)(1+\alpha)-(8H/3\beta)^{1/2}}{(\epsilon/3\beta)(1+\alpha)+(8H/3\beta)^{1/2}}>0,$$
831: $$m'=\frac{(\epsilon/3\beta)(1-\alpha)+(8H/3\beta)^{1/2}}{-(\epsilon/3\beta)(1-\alpha)+(8H/3\beta)^{1/2}}>0.$$
832: $$k^2=\frac{m}{m+m'},\indent
833: C_{1}=c\cdot\frac{64H}{3\beta(m+m')^{1/2}},$$
834: and
835: \begin{eqnarray}
836: c&=&\frac{1}{2\pi}\left[\frac{\epsilon}{3\beta}\left(1+\alpha\right)+\left(\frac{8H}{3\beta}\right)^{1/2}\right]^{-1/2}\nonumber
837: \\
838: & \times &
839: \left[-\frac{\epsilon}{3\beta}\left(1-\alpha\right)+\left(\frac{8H}{3\beta}\right)^{1/2}\right]^{-1/2}\,.\nonumber
840: \end{eqnarray}
841: Here $K$, $E$ and $\Pi$ are the complete elliptic integrals of the
842: first, second and third kind, respectively.
843:
844: \begin{references}
845:
846: \bibitem{Landau} L.D. Landau and E.M. Lifshits, {\it Mechanics}
847: (Pergamon Press, Oxford, 1976).
848:
849: \bibitem{Bogolubov} N.N. Bogolubov and Y.A. Mitropolsky, {\it Asymptotic Methods In The Theory of Non-linear Oscillations}
850: (Gordon and Breach Science Publishers, New York, 1961).
851:
852: \bibitem{Sagdeev} R.Z. Sagdeev, D.A. Usikov, and G.M. Zaslavsky, {\it Nonlinear
853: Physics} (Harwood Academic, Switzerland, 1988).
854:
855: \bibitem{Lichtenberg} A.J. Lichtenberg and M.A. Lieberman, {\it Regular and Chaotic Dynamics}
856: (Springer-Verlag, Oxford, 1992).
857:
858: \bibitem{McMillan} E.M. McMillan, Phys. Rev. {\bf 68}, 143 (1945).
859:
860: \bibitem{Veksler} V. Veksler, J.Phys.(USSR) {\bf 9}, 153 (1945).
861:
862: \bibitem{Bohm1} D. Bohm and L. Foldy, Phys. Rev. {\bf 70}, 249 (1947).
863:
864: \bibitem{Bohm2} D. Bohm and L. Foldy, Phys. Rev. {\bf 72}, 649 (1947).
865:
866: \bibitem{Golovanivsky2} K.S. Golovanivsky, Phys. Scripta {\bf 22}, 126 (1980).
867:
868: \bibitem{Meerson2} B. Meerson, Phys. Lett. A {\bf 150}, 290 (1990).
869:
870: \bibitem{Meerson10} B. Meerson and T. Tajima, Optics Communications
871: {\bf 86}, 283 (1991).
872:
873: \bibitem{Friedland1} L. Friedland, Phys. Plasmas {\bf 1}, 421 (1994).
874:
875: \bibitem{Meerson1} B. Meerson and L. Friedland, Phys. Rev. A {\bf 41},
876: 5233 (1990).
877:
878: \bibitem{molecules} J.M. Yuan and W.K. Liu, Phys. Rev. A {\bf 57},
879: 1992 (1998).
880:
881: \bibitem{Deutsch} M. Deutsch, B. Meerson, and J.E. Golub, Phys. Fluids
882: B {\bf 3}, 1773 (1991).
883:
884: \bibitem{Friedland2} L. Friedland, Phys. Plasmas {\bf 5}, 645 (1998).
885:
886: \bibitem{Aranson} I. Aranson, B. Meerson, and T. Tajima, Phys. Rev. A
887: {\bf 45}, 7500 (1992).
888:
889: \bibitem{Friedland3} L. Friedland and A. Shagalov, Phys. Rev. Lett.
890: {\bf 81}, 4357 (1998).
891:
892: \bibitem{Friedland4} L. Friedland, Phys. Rev. E {\bf 59}, 4106 (1999).
893:
894: \bibitem{Friedland5} L. Friedland and A.G. Shagalov, Phys. Rev.
895: Lett. {\bf 85}, 2941 (2000).
896:
897: \bibitem{Fajans1} J. Fajans, E. Gilson, and L. Friedland, Phys.
898: Rev. Lett. {\bf 82}, 4444 (1999); Phys. Plasmas {\bf 6}, 4497 (1999).
899:
900: \bibitem{Yariv1} B. Meerson and S. Yariv, Phys. Rev. A {\bf 44}, 3570 (1991).
901:
902: \bibitem{Cohen} G. Cohen and B. Meerson, Phys. Rev. E {\bf 47}, 967 (1993).
903:
904: \bibitem{Struble2} R.A. Struble, Quart. Appl. Math. {\bf 21}, 121 (1963).
905:
906: \bibitem{Morozov} A.D. Morozov, J. Appl. Math. Mech. {\bf 59}, 563 (1995).
907:
908: \bibitem{Rabinovich} M.I. Rabinovich and D.I. Trubetskov,
909: {\it Oscillations and Waves in Linear and Nonlinear Systems}
910: (Kluwer Academic Publisher, Dordrecht, 1989).
911:
912: \bibitem{Drazin} P.G. Drazin, {\it Nonlinear Systems} (Cambridge
913: University Press, Cambridge, 1992).
914:
915: \bibitem{Abramowitz} M. Abramowitz, {\it Handbook of Mathematical Functions}
916: (National Bureau of Standards, Washington, 1964).
917:
918: \bibitem{Goldstein} H. Goldstein, {\it Classical Mechanics}
919: (Addison-Wesley, Reading, Mass., 1980).
920:
921: \bibitem{Yariv2} S. Yariv and L. Friedland, Phys. Rev. E {\bf 48},
922: 3072 (1993).
923:
924: \bibitem{Fikhtengolts} G.M. Fikhtengolts, {\it The Fundamentals of Mathematical Analysis}
925: (Pergamon Press, New York, 1965).
926:
927:
928: \end{references}
929:
930: \end{document}
931: