1: %----------------- Settings ----------------------------
2:
3: %\documentclass[preprint, aps, pre, eqsecnum, amsmath,amssymb,showpacs]{revtex4}
4: \documentclass[preprint, aps, pre, preprintnumbers,amsmath,amssymb,showpacs]{revtex4}
5: %\documentclass[draft,twocolumn, aps, pre, eqsecnum,amsmath, amssymb,showpacs]{revtex4}
6: %\documentclass[twocolumn, aps, pre, preprintnumbers,amsmath, amssymb,showpacs]{revtex4}
7:
8:
9:
10: \usepackage[final]{graphicx} % to include figures
11: %\usepackage[draft]{graphicx} % includes only space for figures
12: %\usepackage{psfig}
13:
14: \renewcommand{\arraystretch}{1.5} %faktor 1.5 fuer tabellenzeilenabstand
15: \newcommand{\Nabla}{\mbox{\bf\boldmath $\nabla$}}
16: \newcommand{\ve}[1]{\stackrel{\rightharpoonup}{#1}}
17: \renewcommand{\vec}[1]{{\bf #1}}
18: \newcommand{\Vxi}{\mbox{\bf\boldmath $\xi$}}
19: \newcommand{\Vphi}{\mbox{\boldmath $\varphi$}}
20: \newcommand{\Vzeta}{\mbox{\boldmath $\zeta$}}
21: \newcommand{\VPsi}{\mbox{\boldmath $\Psi$}}
22: \newcommand{\Overrightarrow}[1]{\stackrel{\textstyle\rightarrow}{#1}}
23: \newcounter{saveeqn}
24: \newcommand{\alpheqn}{\setcounter{saveeqn}{\value{equation}}
25: \stepcounter{saveeqn}
26: \setcounter{equation}{0}
27: \renewcommand{\theequation}{\mbox{\arabic{section}.\arabic{saveeqn}\alph{equation}}}
28: }
29:
30:
31: \newcommand{\reseteqn}{\setcounter{equation}{\value{saveeqn}}
32: \renewcommand{\theequation}{\mbox{\arabic{section}.\arabic{equation}}}\\[-3mm]
33: }
34:
35: % info und seite in kopfzeile
36: \pagestyle{myheadings}
37: % datum in kopfzeile. falls 'twoside': {leftside}{rightside}
38: \markboth{\today}{\today}
39:
40: \bibliographystyle{prsty}
41:
42:
43: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
44: \begin{document}
45: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
46:
47: \title{
48: Controlling the stability transfer between oppositely traveling
49: waves and standing waves by inversion-symmetry-breaking
50: perturbations}
51:
52: \author{A.~Pinter\footnote[1]{Electronic address: kontakt@alexander-pinter.de},
53: M.~L\"ucke, and Ch.~Hoffmann}
54: \affiliation{Institut f\"ur Theoretische Physik, Universit\"at des Saarlandes,
55: Postfach 151150, \\ D-66041 Saarbr\"ucken, Germany}
56:
57: \date{\today}
58:
59: %-------------- Abstract --------------------------------
60: \begin{abstract}
61: The effect of an externally applied flow on symmetry degenerated
62: waves propagating into opposite directions and standing waves that
63: exchange stability with the traveling waves via mixed states is
64: analyzed. Wave structures that consist of spiral vortices in the
65: counter rotating Taylor-Couette system are investigated by full
66: numerical simulations and explained quantitatively by amplitude
67: equations containing quintic coupling terms. The latter are
68: appropriate to describe the influence of inversion symmetry breaking
69: perturbations on many oscillatory instabilities with $O(2)$
70: symmetry.
71: \end{abstract}
72:
73: %--------------- PACS -----------------------------------
74: \pacs{47.20.Ky, 47.54.-r, 47.32.-y}
75: %****** aktualisiert am 10.11.2006 *****************************
76: % 47.20.Ky: Nonlinearity, bifurcation, and symmetry breaking
77: % 47.20.-k: Flow instabilities
78: % 47.54.-r: Pattern selection; pattern formation
79: % 47.32.-y: Vortex dynamics; rotating fluids
80: % 47.10.-g: General theory in fluid dynamics
81: % 47.60.+i: Flows in ducts, channels, nozzles, and conduits
82: % 47.20.Lz: Secondary instabilities
83: %****************************************************************
84: \maketitle
85: %---!---10----!---20----!---30----!---40----!---50----!---60----!---70----!---80
86:
87: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
88: %{\it Introduction} --
89: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
90:
91: Many nonlinear structure forming systems that are driven out of
92: equilibrium show a transition to traveling waves (TWs) as a result
93: of an oscillatory instability \cite{CH93}. In the presence of
94: inversion symmetry in one or two spatial directions also a standing
95: wave (SW) solution bifurcates in addition to the symmetry
96: degenerate, oppositely propagating TWs at the same threshold.
97: Moreover, depending on the parameters one can have a stability
98: exchange between TWs and SWs as the driving rate varies. The
99: stability transfer is mediated by mixed patterns that establish in
100: the solution space a connection between a pure TW and a pure SW.
101:
102: Here we investigate how externally tunable symmetry breaking
103: perturbations change stability, bifurcation properties, and the
104: spatiotemporal behavior of the afore mentioned structures. To be
105: concrete we investigate vortex waves in the annular gap between
106: counter rotating concentric cylinders of the Taylor-Couette system
107: \cite{T94,CI94}. To that end we performed full numerical simulations
108: of the Navier-Stokes equations with methods described in
109: \cite{HLP04}. We elucidate that and how these results can be
110: explained quantitatively by amplitude equations which contain
111: only quintic order terms.
112:
113: The perturbation is realized in our system by an externally enforced
114: axial through-flow that can easily be controlled experimentally.
115: However, our results concerning the influence of inversion symmetry
116: breaking perturbations on TWs, SWs, and on the mixed states are more
117: general: the quintic order amplitude equations with small symmetry
118: breaking terms apply to all kinds of $O(2)$ symmetric oscillatory
119: instabilities in the presence of inversion symmetry breaking
120: perturbations.
121:
122: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
123: {\it Structures} --
124: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
125: Without symmetry breaking through-flow the following oscillatory
126: vortex structures occur at small driving \cite{DI84,PLH06}: ({\em
127: i}) Forward bifurcating TWs consisting of left handed spiral
128: vortices (L-SPI) or of right handed spiral vortices (R-SPI) that are
129: mirror images of each other. L-SPI (R-SPI) travel in the annulus
130: between the two cylinders into (opposite to) the direction of the
131: rotation frequency vector of the inner one, i.e., in our notation
132: upwards (downwards) \cite{HLP04}. ({\em ii}) Forward bifurcating SWs
133: that consist of an equal-amplitude nonlinear combination of L-SPI
134: and R-SPI that are called ribbons (RIBs) in the Taylor-Couette
135: literature \cite{TESM89,LPA03}. ({\em iii}) So-called cross-spirals
136: (CR-SPI), i.e., combinations of L-SPI and R-SPI with different
137: amplitudes. They provide a stability transferring connection between
138: TW and SW solution branches \cite{CI94,PLH06}. The vortex structures
139: ({\em i})-({\em iii}) are axially and azimuthally periodic with wave
140: numbers $k=2\pi /\lambda$ and $M$, respectively, with $\lambda=1.3$
141: and $M=2$ throughout this paper. They rotate with characteristic
142: constant angular velocities as a whole into the same direction as
143: the inner cylinder \cite{HLP04}. Thereby they are forced to
144: propagate axially with the exception of RIB vortices which rotate
145: only but do not propagate.
146:
147: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
148: {\it Order parameters} --
149: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
150: Figures
151: \ref{FIG-Bifurcation-Diagram}-\ref{FIG-Bifurcation-Diagram-S-D} show
152: bifurcation diagrams of SPI, CR-SPI, and RIB solutions in a system of radius ratio $\eta=1/2$ for fixed
153: $R_1=240$ versus the reduced distance
154: $\mu=\left(R_2-R_2^0\right)/\left|R_2^0\right|$
155: from the spiral onset $R_2^0$ in the absence of through-flow, $Re=0$. Here,
156: $R_1$ and $R_2$ are the Reynolds numbers defined by the rotational velocities
157: of the inner and outer cylinder, respectively, and $Re$ is the Reynolds number
158: of the imposed axial through-flow. In Figs. \ref{FIG-Bifurcation-Diagram} and
159: \ref{FIG-Bifurcation-Diagram-S-D}, the influence of a small through-flow
160: ($Re=0.02$) is compared with the situation without through-flow. Order
161: parameters in Fig. \ref{FIG-Bifurcation-Diagram} are the squared amplitudes
162: $|A|^2,|B|^2$ of the dominant critical modes
163: \begin{equation} \label{EQ:u2pm1}
164: u_{2,1}(t)=|A|e^{-i\omega_A t}, \quad \quad u_{2,-1}(t)=|B|e^{-i\omega_B t}\,
165: \end{equation}
166: in the double Fourier decomposition of the radial velocity $u$ at
167: mid-gap in azimuthal and axial direction. The indices 2 and $\pm 1$ identify azimuthal and axial modes, respectively. The linear stability analysis of the basic state consisting of a
168: superposition of circular Couette flow and of annular Poiseuille flow in axial
169: direction shows that the growth rates of these modes become
170: positive at the respective bifurcation thresholds $\mu_A$ and $\mu_B$ of $M=2$
171: L-SPI and R-SPI, respectively. Note also that for all relaxed vortex
172: structures investigated here the moduli and frequencies of $u_{2,\pm1}$ in
173: Eq.~(\ref{EQ:u2pm1}) are constant.
174:
175: In Fig.~\ref{FIG-Bifurcation-Diagram-S-D} we show in addition the bifurcation
176: diagrams of the combinations
177: \begin{equation} \label{EQ:SD}
178: S=\left(|A|^2+|B|^2\right)/2 ,\quad \quad D=\left(|A|^2-|B|^2\right)/2\,
179: \end{equation}
180: since they are convenient to describe in particular CR-SPI.
181:
182: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
183: {\it Bifurcation scenario for zero through-flow} --
184: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
185: In the symmetry degenerate case without through-flow L-SPI $(A \neq 0=B, D>0)$
186: and R-SPI $(A=0\neq B, D<0)$ have identical bifurcation properties. They are
187: stable close to onset whereas RIB $(A=B, D=0)$ are initially unstable. In the
188: driving range shown in figures \ref{FIG-Bifurcation-Diagram} and
189: \ref{FIG-Bifurcation-Diagram-S-D} the squared amplitudes of these two states
190: grow practically linearly with the reduced distance $\mu$ from the common
191: onset at $\mu=0$, albeit with different slopes. Then, there appear in a finite
192: supercritical driving interval stable CR-SPI solutions which transfer
193: stability from SPI to RIB. The solution which bifurcates with $B=0$ out of the
194: L-SPI is identified as a L-CR-SPI with $|A| > |B|$, i.e., $D>0$. The symmetry
195: degenerate R-CR-SPI $(|B|>|A|, D<0)$ bifurcates with $A=0$ out of the R-SPI.
196: In the former $|B|$ grows and $|A|$ decreases -- and vice versa in the latter
197: -- until the CR-SPI branches end with $A=B, D=0$ in the RIB solution. The
198: amplitude variations of the CR-SPI solutions, however, are such that the sum
199: $S$ remains practically constant, cf.
200: Fig.~\ref{FIG-Bifurcation-Diagram-S-D}(a). The RIB state loses stability
201: outside the plot range of figures \ref{FIG-Bifurcation-Diagram} and
202: \ref{FIG-Bifurcation-Diagram-S-D} to another type of amplitude-modulated
203: CR-SPI that are not discussed here.
204:
205: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
206: {\it Through-flow induced changes} --
207: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
208: The axial through-flow significantly perturbs and changes structure, dynamics,
209: and bifurcation behavior of the SPI vortex solutions discussed so far \cite{HLP04,CK87,BP90,RL93,AMM06}: RIBs
210: cease to exist in the strict sense, L- and R-SPI are no longer mirror images
211: of each other, and also L-CR-SPI are no longer related to R-CR-SPI by this
212: symmetry operation. However, spirals retain their spatiotemporal structure in
213: the through-flow, i.e., they still do not depend on $\varphi,z,t$ separately
214: but only on the phase combination $\phi_A= M\varphi + kz-\omega_A t$ or
215: $\phi_B= M\varphi - kz-\omega_B t$, respectively. L-SPI (red color in Figs.
216: \ref{FIG-Bifurcation-Diagram}-\ref{FIG-Bifurcation-Diagram-S-D}) bifurcate
217: for small $Re>0$ at a threshold value $\mu_A<0$ prior to R-SPI (orange color)
218: which bifurcate at $0<\mu_B \approx -\mu_A$ out of the basic state.
219:
220: L-SPI are again stable at onset, but then lose stability to L-CR-SPI (violet
221: color) which remain stable in the plotted
222: parameter regime. The L-CR-SPI solution approaches the $(Re=0)$ RIB state with
223: increasing $\mu$, but retains with $A\neq B$ a finite distance $D>0$. On the
224: other hand, R-SPI are unstable for small and large $\mu$, but stable for
225: intermediate $\mu$. Stability is exchanged with a R-CR-SPI (magenta color ) which
226: has a stable as well as an unstable branch resulting from a
227: saddle-node bifurcation at $\mu_S$ in Figs.~\ref{FIG-Bifurcation-Diagram},
228: \ref{FIG-Bifurcation-Diagram-S-D}. For small $\mu$ the unstable R-CR-SPI lies
229: close to the $(Re=0)$ RIB solution and bifurcates with finite $D<0$ out of the
230: R-SPI slightly above $\mu_B$. Note also that the sum of the squared
231: amplitudes $S$ of CR-SPI is no longer constant as for the case of $Re=0$.
232:
233: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
234: {\it Amplitude equations} --
235: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
236: The changes in spatiotemporal, bifurcation, and stability
237: behavior of SPI, CR-SPI, and RIB states by a small through-flow can be
238: explained and described close to onset quantitatively within an
239: amplitude-equation approach. To demonstrate that we focus here on
240: the bifurcation properties of the moduli $|A|, |B|$ of the critical modes.
241:
242: In order to reproduce the bifurcation and stability behavior of the
243: aforementioned vortex states including the CR-SPI one needs coupled equations
244: for $A$ and $B$ of at least quintic order. Higher-order terms that are suggested in
245: \cite{CI94} are not necessary to ensure the existence of CR-SPI solutions.
246: Symmetry arguments \cite{PLH06,CI94,CK87} restrict the form of the equations for
247: the moduli to
248: \begin{subequations}\label{EQ-gekoppelte-AG}
249: \begin{eqnarray}
250: \tau_A\frac{d|A|}{dt}&=&|A|\left[\left(\mu-\mu_A\right)+b_A|A|^2+c_A|B|^2+e_A\left(|A|^2|B|^2-|B|^4\right)\right],\\
251: \tau_B\frac{d|B|}{dt}&=&|B|\left[\left(\mu-\mu_B\right)+b_B|B|^2+c_B|A|^2+e_B\left(|B|^2|A|^2-|A|^4\right)\right]
252: \end{eqnarray}
253: \end{subequations}
254: with real coefficients that depend in general on $Re$.
255: Here we have discarded the quintic terms $|A|^5$ and $|B|^5$ in view of the
256: linear variation of the squared SPI moduli with $\mu$, cf.
257: Fig.~\ref{FIG-Bifurcation-Diagram}. Furthermore, we made a special choice for
258: the coefficients of the last terms $|A||B|^4$ and $|B||A|^4$ in Eq.
259: (\ref{EQ-gekoppelte-AG}) that is motivated by the linear variation of the
260: squared RIB moduli with $\mu$ for $Re=0$ and that suffices to
261: describe the behavior for small $Re$ as well.
262:
263: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
264: {\it Coefficients} --
265: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
266: As a result of the inversion symmetry under $z \leftrightarrow -z$ which
267: includes reverting the through-flow the coupled equations are invariant under
268: the operation $(A,B,Re) \leftrightarrow (B,A,-Re)$ so that the coefficients in
269: Eq.~(\ref{EQ-gekoppelte-AG}) obey relations like, e.g., $c_A(Re)=c_B(-Re)$. To
270: reproduce the bifurcation properties of the moduli for small $Re$ as in
271: Figs.~\ref{FIG-Bifurcation-Diagram}, \ref{FIG-Bifurcation-Diagram-S-D} it
272: suffices to incorporate the $Re$ dependence to linear order in the
273: coefficients
274: \begin{subequations}\label{EQ-coefficients}
275: \begin{eqnarray}
276: \mu_A(Re)=-\mu^{(1)} Re, && \quad \mu_B(Re)=\mu^{(1)} Re \\
277: c_A(Re)=c+c^{(1)}Re, && \quad c_B(Re)=c-c^{(1)}Re
278: \end{eqnarray}
279: only and to ignore any $Re$ dependence of the others by setting
280: \begin{eqnarray}
281: b_A=b_B=b,\quad e_A=e_B=e.
282: \end{eqnarray}
283: \end{subequations}
284: The choice $b_A=b_B=b$ reflects the fact that the linear growth of the squared
285: SPI moduli with the distance from their respective thresholds at $\mu_A$ and
286: $\mu_B$ is unchanged by the through-flow. On the other hand,
287: the downwards shift of the L-SPI onset being for small through-flow of equal
288: magnitude as the upwards shift of the R-SPI onset is reflected by
289: $\mu_A(Re)=-\mu_B(Re)=-\mu^{(1)} Re$ with positive $\mu^{(1)}$. The coupling
290: constant $e_A=e_B=e$ ensures the existence of CR-SPI solutions \cite{PLH06}.
291: The flow induced changes of the coupling constants $c_A$ and $c_B$ reflect the
292: perturbation and destruction of the RIB states and ensures for positive $Re$
293: L-CR-SPI solutions with $D>0$ when $\mu$ is large. The values of the
294: coefficients \cite{coefficients} were obtained from linear stability analyses
295: and by fits to the full numerical nonlinear results.
296:
297: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
298: {\it Fixed points} --
299: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
300: With the coefficients (\ref{EQ-coefficients}) it is straightforward to derive
301: from (\ref{EQ-gekoppelte-AG}) the following relations for the fixed points of
302: $S$ and $D$
303: \begin{subequations}\label{EQ-Fixpunkt-Re}
304: \begin{eqnarray}
305: \mu+\left(b+c\right)S- c^{(1)} Re D -2eD^2&=&0, \\
306: \mu^{(1)}Re +(b-c)D+c^{(1)}Re S +2eSD &=&0.
307: \end{eqnarray}
308: \end{subequations}
309: From these equations we have obtained the lines in the bifurcation plots of
310: $|A|^2=S+D$ and $|B|^2=S-D$ in Fig.~\ref{FIG-Bifurcation-Diagram} and of $S$
311: and $D$ in Fig.~\ref{FIG-Bifurcation-Diagram-S-D}. They compare very well with
312: the symbols from the full numerical simulations of the Navier-Stokes
313: equations, say, up to $Re\simeq 0.2$ or so. Beyond that higher order terms in
314: the $Re$-expansion of the coefficients should be included.
315:
316: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
317: {\it Perturbation of the RIB $\to$ CR-SPI bifurcation} --
318: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
319: The black lines and symbols in Figs.~\ref{FIG-Bifurcation-Diagram},
320: \ref{FIG-Bifurcation-Diagram-S-D} show that without through-flow the RIB state
321: which is stable at large $\mu$ transfers stability to CR-SPI in a pitchfork
322: bifurcation located at $\mu_2$ in Fig.~\ref{FIG-Bifurcation-Diagram},
323: \ref{FIG-Bifurcation-Diagram-S-D}. The symmetry and topology of this
324: unperturbed bifurcation out of the $D=0$ RIB solution at
325: $\mu_2=-(b+c)S_{\ast}$, $S_{\ast}=(c-b)/(2e)$ is best seen in
326: Fig.~\ref{FIG-Bifurcation-Diagram-S-D}(b). This figure shows also most clearly
327: how the through-flow with $Re>0$ perturbs this bifurcation into a continuously
328: varying stable solution branch with $D>0$ (violet) and a pair of stable and unstable
329: solutions (magenta) that are connected by a saddle-node bifurcation at
330: $\mu_S$.
331:
332: The values of $S$ and $D$ at the unperturbed location $\mu=\mu_2$ of the
333: pitchfork bifurcation
334: \begin{equation}\label{EQ-S2-D2}
335: S(\mu_2)=S_2=S_{\ast}+\hat s_2Re^{2/3}+ h.o.t.,\qquad D(\mu_2)=D_2=\hat d_2Re^{1/3} + h.o.t.,
336: \end{equation}
337: grow $\propto Re^{2/3}$ and $\propto Re^{1/3}$, respectively, with constants
338: $\hat d^3_2=-(b+c)(\mu^{(1)} + c^{(1)} S_{\ast})/(4e^2)$ and $\hat s_2=2e \hat d_2^2/(b+c)$.
339: The coordinates of the saddle-node bifurcation, as indicated in
340: Fig.~\ref{FIG-Bifurcation-Diagram-S-D}(b) by the magenta arrows,
341: \begin{equation}\label{EQ-muS-DS}
342: \mu_S=\mu_2+\hat \mu_S Re^{2/3}+h.o.t., \qquad D_S=\hat d_S Re^{1/3}+h.o.t.,
343: \end{equation}
344: vary $\propto Re^{2/3}$ and $\propto Re^{1/3}$ with the perturbation strength
345: $Re$. Here the constants are
346: $\hat d^3_S=\left[(c^2-b^2)c^{(1)}+2e(b+c)\mu^{(1)}\right]/\left(16e^3\right)$ and
347: $\hat \mu_S=\left[(b+c)\mu^{(1)}-\mu_2c^{(1)}+4e^2 \hat d_S^3\right]/(2e\hat d_S)$. The
348: comparison in Fig.~\ref{FIG-power-behavior} of these lowest order
349: approximations (\ref{EQ-S2-D2}, \ref{EQ-muS-DS}) that reflect an additive pitchfork unfolding
350: with the full fixed point result from Eq.~(\ref{EQ-Fixpunkt-Re}) shows good agreement for $Re \lesssim
351: 0.02$.
352:
353: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
354: {\it Larger through-flow} --
355: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
356: Here we investigate the transformation of the unperturbed RIB state into
357: a CR-SPI as a function of the through-flow strength extending to larger Reynolds numbers up to $Re=4$ which are about 200 times larger than those studied so far in this paper. For such large $Re$ the amplitude equations with the simple $Re$ dependence of the coefficients (\ref{EQ-coefficients}) are
358: no longer adequate to describe the behavior of the order parameters $S$ and $D$.
359: However, they continue to follow the same simple power laws with $Re$ as the ones for small through-flow: When $\mu$ is well away from the location $\mu_2$ of the unperturbed pitchfork bifurcation as in Fig.~\ref{FIG-throughflow}(a-c) one has $S=S^{(0)}+S^{(2)}Re^2\,, D=D^{(1)}Re+D^{(3)}Re^3$. On the other hand, at $\mu_2$, as in Fig.~\ref{FIG-throughflow}(d-f) one has $S=S^{(0)}+S^{(2/3)}(Re^{1/3})^2+ S^{(4/3)}(Re^{1/3})^4\,, D=D^{(1/3)}Re^{1/3}+D^{(1)}(Re^{1/3})^3$. The lines in
360: Fig.~\ref{FIG-throughflow} show fits of the data from the full numerical simulations (symbols) to these power law expansions for $S$ and $D$. Fig.~\ref{FIG-throughflow}(c,f) shows plots of the resulting squared moduli $|A|^2=S+D$ and $|B|^2=S-D$.
361:
362: Note that inversion symmetry $z \leftrightarrow -z$ implies $S(-Re)=S(Re)$ and $D(-Re)=-D(Re)$. For $Re>0$ $(Re<0)$ the RIB state is transformed into a L-CR-SPI (R-CR-SPI) with $A$ ($B$) being the major mode and $B$ ($A$) the minor one. With increasing through-flow the major (minor) mode of the CR-SPI increases (decreases) until the minor mode vanishes in the transition to a pure SPI.
363:
364: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
365: {\it Conclusion} --
366: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
367: A small applied flow is an inversion symmetry breaking perturbation
368: for fluid wave structures that is easy to control. It significantly
369: changes the bifurcation and stability transfer scenario involving
370: symmetry degenerate TWs, mirror symmetric SWs, and mixed wave
371: patterns. Numerical simulations of such waves in the
372: counter-rotating Taylor-Couette system are explained and reproduced
373: quantitatively by amplitude equations containing quintic order
374: coupling and symmetry breaking perturbation terms. This approach can
375: be applied to a wide range of $O(2)$ symmetric oscillatory
376: instabilities in the presence of inversion symmetry breaking
377: perturbations.
378:
379: This work was supported by the Deutsche Forschungsgemeinschaft.
380: %----!---10----!---20----!---30----!---40----!---50----!---60----!---70----!---80
381:
382:
383: \newpage
384: %----------------- Bibliography ---------------------
385: \begin{thebibliography}{999}
386:
387: %----------------------------------------------------
388: \bibitem{CH93}
389: M.~C.~Cross and P.~C.~Hohenberg,
390: %{\it Pattern formation outside of equilibrium},
391: Rev.~Mod.~Phys. {\bf 65}, 851 (1993).
392: %----------------------------------------------------
393: \bibitem{T94}
394: R.~Tagg,
395: %{\it The Couette-Taylor Problem},
396: Nonlinear Science Today {\bf 4}, 1 (1994).
397: %-----------------------------------------------------
398: \bibitem{CI94}
399: P.~Chossat and G.~Iooss, {\it The Couette-Taylor Problem},
400: (Springer, Berlin, 1994).
401: %-----------------------------------------------------
402: \bibitem{HLP04}
403: Ch.~Hoffmann, M. L\"ucke, and A. Pinter,
404: %{\it Spiral vortices and Taylor vortices in the annulus between rotating cylinders and the effect
405: %of an axial flux},
406: Phys.~Rev.~E {\bf 69}, 056309 (2004).
407: %----------------------------------------------------
408: \bibitem{DI84}
409: Y.~Demay and G.~Iooss,
410: %{\it Calcul des solutions bifurquées pour le problème de Couette-Taylor avec les deux
411: %cylindres en rotation},
412: J. Mec. Theor. Appl., Spec. Suppl., 193 (1984).
413: %----------------------------------------------------
414: \bibitem{PLH06}
415: A.~Pinter, M.~L\"ucke, and Ch.~Hoffmann,
416: %{\it Competition between Traveling Fluid Waves of Left and Right Spiral Vortices and Their Different Amplitude Combinations},
417: Phys. Rev. Lett. {\bf 96}, 044506 (2006).
418: %----------------------------------------------------
419: \bibitem{TESM89}
420: R.~Tagg, W.~S.~Edwards, H.~L.~Swinney, and P.~S.~Marcus,
421: %{\it Nonlinear standing waves in Couette-Taylor flow},
422: Phys.~Rev.~A {\bf 39}, 3734 (1989).
423: %-----------------------------------------------------
424: \bibitem{LPA03}
425: J.~Langenberg, G.~Pfister, and J.~Abshagen,
426: %{\it Standing waves in flow between finite counterrotating cylinders},
427: Phys.~Rev.~E {\bf 68}, 056308 (2003).
428: %----------------------------------------------------
429: \bibitem{CK87}
430: J.~D.~Crawford and E.~Knobloch,
431: %{On degenerate Hopf bifurcation with broken O(2) symmetry},
432: Nonlinearity {\bf 1}, 617-652 (1988).
433: %-----------------------------------------------------
434: \bibitem{BP90}
435: K.~B\"uhler and N.~Polifke,
436: {\it Dynamical behaviour of Taylor vortices with superimposed axial flow}, in {\it Nonlinear evolution of spatiotemporal structures in dissipative continuous systems}, edited by
437: F.~Busse and L.~Kramer (Plenum Press, New York, 1990), p. 21.
438: %----------------------------------------------------
439: \bibitem{RL93}
440: R.~Raffa\"i and P.~Laure,
441: %{\it The influence of an axial mean flow on the Couette-Taylor Problem},
442: Eur. J. Mech., B/Fluids, {\bf 12}, $n^\circ$ 3, 277-288 (1993).
443: %----------------------------------------------------
444: \bibitem{AMM06}
445: M.~Avila, A.~Meseguer, and F.~Marques,
446: %{\it Double Hopf bifurcation in corotating spiral Poiseuille flow},
447: Phys. Fluids {\bf 18}, 064101 (2006).
448: %----------------------------------------------------
449: \bibitem{coefficients}
450: $b=-2.45\,10^{-3}, c=-1.06\,10^{-2}, e=-3.25\, 10^{-2},\mu^{(1)}=2.27\,
451: 10^{-3}, c^{(1)}=2.8\,10^{-2}$.
452: %----------------------------------------------------
453:
454: \end{thebibliography}
455:
456:
457:
458: %----------------- Figures-------------------------------------------------
459: %----------------- FIG 1 --------------------------------------------------
460: \begin{figure}
461: \includegraphics[clip=true,width=16.7cm, angle=0]{./fig1.eps}
462: \caption{\label{FIG-Bifurcation-Diagram}(Color) Bifurcation diagrams
463: of L-SPI $(A \neq 0, B=0, D>0)$, R-SPI $(A=0, B \neq 0, D<0)$,
464: L-CR-SPI $(|A|>|B|, D>0)$, R-CR-SPI $(|B|>|A|, D<0)$ and RIB (A=B,
465: D=0 -- only for $Re=0$) as functions of $\mu$ and $R_2$
466: for $R_1=240$. Here, $A$ and $B$ are the amplitudes of
467: the dominant modes $u_{2,1}(t)$ and $u_{2,-1}(t)$ (\ref{EQ:u2pm1}),
468: respectively, in the radial
469: velocity field $u$ at mid-gap. Solid (open) symbols denote stable
470: (unstable) solutions of the full Navier-Stokes equations. Full
471: (dashed) lines are the stable (unstable) solutions of the coupled
472: amplitude equations. Red, orange, violet, and magenta show
473: results for $Re=0.02$. The color coding of the different solutions
474: is given in the schematic 3D bifurcation diagram of (a). Therein the
475: black dots denote $Re=0$ solutions without stability information.
476: In (b) we show the projection of bifurcation diagrams onto the
477: $|A|^2-|B|^2$ plane. The color green refers to L-CR-SPI for $Re=0.1$
478: where R-CR-SPI do not exist anymore. Black symbols and lines in (b)-(d) refer to $Re=0$.}
479: \end{figure}
480:
481: %----------------- FIG 2 --------------------------------------------------
482: \begin{figure}
483: \includegraphics[clip=true,width=8cm, angle=0]{./fig2.ps}
484: \caption{\label{FIG-Bifurcation-Diagram-S-D}
485: (Color) Bifurcation behavior of
486: the vortex states of Fig.~\ref{FIG-Bifurcation-Diagram} are shown here in
487: plots of $S$ (a) and $D$ (b) versus $\mu$ and $R_2$, respectively. The color
488: coding is the same as in Fig.~\ref{FIG-Bifurcation-Diagram}.}
489: \end{figure}
490:
491: %----------------- FIG 3 --------------------------------------------------
492: \begin{figure}
493: \includegraphics[clip=true,width=8cm, angle=0]{./fig3.ps}
494: \caption{\label{FIG-power-behavior}
495: Power law behavior of quantities
496: characterizing the break-up of the $Re=0$ pitchfork bifurcation from RIB to
497: CR-SPI versus through-flow strength $Re$ for $R_1=240$. $D_2$ and $S_2$ refer
498: to values at the location $\mu_2$ of the unperturbed bifurcation; $D_S$ and
499: $\mu_S$ are saddle node coordinates, cf.
500: Fig.~\ref{FIG-Bifurcation-Diagram-S-D}(b). Solid lines come from the full
501: fixed point equations (\ref{EQ-Fixpunkt-Re}); dashed lines refer to their
502: leading order approximations (\ref{EQ-S2-D2},\ref{EQ-muS-DS}).}
503: \end{figure}
504:
505: %----------------- FIG 4 --------------------------------------------------
506: \begin{figure}
507: \includegraphics[clip=true,width=8cm, angle=0]{./fig4.eps}
508: \caption{\label{FIG-throughflow} (Color online) Power-law variation of $S,D,|A|^2=S+D,|B|^2=S-D$ with through-flow $Re$ during the transformation from RIB (diamond) at $Re=0$ to CR-SPI (squares and triangles). The latter undergo a transition to SPI (circles) at large $Re$. Symbols denote results from the full Navier-Stokes equations for $R_1=240$, lines show fits explained in the text. Right column: $\mu = \mu_2$, i.e., the unperturbed pitch fork location; left column $\mu=0.011$, i.e., far away from it. Note the different abscissa scaling of the two columns.}
509: \end{figure}
510:
511: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
512: \end{document}
513: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
514: