Você está na página 1de 16

Nonlinear Dynamical Systems,

Their Stability, and Chaos Amol Marathe


Birla Institute of Technology and Science,
Lecture notes from the FLOW-NORDITA Summer School Pilani 333031, India
on Advanced Instability Methods for Complex Flows, e-mail: marathe.amol@gmail.com
Stockholm, Sweden, 2013
Rama Govindarajan
This introduction to nonlinear systems is written for students of fluid mechanics, so connec- TIFR Centre for Interdisciplinary Sciences,
tions are made throughout the text to familiar fluid flow systems. The aim is to present how Tata Institute of Fundamental Research,
nonlinear systems are qualitatively different from linear and to outline some simple proce- Narsingi, Hyderabad 500075, India
dures by which an understanding of nonlinear systems may be attempted. Considerable e-mail: rama@tifrh.res.in
attention is paid to linear systems in the vicinity of fixed points, and it is discussed why this
is relevant for nonlinear systems. A detailed explanation of chaos is not given, but a flavor
of chaotic systems is presented. The focus is on physical understanding and not on mathe-
matical rigor. [DOI: 10.1115/1.4026864]

1 Introduction compared to attached flows such as that shown by the black line
(continuous), so the latter will be manifested.
What distinguishes a nonlinear system from a linear one? For a
Second, nonlinear systems can display finite-time singularities,
stability theorist, a gut reaction on being confronted with a nonlin-
which change the qualitative behavior of the solutions. The worst (or
ear system is to first linearize it about some equilibrium point if
best) thing about nonlinear systems is that the solutions do not add-
he/she can find one, and then study the linear system in detail [1].
up. The dependent variable in a linear partial differential equation
This is an excellent approach for many purposes. When a system
can be Fourier transformed in one of the variables, and you can then
goes from laminar to turbulent or from periodic to chaotic as a
solve for each Fourier mode in isolation. The terms in a nonlinear
control parameter is increased, the first step in this process is often
equation can be Fourier transformed, but every mode will depend on
linear. However, to understand the entire process, we need to
every other. This makes nonlinear systems harder to solve but also
understand how nonlinearities change the answers. Most systems
far richer than linear systems. For example, a cascade or inverse cas-
are nonlinear, and even if a given system is not chaotic, nonlinear-
cade of turbulent kinetic energy, creating small or large scales, is
ity can, and often will, play a big role in the dynamics. We begin
only possible due to different modes exchanging energy.
by discussing how nonlinear systems are completely and qualita-
The most interesting difference between linear and nonlinear
tively different from linear systems. First, nonlinear systems do
systems is that the latter can display chaotic dynamics. We shall
not need to obey the same rules about existence and uniqueness of
distinguish in the section on chaotic dynamics the fundamental
solutions that linear systems do. Let us discuss a counter intuitive
difference between a chaotic system that is deterministic and sim-
example, in a boundary value problem. The FalknerSkan
ply a noisy system.
equation,
We shall discuss several ways to treat a nonlinear system. It is
instructive to first look for equilibrium points, also known as fixed
f 000 ff 00 b1  f 02 0 (1)
points, or steady solutions. These are solutions of the system that
do not change with time.
describing the nondimensional streamfunction f g in the incom-
pressible boundary layer over a solid wedge placed at angle bp to
a flow, is a third-order differential equation in a nondimensional 2 Fixed Points and the Behavior of the System
normal distance g from the solid wall. There are three boundary
conditions, f 0 f 0 0 0, and f 0 1 1. In a linear third- in Their Vicinity
order system with three boundary conditions, we would get either We have already mentioned linearizing the system about a fixed
a unique solution or no solution. The FalknerSkan equation, point and studying the linear stability of the fixed point solution.
however, displays at least two solutions for each b < 0; an exam- A first-order nonlinear system may be written in the form
ple is shown in Fig. 1. Such multiple solutions are not too unusual
in nonlinear boundary value problems such as this one. In con- x_ f x (2)
trast, dynamical systems are most often posed as initial value
problems. In an initial value problem, we have only now way to where x is an n-component vector, and the overdot indicates a
march forward and will obtain a unique solution. However, the time derivative. It is sufficient to study a vector first-order differ-
fact that multiple solutions exist in boundary value problems is ential equation of this kind since, without loss of generality, an
relevant to us. To understand which of these solutions is mani- nth order differential equation can be turned into n first-order
fested in a real flow, we must linearize the NavierStokes equa- equations and written in vector form, as above. A dynamical sys-
tions about each of these solutions and solve the resulting tem is usually presented in the form of an initial value problem,
dynamical system. Usually separated flows (where the flow next with the initial conditions
to the wall is in the direction opposite to the average flow) such as
the profile shown by the red line (dashed) are very unstable x0 x (3)

Second, a system such as that given by Eq. (2), where time t does
Manuscript received July 8, 2013; final manuscript received December 6, 2013; not appear explicitly, is called an autonomous system. An
published online March 24, 2014. Assoc. Editor: Ardeshir Hanifi. n  1th order nonautonomous system, where time appears

Applied Mechanics Reviews Copyright V


C 2014 by ASME MARCH 2014, Vol. 66 / 024802-1

Downloaded From: http://appliedmechanicsreviews.asmedigitalcollection.asme.org/ on 12/29/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use


Fig. 1 A sample solution of the velocity profiles in a boundary
layer, as described by the FalknerSkan equation. Both solu-
tions satisfy all three boundary conditions. Here b 5 20:1. The Fig. 2 Standard fixed points and their phase portraits, along
solid line corresponds to f 00 0 5 0:1644, whereas the dashed with their canonical linear equations. Made using1. The tiny
line corresponds to a separated velocity profile, with arrows indicate the direction in which the solutions move as
f 00 0 5 20:0545. time progresses.

explicitly, can be turned into an nth order autonomous system by


adding time as another component of x; xn t, giving x_n 1.
The solution of an n-dimensional system of the form of Eq. (2)
may be plotted in n-dimensional space known as phase space. We
remark here that trajectories cannot cross in phase space. In other
words, a given equation can have more than one solution as we
have seen, but a given point in phase space can belong to at most
one solution. Each point in phase space can be thought of as an
initial value for the rest of time, so there in a unique way forward,
as we saw earlier. Solutions of a two-dimensional system may be
plotted in a phase-plane. Representative trajectories beginning
from different initial conditions, plotted in phase space, make up a
phase portrait.
Let us suppose that x0 is a fixed point of the above system, i.e.,

f x0 0 (4)

so the system, if initially placed at the fixed point, stays there for
all time. In its immediate neighborhood, we may write
x x0 dxt, where the magnitude of the perturbation is small,
neglect terms containing powers of dx greater than 1 and obtain a
linear system in dx. We drop the d for convenience in future dis- Fig. 3 A cylindrical blob of fluid displaced slightly from its
cussions of linear systems. From the eigenvalues of the linear sys- original location in a stratified fluid will display simple har-
tem, the fixed point may be classified, for example, as a stable or monic motion in the absence of diffusion. The z-axis is upwards
unstable spiral point, a stable or unstable node, a saddle point or a and density increases as z decreases.
center. There are also some other unusual fixed points that we
shall not discuss here. Simple equations that support fixed points
of each type, and phase portraits in the vicinity of the fixed-points, x_1 x2 (5)
are shown in Fig. 2. Such fixed points may be obtained for simple s
systems on a plane, for example, by the MATLAB software available g dq
x_2 N 2 x1 ; where N  (6)
on the Rice University website,1 as done here. Arrows may be q0 dz
drawn on the trajectories shown in the phase portraits to show the
direction of time. Trajectories approach a stable (attracting) fixed
is the BruntVaisala frequency, g the magnitude of acceleration
point, while they emerge out of an unstable or repelling point. The
due to gravity and the z-coordinate increases upwards. If the fluid
reader may wish to attempt the following exercise.
is bottom heavy, the density stratification is negative, and we have
Exercise (i) Work out the connection between the eigenvalues
a real N. This will result in oscillatory motion. On the other hand,
of a system and the nature of the fixed point.
a top-heavy system, with dq=dz > 0, will result in imaginary val-
A simple example of a linear system is that of Fig. 3. The den-
ues of N, one positive and one negative. It is easy to see that in
sity q of the fluid shown is stratified so as to be heavier at the bot-
this case, the perturbation will grow unboundedly.
tom. A small blob of fluid of density q0 is displaced along the
Exercise (ii) It is left to the reader to solve this simple system,
vertical direction z by a small distance dz to a location where the
and to then include a viscous damping term proportional to x_1 in
density of the blob is different from that of the surrounding fluid
Eq. (6), to get a center and a stable spiral, respectively. The direc-
by dq dq=dzdz. Putting dz x1 , the dynamics may be given
tion of density stratification can then be reversed to get an unsta-
by
ble system.
Fixed points whose associated linear systems display eigenval-
1
http://math.rice.edu/dfield/:dfield8.m and pplane8.m. ues with nonzero real part are known as hyperbolic and those with

024802-2 / Vol. 66, MARCH 2014 Transactions of the ASME

Downloaded From: http://appliedmechanicsreviews.asmedigitalcollection.asme.org/ on 12/29/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use


a zero real part of the eigenvalues are nonhyperbolic. Nonhyper-
bolic fixed points are sometimes referred to as elliptic. The
HartmanGrobman theorem assures us that the behavior of a non-
linear system close to a hyperbolic fixed point is very similar to
that of the linear system that approximates it. A saddle-point, for
example, will remain a saddle in the nonlinear system as well, and
moreover, the invariant manifolds emanating from it will look
conformally the same as their linear approximates. Thus, a hyper-
bolic fixed point is robust when faced with the addition of nonlin-
ear terms. However, an elliptic point, or center, of a linear system
is a fragile fixed point. The nonlinear system that it approximates
could display a stable or unstable spiral instead or change the local
flow in other ways. Once we know all the fixed points in a given
domain, if none of them is a center, we may then obtain the quali-
tative features of the nonlinear phase space by knowing the linear Fig. 5 The flow past a cylinder of square cross section. At a
behavior in each neighborhood. Reynolds number Re 5 30, a steady bubble is formed behind
Figure 4 is an example of a nonlinear system with several fixed the cylinder. A Hopf bifurcation occurs at a Reynolds number of
points in the domain shown. The equations solved are about 46, and oscillatory states with Karman vortex streets are
evident at Re 5 50 and 100. Note that the length scale in differ-
x_ y  y3  0:25x x3 y (7) ent at Re 5 30. Figure courtesy: Srikanth Toppaladoddi.
3
y_ x  y  xy xy (8)
Bifurcations of various types frequently occur in fluid mechanical
situations. The transition from a laminar (ordered) state to a turbu-
An examination of this figure reveals that the stable node has a lent (chaotic) state is an extremely complicated one, but in some
basin of attraction, i.e., a region in phase space inside which flows, some parts of this process occur via bifurcations. As one
every trajectory leads to that fixed point. This is characteristic of example, consider a cylinder placed with its axis perpendicular to
every attracting fixed point. There are basin boundaries separating the direction of flow. At low Reynolds numbers, the flow is
these basins of attraction. Repelling fixed points sit on the basin steady. At slightly higher Reynolds numbers, an oscillatory flow
boundaries. A saddle point on the basin boundary contains the is displayed behind the cylinder, with an alternate row of vortices
boundary itself as its attracting direction and has repelling directions moving downstream, as shown in Fig. 5. This transition from a
within the basins of attraction of (typically) two different attracting steady flow to an oscillatory flow is the result of a Hopf bifurca-
fixed points. A trajectory beginning exactly on the basin boundary tion [5]. A Hopf bifurcation is characterized by a change from a
will travel towards the saddle point, and reach it in infinite time. In
steady state to an oscillatory state as a particular parameter, in this
shear flows we often have a situation when the laminar state is line- case the Reynolds number, crosses a critical value. Moreover, as
arly stable but turbulent states exist as well. Here the laminar and tur- is seen in the figure, the amplitude of oscillation gets larger as the
bulent states may each be thought of as attracting fixed points, with control parameter moves further away from critical. At higher
their own basin boundaries. Recent studies, e.g., Refs. [24] find that Reynolds numbers than those shown, other instabilities occur,
the basin boundary between the laminar and turbulent states usually until the wake of the cylinder becomes turbulent. We shall discuss
contains one such saddle, which is termed the edge state. more about the transition to turbulence in the context of a bound-
Exercise (iii) Design a pair of simultaneous nonlinear first-order ary layer below, but let us first discuss some standard bifurcations.
differential equations that contain several fixed points. Plot the The canonical equations for some standard bifurcations in sim-
phase space and mark the basin boundaries. What happens if one ple nonlinear systems and the type of bifurcation that results from
of the fixed points is a center in the linear limit and not in the non- the change of a parameter r are shown in Figs. 6 to 9. A transcriti-
linear limit? cal bifurcation occurs when there are two fixed points, one stable
and one unstable. These exchange their stabilities across the bifur-
cation. Phase portraits for different values of the control parameter
3 Bifurcations r are shown in Fig. 6. In fluid flows, r is the Reynolds number, the
A bifurcation point is one where the dynamics changes in
character. Why are we interested in studying bifurcations?

Fig. 6 Phase portrait in a transcritical bifurcation. The


exchange of stabilities between the two fixed points as r
Fig. 4 A phase plane of a nonlinear system with multiple fixed crosses zero is evident in this figure. In this and subsequent
points. There are two saddles, at (0,0) and (1.1, 21.59), a stable figures, filled circles correspond to stable fixed points, and
node, at 20:59; 20:77, and an unstable node at (0.82,1.17). open circles indicate unstable fixed points.

Applied Mechanics Reviews MARCH 2014, Vol. 66 / 024802-3

Downloaded From: http://appliedmechanicsreviews.asmedigitalcollection.asme.org/ on 12/29/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use


Rayleigh number etc. At the value of r where the bifurcation fork bifurcation is called by that name! In the supercritical pitch-
occurs (here r 0) the stable fixed point loses its stability while fork, a stable fixed point becomes unstable at the bifurcation
the unstable one gains stability. There is thus an exchange of point, but two other stable fixed points are generated, as seen. The
stabilities, as shown in Fig. 7. A saddle-node bifurcation, shown subcritical pitchfork is the same except for a sign change in all the
in Fig. 8, goes by many other names as well. Here a stable and an growth rates. The two unstable fixed points can be stabilized (and
unstable fixed point are neutralized at the bifurcation, so for val- often are, in real systems) by a fifth-order term of the opposite
ues of r beyond the bifurcation point, there are no fixed points. It sign, and such systems display hysteresis. This is an interesting
is not evident in a one-dimensional system why a saddle-node feature of nonlinear systems that is not discussed further here, but
bifurcation goes by this name (this becomes evident in a two- occurs in thermoacoustic [6] and many other systems.
dimensional system), but it is very clear from Fig. 9 why a pitch- In a continuation of exercises from the previous section:
Exercise (iv): Plot trajectories in phase space on either side of
(a) a transcritical bifurcation, (b) a saddle-node bifurcation, (c) a
supercritical pitchfork bifurcation, (d) a subcritical pitchfork
bifurcation with a higher-order stabilization.
Exercise (v): Plot trajectories in phase space on either side
(negative and positive r) of a Hopf bifurcation, given by the
equations

a_ ra  a3 ; h_ 1 a2 (9)

Here a and h are the radial and angular coordinates, respectively.


At r < 0, the steady state a 0 is attained. As r crosses zero, an
oscillatory state, characterized by constant angular velocity
motion on a circle, is attained at large times. Notice that the am-
plitude of oscillation, i.e., the radius of the circle, increases as the
Fig. 7 The nature of fixed points in a transcritical bifurcation.
As r crosses zero, the two fixed points exchange stabilities. In
square root of the parameter r. The Hopf bifurcation can be
such diagrams solid lines correspond to stable fixed points, thought of as an oscillatory version of the supercritical pitchfork.
whereas dashed lines show how unstable fixed points move as In the Hopf bifurcation seen in Fig. 5, the Reynolds number Re
r varies. plays the role of r in the above exercise.
Can we study a simple equation that displays a Hopf bifurca-
tion, and therefore some of the features of the flow past a cylin-
der? A canonical nonlinear system that a fluid mechanicist can
identify with is the van der Pol oscillator, which is described in
Sec. 4. Before that we return to the boundary layer, to contrast it
with the flow past a cylinder, and to reiterate that transition to tur-
bulence is far more complicated than just a series of bifurcations
that can be modeled by simple equations. While the steady state
in the flow past a cylinder gave way to self-sustained oscillations
of a certain frequency, the onset of the Tollmien-Schlichting insta-
bility in a boundary layer (at low free-stream turbulence levels)
marks a transition from a steady state to an oscillatory state. This
seems at first sight similar to a Hopf bifurcation, but there is a ba-
sic difference. These oscillatory solutions are not limit cycles. To
distinguish this basic difference, flows such as this one are
referred to as amplifiers, where a perturbation of a given fre-
quency can be made to amplify by the system. The oscillations
grow in time with a very small growth rate, so the amplitude is
not a fixed function of Reynolds number. At every Reynolds num-
Fig. 8 Phase portrait for a prototypical saddle-node bifurca- ber, or equivalently, every downstream location, the growth rate is
tion. The two fixed points annihilate each other. different. These experimentally observed oscillations are well
predicted by linear theory, whereas a Hopf bifurcation is the result
of a nonlinear saturation to a new steady state. Figure 10 shows
Roddam Narasimhas sketch of the transition process at low
free-stream turbulence levels. The Reynolds number increases
downstream, so the entire route to turbulence can be seen in one
snapshot. The straight solid lines (green in color) represent peaks
in amplitude of the TollmienSchlichting waves that appear peri-
odically and move downstream. On the other hand, globally unsta-
ble systems such as that of Fig. 5 are called oscillators, which
display their inherent frequency and amplitude.
The growth of TollmienSchlichting waves is very slow, so at a
given time and Reynolds number, we may approximate the flow
to one with constant amplitude oscillations. Under certain condi-
tions of free-stream turbulence, this oscillatory state can be fol-
lowed by another instability, to give a new state termed as the
Klebanoff mode, shown by the wavy green lines. This instability
is called the secondary instability and can be predicted by the Flo-
quet approach described later in this article. The new state is still
Fig. 9 Subcritical and supercritical bifurcations; this figure is periodic but has a wavy structure in the spanwise direction. Note
taken from Wikipedia that the term instability is more appropriate here than

024802-4 / Vol. 66, MARCH 2014 Transactions of the ASME

Downloaded From: http://appliedmechanicsreviews.asmedigitalcollection.asme.org/ on 12/29/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use


wires in the wind and several other situations. Wherever one finds
a self-sustained oscillation, one immediately thinks of the van der
Pol oscillator. Other reviews in this issue discuss flows that are
globally unstable, and it may be seen that they often display self-
sustained oscillations of a characteristic frequency. The frequency
and the shape of the disturbance structure are independent of the
details of the disturbance environment. This is a completely non-
linear phenomenon, more commonly known as a stable limit cycle
in phase space. A van der Pol oscillator displays a limit cycle. The
frequency of this limit cycle may be estimated from the linear
approximation, as we shall see. This will give us a basic idea of
how to obtain the periodicity of a nonlinear system by expanding
about the linear answer.
The differential equation

x x2  1x_ x 0 with 0 <  (10)


Fig. 10 Top view of the boundary layer on a flat plate, sketch
of Roddam Narasimha. The thick vertical line marks the transi-
tion from periodic to chaotic flow. describes the van der Pol oscillator (see, e.g., Ref. [9]). This
belongs to a class of planar systems, i.e., systems modeled as
second-order ordinary differential equations, where a special peri-
bifurcation, although the character of the flow has changed from odic solution as described above, namely a limit cycle, exists in the
one type of oscillations to another. One reason is that the new x-x_ plane. (Note that in the phase plane x r cos h.) In Fig. 12, the
oscillations do not constitute limit cycles either. Moreover, the origin is seen to be an unstable equilibrium point of Eq. (10). With
change can occur over a range of Reynolds numbers rather than a  > 0, a trajectory starting sufficiently close to the origin (let us say
sharply identified Reynolds number, depending on the nature and at x; x
_  0:5; 0, shown by the blue dashed-dotted line) will spi-
levels of disturbance in the external flow. As we go further down- ral outward away from the origin, since for this initial condition,
stream a host of not completely understood changes follow, until the damping term has a negative coefficient, and so acts to
turbulent spots appear. The turbulent spots grow and merge as amplify. Note that the damping is given by the coefficient of x. _ On
they move further downstream until the entire flow is turbulent. the other hand a trajectory that starts with an initial condition suffi-
At a slightly higher free stream turbulence than that shown in ciently far from the origin, (e.g., at x; x_  5; 0, shown by the
Fig. 10, we get, instead of the Klebanoff mode, another spanwise pink solid line in the figure), spirals inwards since the damping is
structure called the Craik mode, where the wavy patterns are positive. The two trajectories cannot intersect each other, as dis-
arranged as shown in Fig. 11. It is seen that the pattern is repeated cussed above. Both asymptotically approach the same isolated,
once every two amplitude-maxima, so the period has doubled as simple, closed curve in the phase-plane, in other words, the limit
compared to the previous stage of two-dimensional Tollmien cycle. This curve divides the phase plane into two regions. If
Schlichting waves. We shall talk about period doubling in the instead  in Eq. (10) were to be less than 0, we would have an
context of the logistic map. unstable limit cycle. In this case, the origin will be stable and trajec-
tories from the neighborhood of the limit cycle will tend to go
4 The van der Pol Oscillator away from it as t ! 1.
Limit cycles are to be found in systems that are much more gen-
In Sec. 3, we saw how vortices of a certain amplitude are shed eral than the van der Pol oscillator. How do we characterize a
behind a cylinder at a particular frequency, when the cylinder is limit cycle quantitatively and qualitatively? We need its shape
placed in an oncoming flow at a moderate Reynolds number. The and size as well as its stability.
Hopf bifurcation that occurs there bears similarity to a far simpler
counterpart: the van der Pol equation. The equation and its var-
iants have been used to model lift and drag in such Karman
streets, the flow past a slender cone [7], flow past airfoils with
drag-reducing hair like structures [8], the whistling of electric

Fig. 12 Phase plane of van der Pol oscillator with  5 0.1. The
Fig. 11 A schematic of the Craik mode of secondary instabil- dash-dot trajectory emerges out of the origin and approaches
ity: an example of spatial period doubling. The lines show max- the limit cycle, while the pink line begins from infinity and
ima in the disturbance amplitude, which move downstream as moves inwards towards the limit cycle. The limit cycle is where
well as in the spanwise direction in a sinusoidal manner. the pink and blue trajectories meet.

Applied Mechanics Reviews MARCH 2014, Vol. 66 / 024802-5

Downloaded From: http://appliedmechanicsreviews.asmedigitalcollection.asme.org/ on 12/29/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use


4.1 Period of the Limit Cycle by LindstedtPoincare x0 t A cost (13)
Method. We wish to obtain the time period of one traversal of the
limit cycle for the van der Pol oscillator. We approach this prob- Substituting this into the equation at O; we get
lem via a singular perturbation method called the
LindstedtPoincare. This is a routinely applied method to con- 1 1
x001 x1 A2  1  2x1 A coss A2 cos2s
struct an analytical but approximate periodic solution of a nonlin- 2 2
ear ordinary differential equation.
For applying the perturbation method, we further assume Solving the above for x1(s) gives us the term that grows secularly,
0 <   1. As usual with perturbation methods [9, 10], we i.e., unboundedly in time. This is because we now have a forcing
expand the solution in a power series in , i.e., term, 2x1 A coss, whose frequency [1=(2p)] is the same as
the natural frequency of the oscillator. Such a forcing term is
xt x0 t x1 t 2 x2 t    (11) termed the secular term. As x1(s) grows, after a time O1= (in s
units), the expansion will break down because x1 s will be as
We expect the above expansion to be uniformly valid, i.e., each big (or as small) as x0(s). In order to make the expansion uni-
term added to the series is one order smaller than the previous formly valid, we make the coefficient of coss equal to zero. This
term [11]. Given the assumption imposed on , this is equivalent key step, peculiar to singular perturbation methods, is called the
to saying xk t  O1 for every k and for all time t > 0. removal of secular terms [9]. This gives
Let T be the time period of the periodic solution we are looking
for, and x 2p=T be the corresponding frequency. The key idea x1 0
is to expand x as a power series in  in terms of unknown coeffi-
cients x1 ; x2 , etc. Such an approach is also called the method of After the removal of the secular term, we solve O equation for
strained parameters [10, 12]. We write x1 s with x1 0 0 and x_ 1 0 0 to get
 
x 1 x1 2 x2    (12) 1 2 1 2 1
x1 s 1  A A  1 coss A2 cos2s (14)
2 3 6
where x0 1 is the frequency of the unperturbed oscillator, i.e.,
with  0. This is the most important step in the Lindstedt- Substituting for x0 s from Eq. (13) and x1 s from Eq. (14) into
Poincare method. The unknown xi s are to be found using equation at O2 ; we get
conditions that make the expansion uniformly valid. One of the  
significant differences between linear and nonlinear oscillators is 00 5 3
the independence and dependence, respectively, of their time peri- x2 x2 2A 2x2 A A coss
6
ods on the amplitude of oscillation. The next step in the process is  
1 3 1
to stretch the time by a change of the independent co-ordinate A  A 1 cos2s A3 cos3s
from t to s using 3 6

s xt Removal of another secular term now gives

We rewrite Eq. (10) in stretched time as 5 2


x2 1  A
12
x2 x00 xx2  1x0 x 0
Thus, we obtain an amplitude dependent frequency and the time
where prime denotes differentiation with respect to s. Substituting period of the van der Pol oscillator, correct up to O3 , respec-
x from Eq. (12) and using Eq. (11) in the above equation, collect- tively, as
ing terms of various powers of , we obtain    
5 5
x 1 2 1  A 2 and T 1  2 1  A2
O0 : x000 x0 0 12 12
O : x001 x1 x0 x20  1 2x1 x000
We recall that the van der Pol equation is a good model for
O2 : x002 x2 x20  1x1 x00 x01  2x0 x1 x00  2x1 x001 describing the dynamics on a cylinder in a range of Reynolds
 2x2 x21 x000 numbers, and we can thus find an application for estimating its
frequency of oscillation. There are other applications.
Exercise (vi): Using the above method, obtain the period of the
We solve Eq. (10) for the initial condition (IC) following nonlinear oscillator up to O2 :
x0 A and _
x0 0 x x x2 2 x3 0

_
We choose x0 0, and we can do this without loss of generality, 4.2 Stability of the Limit Cycle via Method of Multiple
as the van der Pol oscillator is autonomous. In other words, we can Scales. Now, we apply another singular perturbation method
choose the origin for time to correspond to a location anywhere on called the method of the multiple scales to the van der Pol oscilla-
the limit cycle, since the oscillator is autonomous. Assuming A to tor. The objective is to get an estimate of the size and shape of the
be of O1; we write down ICs for x0 t; x1 t;    as limit cycle in the phase plane and to know whether the limit cycle
x0 0 A; x1 0 0; x2 0 0;    is stable or not, meaning, do the nearby trajectories get attracted
toward the limit cycle or do they go away from the limit cycle?
and Note that we need not assume here that we are looking for a peri-
odic solution, as we needed to in the LindstedtPoincare method.
x_0 0 0; x_ 1 0 0; x_ 2 0 0;    We begin with the assumption that the solution x(t) we are
looking for depends on various time scales [13]
With above mentioned ICs, solving the O0 equation above, we
get T0 t; T1 t; T2 2 t;   

024802-6 / Vol. 66, MARCH 2014 Transactions of the ASME

Downloaded From: http://appliedmechanicsreviews.asmedigitalcollection.asme.org/ on 12/29/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use


where T0 is the fast time scale and T1 ; T2 ;    are slower ones, so 2
Rt v
!
we have u
u 4
t1  1 et
xt XT0 ; T1 ;    R02

Further the solution X is expanded as a power series in , As t ! 1; Rt ! 2; hence, the limit cycle is stable.
2 It is suggested to the reader that they may consult Ref. [13] to
XT0 ; T1 ;    X0 T0 ; T1 ;    X1 T0 ; T1 ;    O (15) study the method of multiple scales further. To know more about
the van der Pol oscillator and limit cycles, refer to Ref. [14].
The derivatives with respect to t become Let us now consider another canonical nonlinear system, the
Duffing oscillator, and apply the methods we have discussed
d @ @ above to understand it. The Duffing oscillator is a simplified ver-
 O2 (16)
dt @T0 @T1 sion of a simple pendulum, and we will see that under certain con-
d2  @ 2  @ 2  ditions it can display chaotic behavior.
2
2 O2 (17)
dt2 @T0 @T0 @T1
5 Duffing Oscillator, Unforced, and Forced
Substituting Eqs. (15) through (17) in Eq. (10) and expanding and The motion of a simple pendulum is governed by
collecting terms, we obtain at O1 and at O
g
h sin h 0
@ 2 X0 l
X0 0 (18)
@T02
where g is the acceleration due to gravity and l is the length of the
3 5
@ 2 X1 @ 2 X0 @X0 pendulum. Expanding sin h h  hp Oh , changing the
=6
X1 2  X02  1 (19) dependent variable from h to x h= , stretching the independ-
@T02 @T1 @T0 @T0
ent variable from t to s g=l1=2 t, and ignoring O2 terms, we
get
respectively. Remembering that X0 is a function of more than one
variable, Eq. (18) can be solved for X0 as
x00 x ax3 0
X0 AT1 ; cos T0 BT1 ; sin T0
where a 1=6 and prime denotes differentiation with respect to
Substituting for X0 from the above into Eq. (19), we get s. This is the equation for the Duffing oscillator, one of the sim-
plest nonlinear oscillators. The LindstedtPoincare method gives
  the time period T (correct up to O2 ) as
@ 2 X1 @A A 2 2
X 1 2  A A B sin T0  
@T02 @T1 4 3
  T 2p 1   aA2
@B B 2 2 8
2 B  A B cos T0
@T1 4
A 2 B indicating again an amplitude dependence of the period. This am-
A  3B2 sin 3T0 B2  3A2 cos 3T0 plitude dependence is an essential feature of nonlinear oscillators
4 4
and differentiates nonlinear systems from linear ones.
Removing the secular terms, i.e., equating coefficients of the reso- We now consider the damped, externally forced Duffing
nant terms to zero, we get the slow flow oscillator

@A A A 2 x dx_ bx ax3 c cosxt (20)


 A B2
@T1 2 8
@B B B 2 For a 0, the oscillator is linear and has a unique steady state so-
 A B2 lution. For each set of initial conditions, the solution settles down
@T1 2 8
to a unique amplitude. Its frequency is the same as the forcing fre-
quency and its amplitude depends upon the forcing amplitude c,
Transforming to polar coordinates A R cos /; B R sin / and
damping coefficient d and linear stiffness b. Nonlinear systems
using
(e.g., the Duffing oscillator in the a 6 0 case) differ from linear
@A @B systems as they may settle down to multiple periodic solutions,
A_  and B_  stable or unstable, depending upon initial conditions. Figure 13
@T1 @T1 shows two such stable solutions found numerically. The frequency
of each is the same as the forcing frequency, but the solutions
we get, up to O2 , differ in amplitudes, which are decided by their initial conditions,
  here IC1, i.e., x0; x0
_ 1; 0 and IC2, i.e., x0; x0
_
R R3 7; 0. Every stable periodic solution has its own basin of
R_   and /_ 0
2 8 attraction, i.e., a set of initial conditions from which the system
settles down to that particular periodic solution.
Slow flow equilibria correspond to periodic solutions of the sys- We now use the heuristic analytical but approximate method
tem. For the van der Pol oscillator, these are of harmonic balance (MHB) in conjuction with a numerical
arclength-based continuation method to obtain the nonlinear har-
monic response of the Duffing oscillator. MHB, a special case of
R 0 and R2
Galerkin projections, approximates periodic solutions of nonlinear
oscillators using a truncated Fourier series representation. Here,
R 2 is nothing but the limit cycle amplitude as can be seen from we assume the periodic solution to be a single-term harmonic bal-
Fig. 12. We solve equation for R_ with IC R(0) to get ance solution of the frequency of forcing [15],

Applied Mechanics Reviews MARCH 2014, Vol. 66 / 024802-7

Downloaded From: http://appliedmechanicsreviews.asmedigitalcollection.asme.org/ on 12/29/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use


Fig. 13 Multiple periodic solutions for the forced Duffing oscilla-
tor with a 5 1; d 5 0:1; b 5 1; c 5 1; x 5 2:5 with initial conditions
IC1 5 (1,0) (small amplitude) and IC2 5 (7,0) (large amplitude). Fig. 14 Harmonic response of the Duffing equation for differ-
ent strengths of nonlinearity a

x  A sinxt B cosxt

Substituting the above into Eq. (20), multiplying by sinxt and


cosxt, then integrating with respect to t from 0 to 2p=x (which
is nothing but collecting coefficients of sinxt and cosxt), we
get two simultaneous nonlinear algebraic equations

3 3
 Ax2  dxB bA aA3 aAB2  c 0 (21)
4 4

and

3 3
 Bx2 dxA bB aB3 aA2 B 0 (22)
4 4

We can solve Eqs. (21) and (22) simultaneously for A and B for
some fixed values of other parameters. Using a numerical tech-
nique called arclength-based continuation (also called pseudo
arclength continuation
p method) [16, 17], we find the approximate Fig. 15 Nonlinear harmonic response of the Duffing equation
amplitude A2 B2 as a function of x, i.e., the nonlinear har- for different damping coefficients d
monic response of the system. Putting d 0 and c 0 in Eqs. 21
and 22, we obtain B 0 from Eq. (22) and then substituting so in
Eq. (21), we get
r
2 x2  1
A6 (23)
3 a

The above result though approximate in nature, tells us that for


undamped, unforced periodic oscillations, the frequency of oscil-
lations (and, hence, the period) depends on the amplitude. As we
have discussed, this is typical of nonlinear systems. Equation (23)
is the equation of a backbone curve (the continuous curve in
Fig. 16). The case of a > 0 is called a stiffening nonlinearity while
a < 0 case is a softening nonlinearity [18].
Figures 1416 show the effect of varying nonlinearity strength
a, damping d and forcing amplitude c, respectively, on the fre-
quency response of Eq. (20). As can be seen from Fig. 14, increas-
ing a while holding other parameters constant, the resonance
curve along with the backbone curve leans over to the right for a
stiffening nonlinearity and to the left (not shown) for a softening
nonlinearity. Figure 15 shows the effect of varying the damping d
while keeping other parameters constant. The hump in the ampli-
tude versus frequency curve follows the backbone curve in each
case, hence, the importance of the backbone curve. Decreasing d Fig. 16 Harmonic response of Duffing equation for different
raises the hump, i.e., raises the maximum possible response forcing amplitudes c

024802-8 / Vol. 66, MARCH 2014 Transactions of the ASME

Downloaded From: http://appliedmechanicsreviews.asmedigitalcollection.asme.org/ on 12/29/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use


amplitude for the given amplitude of harmonic forcing. For smaller Exercise (viii): On similar lines to the code given, write a code
and smaller values of d and c, the response curve draws closer and for the forced van der Pol oscillator
closer to the backbone curve. Figure 16 shows the effect of varying
forcing amplitude c while keeping other parameters fixed. As we x x2  1x_ x F sinxt
increase the magnitude of forcing amplitude c, the system response
Generate Poincare plot for parameter values
grows. The response follows the backbone as well.
 0:1; F 2; x 5.
Equation (20) is nonautonomous in nature since the independ-
ent variable t appears explicitly on the right hand side. Converting %Function le dufng.m
it to an autonomous equation will increase its dimension by 1, function xdotdufng(t,x)
making it three-dimensional. Therefore, the phase plane is an % dene variables globally,
inadequate tool to analyze this system. The vector field at a given %i.e., across the les
point changes in time, allowing a trajectory in the phase plane to
appear to intersect itself. The fact that the system is three- global gamma omega epsilon GAMMA OMEGA
dimensional allows, as mentioned above, for the possibility of %write 2nd order forced Dufng
chaotic behavior. In order to study the behavior, we now use a %oscillator as two rst order ODEs
Poincare section. Poincare sections [9] are routinely used to
xdot(1)-gamma*x(1)omega 2*x(2)-
reduce the dynamics of such three-dimensional systems to two by
epsilon*x(2) 3GAMMA*cos(OMEGA*t);
introducing a plane transverse to the flow, and marking all points
where a trajectory intersects this plane from a given direction of xdot(2)x(1);
approach. The map that takes the current intersection to the next xdotxdot;
is called the Poincare return map. Studying the map dynamics
helps us to gain insight into the original systems dynamics. For end
example, a fixed point of the Poincare map, where the trajectory %Script le Poincaresection_script.m
cuts the plane of interest at the same location every time, % dene variables globally, i.e.,
corresponds to a periodic motion of the system. A series of never- %across the les
repeating points corresponds usually to a chaotic case. Figure 17 global gamma omega epsilon GAMMA OMEGA
shows such a Poincare section for Eq. (20) generated using script
%numerical values of the parameters involved
file Poincaresection_script.m and function file duffing.m. The
gamma0.1; omega1; epsilon0.25;
function file writes the Duffing oscillator equation in a standard
OMEGA2; GAMMA1.5;
format so that built-in routine ode45 can numerically integrate
the oscillator equation. The script file plots the Poincare section. %number of cycles
For the specified parameter values, the system is chaotic and the N50000;
Poincare section is a fractal set. It is clear that while the points are %using ode45 numerical integration routine,
not repeating, they are not completely randomly distributed in %integrate forced Dufng oscillator
space. The points seem to all lie on a particular pattern occupying [t x]ode45(dufng,
a subset of phase space. Trajectories intersecting the plane some- linspace(0,N*2*pi/OMEGA,100*N),[0 1]);
where on this pattern clearly form an attractor for this dynamics.
This attractor is not a fixed point or a limit cycle, but the three- %plot Poincare section, x(nT)
dimensional extension of the fractal set we see here. Such an %against xdot(nT)
attractor is called a strange attractor, as we shall see. To study the Psectionplot(x(1:100:end,1),x(1:100:end,2),
Duffing oscillator further, the reader is suggested to go through r.);
Refs. [14,18,19]. %label the gure &
The reader, after successful completion of these exercises, will %set x & y label font and fontsize
be able to generate Poincare sections for different nonautonomous set(gca,fontsize,21);
oscillators. set(Psection,MarkerSize,2)
Exercise (vii): As shown in Fig. 17, plot the graph for the fol- xlabel(x(nT)), ylabel(x(nT));
lowing parameter values d 0:5; b 1:3; a 0:1125; c 2;
x 3:14 with the help of the following MATLAB files. 6 Floquet Theory and Hills Equation
When we talked about the Klebanoff and Craik modes, the ob-
servant reader would have wondered how to study the instability
of an oscillatory flow. We do this by Floquet theory, and a canoni-
cal example is discussed below.
6.1 Floquet Theory. Consider an n-dimensional, linear non-
autonomous system
x_ Atx (24)
where the dot denotes differentiation with respect to time t, and
A(t) is periodic in time with a fixed time period T, i.e.,
At T At for all t. Let X(t) be a fundamental solution ma-
trix (FSM) to this system, i.e.,
Xt x1 t; x2 t; ; xn t
where x1 t; x2 t; ; xn t are n linearly independent solutions
of Eq. (24).
It is straightforward to see that if x(t) is a solution to Eq. (24),
then so is x(t T). Since x(t) is a solution, we have

Fig. 17 Poincare  section for the forced Duffing oscillator for d


xt Atxt
d 0:1; b 1; a 0:25; c 1:38; x 2 dt

Applied Mechanics Reviews MARCH 2014, Vol. 66 / 024802-9

Downloaded From: http://appliedmechanicsreviews.asmedigitalcollection.asme.org/ on 12/29/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use


Replacing t by t T in the above equation, we get 6.2 Hills Equation. Consider an inverted pendulum of
length l and mass m as shown in Fig. 18, and that its pivot point is
d given a displacement u(t) in the y-direction. The angular displace-
xt T At Txt T
dt T ment h(t) is a generalized coordinate describing the motion of the
pendulum. Acceleration due to gravity g acts in the negative
Since dt T dt and A(t) is periodic with period T, we rewrite y-direction. Taking the origin of the coordinate system at some
the above as point on the h 0 line, the coordinates of the center of mass of the
pendulum are
d
xt T Atxt T x l sin h; y ut l cos h
dt

It follows that if X(t) is a FSM to Eq. (24), then so is X(t T). where u(t) is the imposed displacement of the pivot. Differentiat-
Therefore, these two must be related to each other via some con- ing with respect to time, we have
stant nonsingular matrix C. We then have
_
x_ l cos hh; y_ u_  l sin hh_
Xt T XtC
The Lagrangian [20] of the system is given by
~ we denote a special
Setting t 0, we get XT X0C. By Xt
~ I , where I denotes the n n
FSM to Eq. (24) such that X0 1  
~ L T  V m x_2 y_ 2 mgl cos h ut
identity matrix. Then we have C XT [9] and, therefore, 2
~
Xt T XtXT (25) and the EulerLagrange equation governing the angular displace-
ment h(t) is given by [20]
We now consider another FSM Z(t) to Eq. (24) such that  
d @L @L
 0 (27)
Zt XtR dt @ h_ @h

where R is arbitrary. Substituting for X(t) from the above equation which in the present case gives
in Eq. (25), we get  
u g
h  sin h 0
Zt T Zt R1 XTR
~ l l

~
We now choose R such that R1 XTR is diagonal.2 Then we Linearizing the trigonometric term for small h, we get
write n decoupled equations as  
u g
zi t T ki zi t; i 1; ; n (26) h  h 0
l l
~
ki is ith eigenvalue of XT. Each zi t is a linear combination of Assuming the pivot displacement to be harmonic, i.e.,
x1 t; x2 t; ; xn t. Equation (26) is a functional equation. We u  cost, the above equation becomes
assume its solution to be of the form
 
g  cost
zi t kkt
i pi t
h h0
l l
where pi t is some function periodic with period T, the same as
that of A(t). k is an unknown to be determined. Substituting the
assumed form of zi(t) into Eq. (26), we get k 1=T. Therefore,
t=T
zi t ki pi t

At the end of each time period, zi attains the values shown below:

t 0; zi 0 pi 0
t T; zi T ki pi 0 ki zi 0
..
.
t mT; zi mT km m
i pi 0 ki zi 0

~
Eigenvalues of XT, i.e., ki s, determine the long-term behavior
of solutions to Eq. (24). If jki j > 1 for some i, then the corre-
sponding solution grows unboundedly and the system is unstable.
If jki j < 1 for all i 1;    ; n, then all solutions remain bounded
as t ! 1 and the system is stable. The matrix T is called the
Floquet matrix and its eigenvalues ki s are called Floquet
multipliers.

2 ~ is diagonalizable.
We assume that Xt Fig. 18 Inverted pendulum, harmonically excited in y-direction

024802-10 / Vol. 66, MARCH 2014 Transactions of the ASME

Downloaded From: http://appliedmechanicsreviews.asmedigitalcollection.asme.org/ on 12/29/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use


The above may be considered as an oscillator whose frequency and the other decays exponentially. The solution in general
varies in a periodic fashion with respect to time. In a more com- grows and, hence, the system is unstable.
pact setting, we can write such an oscillator in a form ~
(2) jtrXTj 2: k1;2 1 or k1;2 1. For k1;2 1, out of
two linearly independent solutions x1 t and x2 t, one
x f tx 0 (28) grows linearly and the other is periodic with period T i.e.
the period of f(t). For k1;2 1, one grows linearly and the
with f t T f t for all t. Equation (28) is called Hills equa- other is periodic with period 2T. In both cases, the solution
tion, which appeared in a memoir on the motion of the lunar peri- in general grows and, hence, the system is unstable.
gee by Hill in 1886 [21]. ~
(3) jtrXTj < 2: k1 and k2 are a complex conjugate pair. The
We study the asymptotic stability of solutions to Eq. (28) with two lie on a unit circle since k1 k2 1. Both linearly inde-
the help of Floquet theory. Writing Eq. (28) as a system of two pendent solutions are periodic but with incommensurate
first-order ordinary differential equations, we get frequencies; so the solution in general is quasi-periodic.
( ) " #( ) The system is stable.
d x1 0 1 x1
Note that solutions x1 t and x2 t cannot be expressed in terms

dt x2 f t 0 x2 of elementary functions such as sine and cosine. However, since
Eq. (28) is linear, the principle of superposition holds and the
where x1 x and x2 x.
_ By most general solution is given by a linear combination of x1 t
and x2 t. Floquet multipliers move on the real line or on the unit
" # circle in the complex plane with respect to parameters of function
x11 t x12 t
Xt x1 t x2 t  f(t) as indicated in Fig. 19.
x21 t x22 t The aim of the following example is to obtain the Floquet mul-
tipliers of the Hills equation for a given set of parameter values.
we denote a fundamental solution matrix (FSM) to Eq. (28). By Example: if we set g l  1 for the inverted pendulum, we
definition, we have have
" # " #" #
1 1
d x1 t x2 t 0 1 x11 t x12 t h 1 costh 0 (29)

dt x21 t x22 t f t 0 x21 t x22 t
We integrate Eq. (29) in MATLAB with ode45 for a period T 2p
with ICs f1; 0gT and f0; 1gT using function file mathieu.m and
i.e., x_11 t x21 t; x_ 12 t x22 t; x_21 t f tx11 t and x_22 t script file FloquetMultipliers_script.m given below.
f tx12 t
We have %Function le mathieu.m
function xdotmathieu(t,x)
det Xt x11 tx22 t  x21 tx12 t %parameter values
delta 1;
Therefore, epsilon 1;
%Mathieu equation written as
d
det Xt x_11 tx22 t x11 tx_ 22 t  x_ 21 tx12 t  x21 tx_ 12 t %two rst order ODEs
dt xdot[x(2);-(deltaepsilon*cos(t))*x(1)];

Substituting for time derivatives from the above four equations, %Script le FloquetMultipliers_script.m
we get %set absolute and relative tolerances
optionsodeset(AbsTol,1e-7,RelTol,1e-8);
d
det Xt 0 and therefore det Xt constant % Numerical integration of Mathieu equation
dt %for IC (1,0)
~ be the FSM which evolves from the [t,x]ode45(mathieu,[0 2*pi],[1 0],options);
As denoted earlier let Xt x1x(end,:);
identity matrix at t 0. Therefore,

~ det X0
det Xt ~ 1 det XT
~

According to Floquet theory, the asymptotic stability of solutions


to the Hills equation is governed by the Floquet multipliers, i.e.,
~
the eigenvalues of XT. These satisfy the quadratic equation

~
k2  trXTk 10

~
where 1 det XT. So the Floquet multipliers are
q
~
trXT6 ~ 24
trXT
k1;2
2
~
Based on numerical value of trXT, we have the following cases:
~
(1) jtrXTj > 2: Both k1 and k2 are real. If k1 > 1, then
0 < k2 < 1 since k1 k2 1. The second possibility is
k1 < 1, so 0 > k2 > 1. Out of the two linearly inde- Fig. 19 Movement of Floquet multipliers with respect to
pendent solutions x1 t and x2 t, one grows exponentially parameters of f(t)

Applied Mechanics Reviews MARCH 2014, Vol. 66 / 024802-11

Downloaded From: http://appliedmechanicsreviews.asmedigitalcollection.asme.org/ on 12/29/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use


% Numerical integration of Mathieu equation we mix sugar into tea. By this, every region of phase space will
%for IC (0,1) overlap at some time or other with every other region.
[t,x]ode45(mathieu,[0 2*pi],[0 1],options); When we talk ofunpredictability, we are usually comfortable
x2x(end,:); with the idea that a system of many variables, such as our lives,
% Combine x1 and x2 to create Floquet Matrix shows unpredictable turns. On the other hand, when a system that
X[x1;x2];
contains only a few independent unknowns (degrees of freedom)
det(X), trace(X) displays chaotic behavior (three is the minimum number required
for chaos), it is very surprising. In turbulence, although there are
% Floquet multipliers are an infinite number of variables (the velocity, temperature, etc., at
%eigenvalues of Floquet Matrix each point in space is a variable), it seems that some aspects of
Fmeig(X); the physics can be reduced to that dominated by a much smaller
set of variables. We should, however, warn the reader of the fol-
The Floquet multipliers for the given choice of parameters are lowing: at high Reynolds numbers, it is not clear what the dimen-
k1 1:7716 and k1 0:5645. So the system is unstable. sionality of turbulence is. Even at moderate Reynolds numbers,
Exercise (ix): Obtain Floquet multipliers for g l 10 and we often need quite a high dimensionality in our models to get
 0.1 for the inverted pendulum. Determine whether the system reasonable answers. In other words, we have not clearly under-
is stable for these parameter values. stood even what we need in order to understand turbulence as the
To study Hills equation further, it is suggested to the reader to Reynolds number goes higher and higher. In spite of this, low-
go through Ref. [22]. dimensional models are instructive to study and make a good
starting point toward the understanding of more complicated prob-
lems such as turbulence, and life. The section on the Lorenz sys-
tem will describe the most famous example of low-dimensional
7 Chaotic Systems chaos, written for a toy weather problem.
When we obtained a Poincare section with nonrepeating inter- Systems often go through a series of bifurcations to go from or-
section points for the forced Duffing oscillator, we termed it a cha- dered to chaotic. The routes to chaos can most often be classified
otic system but did not say why. When do we call the dynamics of under the period-doubling route, the RuelleTakensNewhouse
a system chaotic? The most often quoted property of a chaotic quasi-periodic route, the crisis route, and the intermittency route.
system is that it is sensitive to initial conditions. By this term, We will talk only about the first here. We had seen that period
we mean that trajectories beginning from two initial conditions doubling occurs in a Craik mode of disturbance growth in the
that are arbitrarily close to each other will diverge exponentially, boundary layer, although it was not via a standard bifurcation.
so that, prescribing an initial condition to any degree of accuracy The following section gives the most famous example of this
is insufficient to predict the complete dynamics of the system. For route to chaos. The similarity in detail of this simple system to
example, let us say Eq. (2) describes chaotic dynamics that we many far more complicated real systems is simply amazing. The
wish to compute the value of x at all times and that we know the reader is encouraged to try out some of these.
initial condition x(0) correct to 10 decimal places. Even if we
have a completely accurate solver, our predictions will go very
wrong after some time t. If we know x(0) correct to 20 decimal 7.1 The Logistic Map: A Prototype for Period-Doubling
places, our predictions would be good for longer, but again at Bifurcations. Human population growth seems unlimited even
some long time we would go very wrong. Chaotic dynamics thus today. It appears to be following what Malthus described as an ex-
cannot be predicted completely. Such systems by definition will ponential growth. Malthusian law does not take into consideration
display trajectories that are not periodic, because periodic dynam- the finiteness of natural resources. Verhulst suggested a different
ics, however long the period, is completely predictable once we law where growth rate is proportional to the existing population
obtain the dynamics within one period. All chaotic systems, how- and at the same time is limited by the carrying capacity of the
ever, do contain many orbits that are periodic and have various environment. Populations of almost all species appear to follow
periods. In fact, one feature of a chaotic system is that it is dense this law. The discrete version of this law is the famous logistic
in periodic orbits. We note that these periodic orbits are unstable, map
unlike limit cycles.
It is pertinent to mention here that not all dynamical systems xn1 rxn 1  xn (30)
that are not periodic are chaotic. An Hamiltonian system of n
degrees of freedom and n integrals or constant of motion, will in This is a quadratic first-order difference equation. Population evo-
general display quasi-periodic motion on the surface of an n-torus. lution as predicted by Eq. (30) strongly depends on the value of
Such a system is termed as integrable. Quasi-periodicity is created the parameter r. For r < 2.992 trajectories evolve toward a fixed
by the existence of two or more time scales that are incommensu- point x*. The approach to the fixed point is monotonic when
rate (do not have a rational ratio). Trajectories winding around the r < 2.35 (refer to Fig. 20(a)). For r lying in the range
torus will cover its surface densely but will not return to the point 2.35 < r < 2.992, the trajectory still approaches its stable fixed
they started from. If on the other hand such a system has fewer point, but in an oscillatory manner (Fig. 20(b)). As we increase r
than n integrals, it is likely to display chaotic or nonintegrable further till it reaches the value 2.992, oscillation amplitudes
behavior. A convincing explanation of this is beyond the scope of increase and the decay rate decreases. The fixed point x*
this article, but the interested reader is directed towards any good approaches a value of 0.643 as r approaches the numerical value
text book on classical mechanics, such as Ref. [23]. 2.992 from below. As r increases steadily past the value of 2.992,
The discussion above brings to light a fundamental difference trajectories of increasing period, doubling at every bifurcation
between a system that is deterministic but chaotic, and one that is point, are displayed. A period-k orbit is a trajectory that eventually
noisy. In the former, we know the equations governing the system settles down to dynamics given by xnk xn for every n > n0 for
exactly but since the smallest changes in initial conditions take us some finite n0. It may be given by, for example,
ultimately on completely different trajectories, we cannot predict xn0 1 ; xn0 2 ; ; xn0 k . The period is 2 for r up to 3.448. Beyond
the dynamics completely. In a noisy system, we do not know the r 3.448, the dynamics changes from that on a period-2 orbit
equations of motion exactly, i.e., there are small external forces, (Fig. 20(c)) to that on a period-4 orbit (Fig. 20(d)). Again as r
usually random, which produce changes in the dynamics. Of big- crosses 3.544, the steady state changes from period-4 orbit to
gest interest to fluid dynamicists is the feature of topological mix- period-8 orbit (Fig. 20(e)). Similarly as r crosses 3.568, the steady
ing that chaotic systems possess. The effect is the same as when state changes from period-8 orbit to period-16 orbit. At r

024802-12 / Vol. 66, MARCH 2014 Transactions of the ASME

Downloaded From: http://appliedmechanicsreviews.asmedigitalcollection.asme.org/ on 12/29/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use


rar(i);
a0.1;
for j1:900
afeval(@logis,a);
end;
a(1)a;
for j2:length(r)
a(j)feval(@logis,a(j-1));
end
rpra*ones(1,length(r));
bdplotplot(rp,a,.);
end;
%font, fontsize of title and labels
set(gca,fontsize,18);
set(bdplot,MarkerSize,1)
xlabel(r), ylabel(steady state);
title(Bifurcation diagram of a Logistic map)
axis([2.6 4 0 1.1]);
Fig. 20 Trajectories of the logistic map for different values of r
8 The Lorenz System
Lorenz in his remarkable 1963 paper described a simple three-
approximately 3.56995 is the onset of chaos, and the end of the dimensional approximation of an infinite-dimensional model of a
period-doubling cascade. This cascade of period-doubling bifurca- weather system, which is now famous as the Lorenz equations. It
tions is one of the major routes from nonchaotic to chaotic behav- describes the convective motion of a 2D fluid cell that is warmed
ior. If we denote by rk the value of r for which the trajectory from below and cooled from above. The resulting partial differen-
settles down to period-2k cycle for the first time, then we have the tial equation is approximated as
Feigenbaum constant defined as x_ ry  x
rn  rn1 y_ rx  y  xz
d lim 4:66920   
n!1 rn1  rn z_ xy  bz

It is seen that period doubling bifurcations occur in increasingly where x physically describes the rate of convective overturning, y
quick succession for larger periods. This happens in all period is the horizontal temperature variation, and z the vertical tempera-
doubling bifurcations, and the Feigenbaum constant is the same ture variation. Real positive parameters r and r are proportional to
for a wide variety of systems! To generate Fig. 20(a), we run the the Prandtl number, Rayleigh number, while b is proportional to
script file bdlogis.m, which in turn calls function file logis.m. The physical dimensions of the system under consideration (Fig. 21).
function file just describes the logistic map in MATLAB syntax, Since the system is dissipative all trajectories eventually settle
while the script file runs the function file for parameter r values down to a bounded set of zero volume lying in phase space. This
from 2.6 to 4 in step of 0.004 and plots the solution after transients set may be an equilibrium point, a periodic orbit, or some compli-
die out. cated set of noninteger dimension. The last of these is called as a
The aim of the following exercise is to generate a bifurcation strange attractor. Figure 22 shows a trajectory of the system start-
diagram of a given map. ing with point (0.01,0,0) in the phase portrait. The trajectory is not
Exercise (x): Modifying the routines bdlogis.m and logis.m periodic. Continuing the numerical integration, the trajectory con-
given below appropriately, generate a bifurcation diagram for the tinues to wind around on one side and then switches from that
following cubic map: side to the other, and continues to wind around that side until the
next switch, without settling down to either periodic or stationary
behavior. This aspect is not a function of a particular choice of
xn1 rxn 1  x2n

for r varying from 0 to 3. Set the x-axis to [0.6 3] and y-axis to


[1.2 1.2] while plotting.
%Function le logis.m
function flogis(x)
global ra
fra*x*(1-x);
%Script le bdlogis.m
%set ra globally so that
it can be used across the les
global ra;
%range of the parameter r
r2.6:0.004:4;
hold off;
gure(1)
hold on
% for loop to vary parameter r
%in steps of 0.004
for i1:length(r); Fig. 21 Bifurcation diagram of a logistic map

Applied Mechanics Reviews MARCH 2014, Vol. 66 / 024802-13

Downloaded From: http://appliedmechanicsreviews.asmedigitalcollection.asme.org/ on 12/29/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use


Fig. 22 Phase portrait for r 5 10:0; b 5 8=3; r 5 28:0 with IC Fig. 23 Two trajectories for the Lorenz system with ICs
(0.01,0,0) (1.5,0,0) and (1.5 1 0.001,0,0) diverge appreciably after t 5 25

initial conditions or the choice of the numerical scheme employed


%Function le Lorenz.m
but only of the choice of parameter values. Note that a choice of
function ydotLorenz(t,y)
r 28 makes the system operate in a parameter regime far from
the original fluid dynamical problem, for which r  1. Lorenz global z
equations are chaotic for the particular choice of parameters % parameter values
shown in the phase portrait of Fig. 22. The figure is generated by sigma10;
running the script file Butterfly_script.m, which calls the function r28;
file Lorenz.m. The function file writes down Lorenz system in b8/3;
MATLABs built-in ode45 syntax and the script file generates the
% Lorenz system
phase space for Lorenz system using plot3 command. ydot[sigma*(y(2)-y(1));y(1)*(r-y(3))-y(2);
The origin is an equilibrium point for all parameter choices. y(1)*y(2)-b*y(3)];
Keeping r 10 and b 8=3; we see interesting dynamics by
changing r. For 0 < r < 1, the origin is globally stable and all tra- %Script le sensitivity_IC_script.m
jectories are attracted to it. Stability analysis is based on the line- epsil0.001;
arized flow around an equilibrium point and is determined by the
% numerical integration of Lorenz system
positive or negative real parts of the eigenvalues of the Jacobian
evaluated at this point. At r 1, a supercritical pitchfork bifurca- %for initial condition (1.5,0,0)
[t,x]ode45(Lorenz1,linspace(0,50,1000)
tionpoccurs
and
p two new equilibrium points appear, given by

6 br  1; 6 br  1; r  1 [24]. The origin loses its sta- ,[1.5 0 0]);
bility. At r 13:926    ; unstable limit cycles are created (due to %plotting the solution
a homoclinic bifurcation, which we have not discussed). At this plot(t,x(:,1));
point, a strange invariant set is born. For 1 < r < 24:74    ; the hold on;
other two equilibrium points are stable and the origin is unstable. % numerical integration of Lorenz system
At r 24:74    ; a subcritical Hopf bifurcation occurs, the two %for initial condition (1.501,0,0)
equilibrium points other than the origin lose their stability by [t,x]ode45(Lorenz1,linspace(0,50,1000)
absorbing unstable limit cycles. As we increase r further, the Lor- ,[1.5epsil 0 0]);
enz attractor is displayed and the system continues to be chaotic.
Finally as r is increased beyond 24:74    ; all three equilibrium %plotting the solution over
points are unstable, the system undergoes period doubling bifurca- the earlier solution
tions and intermittent chaos is observed. plot(t,x(:,1),r);
The dynamics on the strange invariant set shows sensitive de- %Script le Buttery_script.m
pendence on initial conditions, usually quantified by Lyapunov %numerically integrate Lorenz
exponents. Figure 23 shows the divergence of two trajectories that equations using ode45
start quite close to each other. The figure is generated by running [t,x]ode45(Lorenz1,linspace(0,50,30000)
script file sensitivity_IC_script.m, which calls function file Lor- ,[0.01 0 0]);
enz.m. The script file numerically integrates the Lorenz system
with two nearby initial conditions. Lyapunov exponents are a %3-D plot of the phase space
Butteryplot3(x(:,1),x(:,2),x(:,3));
measure of the exponential rate of divergence of two trajectories
that begin arbitrarily close to each other. A Lyapunov exponent %setting font and fontsize of label and title
greater than zero signifies the system to be chaotic. set(gca,fontsize,18); set(Buttery,
The aim of this exercise is to generate phase portraits of the MarkerSize,5);
Lorenz system for various parameter values. xlabel(x(t)), ylabel(y(t)),
Exercise (xi): Generate phase portraits of the Lorenz equations zlabel(z(t));
for the parameter values r 10:0; b 8=3 and for different r title(Phase portrait for Lorenz equations)
values such as r 0:5; 2; 15 using Lorenz.m for initial conditions axis([10,20,20,20,0,50]);
such as (0.01,0,0). grid on;

024802-14 / Vol. 66, MARCH 2014 Transactions of the ASME

Downloaded From: http://appliedmechanicsreviews.asmedigitalcollection.asme.org/ on 12/29/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use


s 0ek1 s ;
s k2 s
2s e ; ;
jjsjj
n  1s kn s
ns e
jjn  1sjj

We define k as

1X
n
1X
n
Fig. 24 Numerical calculation of the largest Lyapunov k lim lnjjksjj lim kk
n!1 ns n!1 n
exponent k1 k1

Figure 25 shows all the three Lyapunov exponents for the Lorenz
Consider two trajectories that start from nearby initial condi- system. The largest is greater than zero and, therefore, the system
tions. Let t be the distance between them. We write it in the lin- is chaotic. The system displays a strange attractor for the chosen
earized form values of the parameters, in accordance with the largest exponent
being positive, the second one zero and the smallest negative. The
_ Jt (31) magnitude of the largest negative exponent is greater than that of
the largest positive one, given that the Lorenz system is dissipa-
with initial condition 0 assumed to be a very small quantity. tive. To learn more about the Lorenz system and related chaos
The Jacobian for the Lorenz system is given by theory, it is suggested to the reader to study Ref. [24].
2 3
r r 0
6 7 9 Conclusions
Jt 4 r  zt 1 xt 5
Nonlinear dynamics is too wide a subject to be treated fairly in
yt xt b a brief review. Our objective has only been to ignite the readers
enthusiasm for this field. Apart from the references we recom-
We solve Eq. (31) to get mend above at the end of some of the subsections, there is an
enormous amount of literature on this topic available in Wikipe-
dia, Scholarpedia, and in many other cites, journals, and books.
t eJt 0 The difference in the present treatment is that we have tried to
make connections with flow stability and transition to turbulence.
For further reading on this vast subject, we recommend for gen-
To get the Lyapunov exponent numerically, we integrate Eq. (31)
eral reading excellent books [2628] and for advanced treatment
for a large number of cycles (n), each for small but fixed time
on a few topics touched upon in this document, we recommend
interval s. We begin with the IC 0 as shown in Fig. 24. At the
Ref. [30].
end of the first cycle, we store the natural logarithm of the Euclid-
ean norm (jj  jj) of the end condition s (Euclidean distance
between two given points a and b in terms of Euclidean norm is
Acknowledgment
given by jja  bjj). We normalize s to unit norm (refer to
Fig. 24) to use this normalized end condition as the IC for the next Sharath Jose and Mamta Jotkar are thanked for their contribu-
cycle to integrate Eq. (31). We repeat this process n times. An tions. The anonymous referees have contributed significantly to
arithmetic average of these stored logarithms of end conditions in improving the article.
the limit of number of cycles n tending to infinity is nothing but
the largest Lyapunov exponent k [25].
References
[1] Strogatz, S., 2001, Nonlinear Dynamics and Chaos: with Applications to Physics,
Biology, Chemistry and Engineering, 1st ed., Westview Press Boulder, CO.
[2] Schneider, T., Marinc, D., and Eckhardt, B., 2010, Localized Edge States Nu-
cleate Turbulence in Extended Plane Couette Cells, J. Fluid Mech., 646, pp.
441451.
[3] Willis, A., and Kerswell, R., 2009, Turbulent Dynamics of Pipe Flow Captured
in a Reduced Model: pu? Relaminarization and Localized Edge States,J. Fluid
Mech., 619, pp. 213233.
[4] Duguet, Y., Schlatter, P., and Henningson, D. S., 2009, Localised Edge States
in Plane Couette Flow, Phys. Fluids, 21, p. 111701.
[5] Manneville, P., and Pomeau, Y., 2009, Transition to Turbulence, Scholarpe-
dia, 4(3), p. 2072.
[6] Subramanian, P., Mariappan, S., Sujith, R. I., and Wahi, P., 2010, Bifurcation
Analysis of Thermoacoustic Instability in a Horizontal Rijke Tube, Int. J.
Spray Combus. Dyn., 2, pp. 325355.
[7] Gaster, M., 1969, Vortex Shedding From Slender Cones at Low Reynolds
Numbers, J. Fluid Mech., 38(3), pp. 565576.
[8] Venkatraman, D., 2013, Computational Models of Dorsal Coverts on Birds
Wings, Ph.D. thesis, University of Genova, Italy.
[9] Rand, R., Lecture notes on Nonlinear Vibrations. Version 52. Available at:
http://www.tam.cornell.edu/randdocs/ed
[10] Hinch, E. J., 1991, Perturbation Methods, Cambridge University Press, Cam-
bridge, UK.
[11] Johnson, R. S., 2005, Singular Perturbation Theory, Springer ebooks.
[12] Holmes, M. H., 1991, In Introduction to Perturbation Methods, Springer-Verlag,
New York.
Fig. 25 Three Lyapunov exponents of the Lorenz system for [13] Kevorkian, J., and Cole, J. D., 1996, Multiple Scale and Singular Perturbation
r 5 10:0; b 5 8=3; r 5 28:0 Methods, Springer-Verlag, New York.

Applied Mechanics Reviews MARCH 2014, Vol. 66 / 024802-15

Downloaded From: http://appliedmechanicsreviews.asmedigitalcollection.asme.org/ on 12/29/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use


[14] Mickens, R. E., 1981, An Introduction to Nonlinear Oscillations, 1st ed., Cam- [23] Rana, N., and Joag, P., 2011, Classical Mechanics, 30th ed., Tata-McGraw
bridge University Press Cambridge, UK. Hill.
[15] Mickens, R. E., 1996, Oscillations in Planar Dynamic Systems, World Scien- [24] Sparrow, C., 1982, The Lorenz Equations: Bifurcations, Chaos and
tific Publishing Co., River Edge, New Jersey. Strange Attractors, Applied Mathematical Sciences 41, Springer-Verlag,
[16] Chatterjee, A., 2002, An Elementary Continuation Technique, http://home.iitk. New York.
ac.in/anindya/continuation.pdf [25] Rugonyi, S., and Bathe, K., 2003, An Evaluation of the Lyapunov Characteris-
[17] Morgan, A., 2009, Solving Polynomial Systems using Continuation for Engi- tic Exponent of Chaotic Continuous Systems, Int. J. Numer. Methods Eng., 56,
neering and Scientific Problems, Classics in Applied Mathematics, SIAM. pp. 145163.
[18] Stoker, J. J., 1950, Nonlinear Vibrations in Mechanical and Electrical Systems, [26] Wiggins, S., 2003, Introduction to Applied Nonlinear Dynamics and Chaos,
John Wiley & Sons, New York. 2nd ed., Springer, New York.
[19] Verhulst, F., 1996, Nonlinear Differential Equations and Dynamical Systems, [27] Hale, J. K., 1963, Oscillations in Nonlinear Systems, 1st ed., Dover Publications
2nd ed., Springer, New York. Mineola, NY.
[20] Lanczos, C., 1986, The Variational Principles of Mechanics, Dover Publica- [28] Drazin, P. G., 1992, Nonlinear Systems, 1st ed., Cambridge University Press,
tions, New York. Cambridge, UK.
[21] Hill, G. W., 1886, On the Part of the Motion of the Lunar Perigee is a Function [29] Grimshaw, R., 1991, Nonlinear Ordinary Differential Equations, Taylor &
of the Mean Motions of the Sun and the Moon, Acta Math., 8, pp. 136. Francis, London.
[22] Magnus, W., and Winkler, S., 2004, Hills equation, Dover Publications, Mine- [30] Glendinning, P., 1994, Stability, Instability and Chaos, Cambridge University
ola, NY. Press, Cambridge, UK.

024802-16 / Vol. 66, MARCH 2014 Transactions of the ASME

Downloaded From: http://appliedmechanicsreviews.asmedigitalcollection.asme.org/ on 12/29/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use

Você também pode gostar