DOING PHYSICS WITH MATLAB

DYNAMICS OF OSCILLATING AND CHAOTIC SYSTEMS

POINCARE SECTIONS: DUFFING OSCILLATOR

 

Ian Cooper

matlabvisualphysics@gmail.com

 

DOWNLOAD DIRECTORY FOR MATLAB SCRIPTS

 

chaos02.m

Runga-Kutta method for solving the equation of motion for a Duffing oscillating with viscous damping and forced motions. Computation for the Poincare section for a phase space plot. All parameters are changed within the script.

 

chaos01.m

Runga-Kutta method for solving the equation of motion for a Duffing oscillating: free, viscous damping and forced motions: time; displacement; and phase space plots. All parameters are changed within the script. Animations of the trajectories are saved as animated gif files. The script could be altered so the animations are saved as avi files.

 

 

 

 

 

POINCARE SECTIONS

There is a way to study chaotic motion that is better than watching a trajectory wander around in phase space. For a system which includes the viscous damping and an externally applied driving force, the Poincare section is constructed on the phase plot (x vs v graph) by only plotting points every   where  is the angular frequency of the driving force and  is its period.

 

If the orbit in phase space is periodic, with this period, then we will get only one point displayed on the plot. If the orbit has a period equal to two times the period of the driving force, then the Poincare section will show two points, and so on. If the system is chaotic, the Poincare section will consist of a pattern of points called the attractor. The attractor has a structure that is often beautiful. A surprising result is that a deterministic system can exhibit unpredictability and apparent chaos and at the same time preserve a coherent global structure.

 

 

Simulations #1     Poincare sections          chaos02.m

A useful way of analysing chaotic motion is to look at what is called the Poincare section. Rather than considering the phase space trajectory for all times, which gives a continuous curve, the Poincare section is just the discrete set of phase space points of the particle at every period of the driving force, i.e. at 

 

Input parameters

     c(1) = 0.1     c(2) = 1   c(3) = -1   c(4) = 0.38   c(5) = 1.4

    m = 1.0     nT = 501   nP = 24000   nS = 1

    x(1) = 0     v(1) = 0

 

nT is the number of calculations before another point is plotted.

nP is the number of points plotted at time intervals of .

nS is the start number for plotting the points.

 

For figure 1, nP =24000, which is a very large number. nP must be large enough to plot enough points to show the structure of the Poincare section. It took about 4 minutes to calculate and plot the Poincare section.

Fig. 1.   Poincare section of the Duffing two-well oscillator. This strange diagram is the strange attractor. It is the limiting set of points to which the trajectory tends to after every period of the driving force (after the initial transient). Notice the that the structure is complicated but not completely random, we see structure.

 

 

The Fractal Nature of Chaotic Attractors

 

The structure of a Poincare section can be viewed in more detail by using the Zoom In button in the Figure Window. When you Zoom In it will take some time for the screen to update because of the large number of points that are plotted. Zoom In on a small region of the strange attractor and examine its structure. Zoom In again and view the structure of the strange attractor. You will notice that the features of the strange attractor at the smaller scale are similar to those features at the larger scale – there appears to be a fine structure. Having the same features appearing in different parts of a figure and at different scales is a characteristic feature of a fractal. An attractor is fractal if a set of points that shows self-similarity as the scale being used decreases. In practice, you cannot keep zooming in, because fewer and fewer points will be displayed and if you keep increasing the number of points, the computation time becomes very large.

 

Fig. 2.  Expanded views of the Poincare section displayed in figure 1 for the two-well Duffing oscillator.

 

 

Simulations #2     Poincare sections          chaos02.m

 

Input parameters

     c(1) = 0.24     c(2) = 1   c(3) = -1   c(4) = 0.88   c(5) = 1.7

    m = 1.0     nT = 501   nP = 24000

    x(1) = 1     v(1) = -1

Fig. 3.  Poincare section of the Duffing two-well oscillator and an expanded view of the lower left portion of the attractor. Note the structure that emerges.

 

 

Simulations #3     Poincare sections – period doubling: the road to chaos         chaos02.m

 

Input parameters

     c(1) = 0.10     c(2) = 1   c(3) = -1   c(4) = 0.1   c(5) = 1.4

    m = 1.0     nT = 501   nP = 4000    nS = 3000

    x(1) = 0     v(1) = 0

Fig. 4.   The period of the oscillator is T. The Duffing oscillator oscillates at the driving frequency. Hence, the driving period is equal to the period of the oscillation. Only one point is displayed in the Poincare section.

 

View chaos01, simulation #4.1 Fig. 6.

 

 

Input parameters

     c(1) = 0.10     c(2) = 1   c(3) = -1   c(4) = 0.32   c(5) = 1.4

    m = 1.0     nT = 501   nP = 4000    nS = 3000

    x(1) = 0     v(1) = 0

Fig. 5.   The period T of the Duffing oscillator is now twice the period of the driving force . Hence, two points are displayed in the Poincare section.

 

View chaos01, simulation #4.2 Fig. 7.

 

 

Input parameters

     c(1) = 0.10     c(2) = 1   c(3) = -1   c(4) = 0.34   c(5) = 1.4

    m = 1.0     nT = 501   nP = 4000    nS = 3000

    x(1) = 0     v(1) = 0

Fig. 6.   Expanded view. The period T of the Duffing oscillator is now four times the period of the driving force . Hence, four points are displayed in the Poincare section.

 

View chaos01, simulation #4.3 Fig. 8.

 

 

Input parameters

     c(1) = 0.10     c(2) = 1   c(3) = -1   c(4) = 0.3425   c(5) = 1.4

    m = 1.0     nT = 501   nP = 4000    nS = 3000

    x(1) = 0     v(1) = 0

 

Fig. 6.   Expanded view. The period T of the Duffing oscillator is now eight times the period of the driving force . Hence, four points are displayed in the Poincare section.

         

 

Input parameters

     c(1) = 0.10     c(2) = 1   c(3) = -1   c(4) = 0.35   c(5) = 1.4

    m = 1.0     nT = 501   nP = 4000    nS = 3000

    x(1) = 0     v(1) = 0

Fig. 5.   The motion is no longer periodic – the motion has become chaotic.

 

Note: the slight change in the strength of the diving force given by the variable c(4) results in a transition from periodic motion to chaotic motion.

 

    

       

         

    

           

 

 

 

 

Logistic maps can produce fractal patterns. Integrating a differential equation requires much more time than iterating a logistic map. Therefore, people have investigated maps which have similar behaviour to that of driven, damped differential equations like the Duffing equation.

 

Thus, we can come to a quite an exciting conclusion – the beautiful world of fractals tentatively touches the fascinating field of chaotic dynamics.