Python4Physics LAB
Classic v1
Home > Visualization > Classical Mechanics: Double Pendulum Dynamics & Chaos

Module 3.2 Deterministic Chaos & The Butterfly Effect: Sensitivity to Initial Conditions

Modules Index
Simulation 1

Deterministic Chaos & The Butterfly Effect: Sensitivity to Initial Conditions

Python 3.11 with Matplotlib Slider Ready

Interactive Physics Ready

Click "Run Code" to launch simulation and interactive parameter sliders.


                                

                            


                            
                                                        

                                

                                                                            

                                            
 Mathematical Problem Formulation

                                            
Deterministic Chaos, Butterfly Effect & Lyapunov Divergence in Dual Pendulums

                                        

                                    
                                                                            

                                            
 Theoretical Background & Explanation

                                            

    

        

             Deterministic Chaos & The Butterfly Effect
        

        

            The concept of deterministic chaos implies that a physical system governed by completely exact, deterministic equations of motion without any stochastic randomness can nonetheless produce fundamentally unpredictable long-term trajectories due to exponential sensitivity to initial conditions.
        

    


    
    

        Deterministic Chaos and Lyapunov Divergence in Double Pendulums
        

            Figure 3.2.1: Demonstration of chaotic divergence: two identical double pendulums released with an initial discrepancy $\Delta\theta_0 = 10^{-3}$ rad stay aligned during the initial predictability window before exponentially flying apart ($\Delta r(t) \sim e^{\lambda t}$).
        

    


    

        1. The Maximal Lyapunov Exponent ($\lambda$)
    

    

        Consider two initial states in 4D phase space $\vec{X}_A(0)$ and $\vec{X}_B(0)$ separated by an infinitesimal Euclidean perturbation $\|\Delta \vec{X}(0)\| = \delta_0 \ll 1$. In a chaotic dynamical system, the separation distance evolves exponentially in time:
    

    

        $$\|\Delta \vec{X}(t)\| \approx \delta_0 \, e^{\lambda t}$$
    

    

        The maximal Lyapunov exponent $\lambda$ is formally defined by the asymptotic limit:
    

    

        $$\lambda = \lim_{t \to \infty} \lim_{\delta_0 \to 0} \frac{1}{t} \ln \left( \frac{\|\Delta \vec{X}(t)\|}{\delta_0} \right)$$
    

    

        • $\lambda < 0$: Regular dissipative system converging to a stable fixed-point attractor.

        • $\lambda = 0$: Conservative, integrable periodic/quasi-periodic orbits (e.g. harmonic oscillators or planetary orbits without resonances).

        • $\lambda > 0$: The system is chaotic. Adjacent trajectories diverge exponentially, destroying deterministic long-range forecastability!
    


    

        2. The Predictability Horizon & Geometric Saturation
    

    

        Because physical pendulums are mechanically bounded to a maximum possible distance $D_{\text{max}} = 2(L_1 + L_2)$, the exponential growth $\Delta r(t) \sim e^{\lambda t}$ cannot continue infinitely. It reaches a saturation boundary when the two bobs become as separated as physically possible in space:
    

    

        $$t_{\text{horizon}} \approx \frac{1}{\lambda} \ln\left( \frac{D_{\text{max}}}{\Delta r_0} \right)$$
    

    

        Notice the logarithmic dependence: even if you improve your measurement precision by a factor of $1,000$ ($\Delta r_0 \to 10^{-3} \Delta r_0$), your forecast horizon $t_{\text{horizon}}$ only increases additively by $\frac{\ln(1000)}{\lambda} \approx \frac{6.9}{\lambda}$ seconds! This mathematical fact is why weather forecasting beyond two weeks is fundamentally impossible, as discovered by Edward Lorenz in 1963.
    


    

        3. Energy Threshold: Transition from Integrability to Chaos
    

    

        The double pendulum does not exhibit chaos at all energies:
        
• Low Energy Regime ($E < m_2 g L_1$): The potential well confines both angles to small amplitudes. KAM (Kolmogorov-Arnold-Moser) tori dominate phase space, and trajectories remain regular and quasi-periodic.
        
• Critical Transition ($E \approx m_2 g L_1$): The lower pendulum has sufficient energy to reach the inverted saddle point $\theta_2 = \pi$. Separatrix crossing creates homoclinic chaos.
        
• Fully Chaotic Regime ($E \gg m_2 g L_1$): KAM tori are shattered into a sea of chaos, and the phase space portrait exhibits ergodicity and dense orbital mixing.