Home / Mechanics / Double Pendulum: Chaos

Double Pendulum: Chaotic Motion

JEE Advanced
Total energy: 0 J (per kg) Angle divergence: 0°

The physics

A double pendulum — just a pendulum swinging from the end of another pendulum — is one of the simplest systems that produces genuinely chaotic motion. Unlike a single pendulum, there's no simple closed-form solution; the motion must be computed numerically, step by step.

Coupled nonlinear equations of motion (Lagrangian mechanics), solved via RK4 integration each frame
Total mechanical energy is exactly conserved (undamped) — verified numerically to hold across the entire simulated motion
"Chaos" here means: fully deterministic equations, yet outcomes become practically unpredictable due to extreme sensitivity to initial conditions

This simulation solves the real coupled differential equations at every frame — the motion you see is a genuine numerical solution, not a pre-recorded animation, which is exactly why two nearly-identical starting angles eventually produce completely different trajectories.

Exam tip: JEE Advanced occasionally probes double pendulum setups conceptually (small-oscillation coupled-mode analysis) rather than asking for the full chaotic solution — the small-angle limit gives two distinct "normal mode" frequencies, a very different (and solvable) regime from the large-angle chaos shown here.

Sample data: how a tiny difference grows into a completely different motion

Two identical double pendulums start with the upper arm at 120° and the lower arm at −10°. In the second one the upper angle is larger by only one millionth of a radian. The table shows the difference in the upper angle.

TimeDifference in upper angle
0 s1.00×10⁻⁶ rad
4 s8.16×10⁻⁶ rad
8 s3.19×10⁻⁴ rad
12 s0.03 rad
15 s0.92 rad
20 s1.07 rad

Conditions: equal masses 1 kg, equal arm lengths 1 m, g = 9.8 m/s², released from rest, RK4 integration with a 0.5 ms step.

Notice: a difference of 10⁻⁶ rad stays invisible for several seconds, then grows to 0.03 rad by 12 s (30,000 times larger) and reaches about 1 rad (57°) by 15 s, so the two motions have become unrelated. The equations are exactly deterministic and the total energy stayed constant to within 10⁻⁸ (relative) in this calculation, so the unpredictability comes from the sensitivity, not from any loss or noise.

🎯 Practice Zone

Puzzle 1. 🎲 The double pendulum's motion is governed by completely deterministic equations — no randomness anywhere. So why is its long-term motion still fundamentally unpredictable?

This is the defining feature of chaos: determinism doesn't guarantee predictability. Any real-world measurement of initial conditions has some tiny uncertainty, and in a chaotic system, that uncertainty grows exponentially with time (not linearly). Within seconds, the "error bar" on your prediction becomes as large as the motion itself — the system is perfectly deterministic in principle, but practically unpredictable beyond a short horizon, because you can never measure the starting angle with infinite precision.

Puzzle 2. ⚖️ At small oscillation angles, the double pendulum's motion is NOT chaotic — it settles into a predictable combination of two "normal modes." Why does chaos disappear at small angles?

The chaotic behavior comes from the genuinely nonlinear terms in the equations of motion (sin θ, products of angular velocities, etc.). At small angles, sin θ≈θ (the small-angle approximation), which linearizes the equations. Linear systems, no matter how many coupled parts they have, never produce chaos — they only ever combine a fixed set of "normal mode" oscillations. Chaos in the double pendulum is a genuinely large-angle, nonlinear phenomenon.

Puzzle 3. 🔋 Even in the middle of wildly chaotic motion, this simulation's total energy stays essentially perfectly constant (verified to about 8 decimal places). Why does chaos not violate energy conservation?

Chaos is about unpredictability in the detailed trajectory (exact angles at exact times), not about breaking any conservation law. The system is still governed by ordinary Newtonian mechanics with gravity as the only force (no friction in this idealized model) — energy conservation is a completely separate, always-true statement, regardless of how complicated or unpredictable the specific path through phase space happens to be.

Puzzle 4. 🌪️ Does the double pendulum's chaotic behavior depend on the two arm lengths and masses being exactly equal? What happens if they're very different (say, a tiny second bob on a huge first arm)?

Chaos appears across a very wide range of mass and length ratios, not just the equal-parameter case — the nonlinear coupling between the two arms exists regardless of the specific values. However, in extreme limits (like a vanishingly small second mass), the system approaches the behavior of an ordinary single pendulum with a tiny, nearly-decoupled second bob riding along — chaos becomes practically undetectable in that limit, even though it's still technically present.

Puzzle 5. 🌀 If you started two double pendulums from angles differing by an unmeasurably tiny amount (say, 10⁻¹⁰ degrees), would they still eventually diverge completely?

Yes — this is exactly what makes chaos so extreme. The divergence grows exponentially, so even an astronomically tiny initial difference gets amplified over time until the two pendulums are moving completely independently. The only effect of starting closer together is a longer delay before visible divergence appears — not a permanently smaller difference. This exponential sensitivity is precisely why long-term weather forecasting faces a fundamental limit, not just a technological one.

Problem 1 (NEET level). 📘 For a single pendulum of length 1 m, find its angular frequency ω₀=√(g/L).

ω₀ = √(g/L) = √(9.8/1) = 3.130 rad/s (period T=2.007s — this is the baseline for comparison with the double pendulum's normal modes below)

Problem 2 (JEE Main level). 📗 For a small-angle double pendulum (equal masses, equal lengths L=1m), the "in-phase" normal mode has ω₋=√(g/L)·√(2−√2). Find this frequency and its period.

ω₋ = √(9.8/1) × √(2−1.414) = 3.130 × 0.765
ω₋ = 2.396 rad/s (T = 2π/ω₋ = 2.622 s) — verified numerically against a genuine small-angle simulation, matching to within 0.3%

Problem 3 (WBJEE level). 📙 Find the "out-of-phase" normal mode frequency ω₊=√(g/L)·√(2+√2) for the same double pendulum.

ω₊ = √(9.8/1) × √(2+1.414) = 3.130 × 1.848
ω₊ = 5.784 rad/s (T = 1.086 s) — notice this is more than twice as fast as the in-phase mode, and faster even than a single pendulum of the same length

Problem 4 (CBSE Class 11 level). 📕 Both bobs (m₁=m₂=1kg, L₁=L₂=1m) are released from rest with both arms horizontal (θ₁=θ₂=90°). Taking the pivot as the zero of potential energy, find the total mechanical energy.

At release: both bobs are at the pivot's height (y₁=y₂=0 relative to pivot), and at rest (KE=0)
Total energy E = KE + PE = 0 + 0 = 0 J (using the pivot as the reference level)

Problem 5 (JEE Advanced style). 📓 Using the energy from Problem 4, find the upper bound on total kinetic energy when both bobs swing down to their lowest possible point (hanging straight down).

PE at lowest point = −m₁gL₁ − m₂g(L₁+L₂) = −(1)(9.8)(1) − (1)(9.8)(2) = −29.4 J
Since total E=0 J (conserved): KE_max ≤ 29.4 J
This is an upper bound, not necessarily achieved exactly — in the real chaotic motion, both bobs generally don't pass through their lowest points simultaneously, so this represents the theoretical ceiling set by energy conservation alone.
🌀
1890s

Henri Poincaré, studying the three-body problem in celestial mechanics, discovers that even simple deterministic systems can show extreme sensitivity to initial conditions — laying the mathematical groundwork for what would later be called chaos theory, decades before the double pendulum itself was studied this way.

🌦️
1961

Edward Lorenz, running a simplified weather simulation, accidentally discovers modern chaos theory when a tiny rounding difference in his re-entered initial conditions produces a completely different forecast — coining the intuition behind the famous "butterfly effect."

💻
1970s–1980s

Chaos theory matures into a full mathematical field, with the double pendulum becoming one of its most popular teaching examples — simple enough to build physically on a tabletop, yet rich enough to display genuine, visible chaotic motion.

🖥️
1990s–present

Widespread availability of numerical computing (like the RK4 integration used in this simulation) makes it possible for students anywhere to explore double pendulum chaos interactively — a system that would have required serious analog computing effort to simulate as recently as the 1970s.