Beyond the Parabola Slides Beyond the Parabola
Lesson 1: The Anharmonic Oscillator
NONLINEAR DYNAMICS UNIT
The Linear Assumption
The Simple Harmonic Motion (SHM) model assumes a perfectly quadratic potential well:
\[ V(x) = \frac{1}{2} k x^2 \]
Restoring force is purely linear: \( F = -kx \)
Frequency is independent of amplitude.
[Visual: Parabolic Potential Well]
Deviations from Ideality
Taylor Expansion of a General Potential \( V(x) \)
\[ V(x) = V(0) + V'(0)x + \frac{1}{2}V''(0)x^2 + \frac{1}{6}V'''(0)x^3 + \dots \]
Quadratic
The SHM term. Leads to linear restoration.
Cubic (\( x^3 \))
Breaks symmetry. Creates asymmetric potentials.
Quartic (\( x^4 \))
Modifies the "stiffness" at high amplitudes (Soft/Hard springs).
Finding the Period Exactly
Conservation of energy allows us to find the period \( T \) for any conservative potential \( V(x) \) with energy \( E \).
\[ T(A) = 2 \int_{x_{min}}^{x_{max}} \frac{dx}{\sqrt{\frac{2}{m}(E - V(x))}} \]
1
For \( V(x) \propto x^2 \), the integral yields a constant period \( 2\pi\sqrt{m/k} \).
2
For nonlinear potentials, \( T \) depends on the amplitude \( A \).
The Pendulum "Violation"
The exact potential for a pendulum is: \[ V(\theta) = mgL(1 - \cos\theta) \]
Small angle approximation: \( \cos\theta \approx 1 - \frac{\theta^2}{2} \)
What happens when \( \theta \) is large?
Period Expansion
\[ T = T_0 \left( 1 + \frac{1}{16}\theta_0^2 + \dots \right) \]
The clock runs slower as the amplitude increases. This is our first encounter with frequency shifting .
Potential Pitfalls Worksheet Potential Pitfalls
Nonlinear Dynamics & Perturbation Methods
Student:
Date:
Problem Statement
Consider a particle of mass \( m \) moving in a quartic potential well defined by \( V(x) = \frac{1}{2} k x^2 + \frac{1}{4} \epsilon x^4 \), where \( \epsilon > 0 \) is a small parameter representing the nonlinearity. In this workshop, you will derive the period of oscillation using energy conservation.
1
The Energy Integral Setup
Using the conservation of energy \( E = \frac{1}{2}m\dot{x}^2 + V(x) \), express the velocity \( \dot{x} \) as a function of position and total energy. Then, write the integral expression for the period \( T \) for an oscillation with amplitude \( A \). Assume the motion is symmetric about \( x=0 \).
2
Small Perturbation Approximation
Assume \( \epsilon \) is very small. Perform a Taylor expansion of the integrand in your period expression from Problem 1. Use the approximation \( (1 + z)^{-1/2} \approx 1 - \frac{1}{2}z \) for small \( z \).
3
Evaluating the Shift
Evaluate the resulting integrals to find the first-order correction to the period \( T(A) \). Recall that for the harmonic case (\( \epsilon = 0 \)), \( T_0 = 2\pi\sqrt{m/k} \). Show that the period decreases as amplitude increases for a "hard" spring (\( \epsilon > 0 \)).
Critical Thinking
Physical Interpretation: If the potential was cubic (\( V(x) = \frac{1}{2}kx^2 + \frac{1}{3}\alpha x^3 \)), how would the symmetry of the oscillation be affected? Would the period shift differently than in the quartic case?
Anharmonic Facilitation Guide Instructor Resource
Anharmonic Facilitation Guide
Lesson 1: The Anharmonic Oscillator and Potential Wells
Pacing
90 min
Learning Objectives
Contrast linear vs. nonlinear restoring forces in terms of potential expansion.
Calculate period shifts for small nonlinearities using energy integrals.
Explain the physical phenomenon of frequency dependence on amplitude.
Key Terms
Potential Well Anharmonicity Hard Spring Energy Integral
1
The Hook: The Pendulum Problem (15 min)
Start by asking: "In undergraduate physics, we say the period of a pendulum is \( T = 2\pi\sqrt{L/g} \). If I swing the pendulum at a 60-degree angle, is this still true?"
Key Insight: The formula is an approximation based on \( \sin\theta \approx \theta \). For graduate level, we need to handle the "residual" forces that this approximation ignores.
2
Technical Lecture: Taylor Expansion (25 min)
Use the slides to show how any potential can be expanded. Emphasize that \( V'(0) = 0 \) for an equilibrium point and \( V''(0) \) determines the harmonic frequency \( \omega_0 \).
Common Misconception:
Students often think the cubic term (\( x^3 \)) doesn't affect the period. While it doesn't at first order in symmetric oscillations, it shifts the equilibrium point, which is a subtle but important effect.
3
Workshop: Potential Pitfalls (40 min)
Hand out the "Potential Pitfalls" worksheet. Students work in pairs to derive the period of a quartic oscillator.
Answer Key Highlight
\[ T \approx T_0 \left( 1 - \frac{3\epsilon A^2}{8k} \right) \] Note: For \( \epsilon > 0 \) (Hard Spring), the period decreases (frequency increases) as amplitude increases.
4
Debrief: Physical Realism (10 min)
Discuss real-world "hard" and "soft" springs. Example: A guitar string increases in tension as it is plucked harder, acting as a hard spring.
The Secular Crisis Slides The Secular Crisis
Lesson 2: Perturbation Theory Failures
NONLINEAR DYNAMICS UNIT
Regular Perturbation Theory
For an equation like the Duffing Equation:
\[ \ddot{x} + x + \epsilon x^3 = 0 \]
We assume a power series solution in \( \epsilon \):
\[ x(t) = x_0(t) + \epsilon x_1(t) + \epsilon^2 x_2(t) + \dots \]
The Procedure
Substitute expansion into the ODE.
Group terms by powers of \( \epsilon \).
Solve the resulting sequence of linear ODEs.
This works beautifully for steady-state algebraic problems. For dynamics , we run into a wall.
The Birth of Secular Terms
Analyzing the First-Order Correction \( x_1(t) \)
At order \( \epsilon^1 \), we often see:
\[ \ddot{x}_1 + x_1 = -x_0^3 \]
Substituting \( x_0 = \cos t \):
\[ \ddot{x}_1 + x_1 = -\frac{3}{4}\cos t - \frac{1}{4}\cos 3t \]
The Resonance Problem
The term \( \cos t \) is a driving force at the natural frequency of the left-hand side.
Solution: \( x_1(t) \sim t \sin t \)
As \( t \to \infty \), the solution grows without bound. This is a Secular Term .
Why This Fails Physically
!
A real Duffing oscillator with \( \epsilon > 0 \) is bounded. Energy is conserved; it cannot go to infinity.
?
The secular term \( t \sin t \) suggests a breakdown in our time scales .
"Regular perturbation theory assumes the nonlinearity only affects the shape, not the speed. But we know nonlinearity shifts the frequency."
[Visual: Comparison of true bounded oscillation vs. divergent secular approximation]
The Solution?
We need a method that accounts for the accumulation of small errors over long times.
Poincaré-Lindstedt
Expand the frequency \( \omega \) as well as the solution \( x \).
Multiple Scales
Introduce slow and fast time variables (\( T_0, T_1, \dots \)).
Next Lesson: Mastering the Scales
Perturbation Puzzles Worksheet Perturbation Puzzles
Regular Perturbation & The Secular Problem
Researcher:
Project ID:
"In physics, we often find that a small correction leads to a large catastrophe. Here, you will witness the mathematical breakdown of regular perturbation theory when applied to oscillatory systems."
The Asymmetric Oscillator
Consider the equation \( \ddot{x} + x + \epsilon x^2 = 0 \), where \( \epsilon \ll 1 \). This represents an oscillator with a small quadratic nonlinearity.
(a) Expansion Setup
Assume \( x(t) = x_0(t) + \epsilon x_1(t) + O(\epsilon^2) \). Substitute this into the ODE and derive the equations for \( x_0(t) \) and \( x_1(t) \).
(b) Solving for the Correction
Using the initial conditions \( x(0) = A, \dot{x}(0) = 0 \), solve for \( x_0(t) \) and then find the particular solution for \( x_1(t) \). Identify any secular terms if they exist.
The Failure Analysis
Repeat the procedure for the Duffing Equation : \( \ddot{x} + x + \epsilon x^3 = 0 \).
(a) Identifying Resonance
Show that the driving term for the \( x_1 \) equation contains a component \( \cos(t) \). Why does this term specifically cause a secular divergence?
(b) Validity Limits
If the secular term is of the form \( \frac{3A^3}{8}t \sin t \), at approximately what time \( t \) does the first-order correction become as large as the zeroth-order solution? Express this in terms of \( \epsilon \) and \( A \).
Theoretical Reflection
Mathematically, we know the true solution to the Duffing equation is periodic and bounded (it can be written in terms of Jacobi Elliptic functions). Why does a power series in \( \epsilon \) fundamentally fail to capture this periodic nature over long times?
Secular Solutions Key Answer Key & Teacher Notes
Secular Solutions
Lesson 2: Perturbation Theory & Secular Terms
Topic
Duffing Failures
Problem 1 Solutions
1(a) Expansion Equations
\( O(\epsilon^0): \ddot{x}_0 + x_0 = 0 \)
\( O(\epsilon^1): \ddot{x}_1 + x_1 = -x_0^2 \)
1(b) Result & Secular Check
\( x_0 = A \cos t \). Substituting into \( x_1 \):
Note: There are NO secular terms here because 1 and \(\cos 2t\) do not resonate with the LHS frequency (1). The solution remains bounded at \( O(\epsilon) \).
Problem 2 Solutions
2(a) The Duffing Secular Term
\( \ddot{x}_1 + x_1 = -x_0^3 = -A^3 \cos^3 t = -\frac{3A^3}{4}\cos t - \frac{A^3}{4}\cos 3t \).
The term \( \frac{3A^3}{4}\cos t \) resonates.
Particular solution: \( x_{1,p} = -\frac{3A^3}{8} t \sin t \).
2(b) Breakdown Scale
Breakdown occurs when \( \epsilon x_1 \approx x_0 \).
Pedagogical Strategy
The "Aha!" Moment
Ensure students realize that the quadratic nonlinearity (Problem 1) shifts the center of oscillation but doesn't cause divergence at first order, whereas the cubic nonlinearity (Problem 2) immediately breaks the method. This contrast is vital.
Transitioning to Lesson 3
Use the "Validity Limits" result to motivate Multiple Scales. Tell them: "We need a method that can handle the time scale \( t \sim 1/\epsilon \)."
Time Speed Slides Time Speed Splitting
Lesson 3: The Method of Multiple Scales
NONLINEAR DYNAMICS UNIT
The Core Concept: Hierarchical Time
We treat time as a collection of independent variables representing different scales of motion :
T₀
Fast Time: \( T_0 = t \)
Governs the rapid oscillations.
T₁
Slow Time: \( T_1 = \epsilon t \)
Governs slow amplitude/phase drift.
The Derivative Chain
\[ \frac{d}{dt} = \frac{\partial}{\partial T_0} + \epsilon \frac{\partial}{\partial T_1} + \dots \] \[ \frac{d^2}{dt^2} = D_0^2 + 2\epsilon D_0 D_1 + \epsilon^2 D_1^2 + \dots \]
Eliminating the Divergence
In the Duffing equation at \( O(\epsilon) \), we get:
\[ D_0^2 x_1 + x_1 = -2 D_0 D_1 x_0 - x_0^3 \]
The Strategy
Instead of solving for \( x_1 \) and finding a secular term, we require that the resonant driving terms on the RHS vanish.
"Solve the slow-time dynamics to keep the fast-time solution bounded."
The Ultimate Result
The solvability condition (no secular terms) leads to an evolution equation for the complex amplitude \( A(T_1) \).
For the Duffing Oscillator:
\[ \omega(A) \approx 1 + \frac{3\epsilon A^2}{8} \]
Physical Insight
The frequency "stretches" with amplitude.
This is the fundamental reason regular perturbation fails: it doesn't allow for frequency shifting.
Scales of Motion Derivation Scales of Motion
Poincaré-Lindstedt & Multiple Scales Masterclass
TARGET: ω(A) RELATION
The Problem
The Duffing equation \( \ddot{x} + x + \epsilon x^3 = 0 \) produces secular terms \( t \sin t \) in regular perturbation. We resolve this by assuming the frequency \( \omega \) is not fixed at 1, but rather shifts with \( \epsilon \).
Goal
Find the second-order correction to frequency such that the solution remains uniform in time.
1
Poincaré-Lindstedt Transformation
Introduce a new time variable \( \tau = \omega t \), where \( \omega = 1 + \epsilon \omega_1 + \epsilon^2 \omega_2 + \dots \). Rewrite the Duffing equation in terms of \( \tau \).
DERIVATION AREA: Transformation & Expansion
2
The Solvability Condition
At order \( \epsilon \), you will find a term of the form \( (2\omega_1 + \frac{3}{4}A^2) \cos \tau \). For a uniform solution, this term must be zero. Solve for \( \omega_1 \).
DERIVATION AREA: \(\omega_1\) Calculation
3
The Method of Multiple Scales (MMS)
MMS is more powerful because it can handle damping. Using \( T_0 = t \) and \( T_1 = \epsilon t \), derive the evolution equation for the amplitude \( A(T_1) \). Show that for the undamped case, it yields the same frequency shift as Poincaré-Lindstedt.
"Treat A as a function of slow time T₁..."
Computational Check
If we were to include a damping term \( \epsilon \gamma \dot{x} \), Poincaré-Lindstedt would fail because the solution is no longer periodic. Why does the Method of Multiple Scales survive this addition?
Memory and Jumps Slides Memory and Jumps
Lesson 4: Hysteresis in Driven Systems
NONLINEAR DYNAMICS UNIT
The Driven Duffing Oscillator
When we add periodic driving and damping:
\[ \ddot{x} + \gamma\dot{x} + x + \epsilon x^3 = F \cos(\Omega t) \]
where \( \Omega \) is the driving frequency and \( F \) is the force amplitude.
Key Question
How does the steady-state amplitude \( A \) respond to the driving frequency \( \Omega \)?
In linear systems, we get a simple resonance peak. In Duffing systems, the peak leans .
Frequency Response Curves
Linear (\( \epsilon = 0 \))
Symmetric Peak
Single-valued amplitude for every frequency.
NONLINEAR
Hard Spring (\( \epsilon > 0 \))
Peak Leans Right (High-Freq)
For a range of frequencies, three steady-state solutions exist: two stable, one unstable.
Hysteresis: The Path Matters
↑
Ramping Up
The amplitude follows the lower branch until it reaches the edge of the 'lean', then jumps to the high-amplitude branch.
↓
Ramping Down
The system stays on the high branch past the original jump point, before eventually dropping back down.
Bistability
The system 'remembers' where it came from.
Hysteresis Hunt Lab Hysteresis Hunt
Exploring Bistability in the Driven Duffing Oscillator
Sim/Lab Activity
Objective
Map the frequency-response curve of a Duffing oscillator. Identify the bistable region and record the jump frequencies \( \Omega_{up} \) and \( \Omega_{down} \).
System Parameters
Mass (m): 1.0
Damping (\( \gamma \)): 0.1
Stiffness (k): 1.0
Nonlinearity (\( \epsilon \)): 0.25
Driving Force (F): 0.3
The Upward Sweep
Increase \( \Omega \) from 0.5 to 2.0 in increments of 0.05. Wait for transients to die out. Record the steady-state amplitude \( A \).
Table: Frequency vs. Amplitude (Up)
The Downward Sweep
Decrease \( \Omega \) from 2.0 back down to 0.5. Does the jump occur at the same frequency? Note the region where the "Up" and "Down" data do not overlap.
Upward Jump Point (\( \Omega_{up} \))
Downward Jump Point (\( \Omega_{down} \))
Post-Lab Analysis
1. Stability Calculation
Based on your MMS derivation in Lesson 3, the steady-state amplitude satisfies the algebraic equation: \[ A^2 \left( (\Omega - \omega(A))^2 + \gamma^2 \right) = (F/2)^2 \] Compare your experimental jump frequencies to the theoretical points where \( dA/d\Omega = \infty \).
2. Physical Interpretation
Imagine you are designing a MEMS resonator. Why would the existence of hysteresis be a "feature" in some applications but a "failure" in others?
Fractal Footprints Slides Fractal Footprints
Lesson 5: Poincaré Sections & Chaos
The End of Predictability
When Perturbation Fails
As the nonlinearity \( \epsilon \) or driving force \( F \) increases, the "slow and fast" assumption breaks down.
The Limit of Mastery
"Analytical methods can approximate the orbit, but they cannot predict the sensitivity to initial conditions that characterizes chaos."
Loss of periodic stability
Period doubling bifurcations
The Strange Attractor
The Poincaré Section
Slicing through High-Dimensional Motion
Instead of watching the continuous path in phase space \((x, \dot{x})\), we take a "snapshot" every period of the driving force:
tₙ = n(2π/Ω)
If the motion is periodic, we see a single point. If it's chaotic, we see a fractal dust .
[Visual: A Poincaré Map of the chaotic Duffing oscillator]
Period Doubling
Period 1
System repeats every drive cycle.
Period 2
System repeats every TWO cycles.
Period 4
Doubling sequence accelerates.
CHAOS
Deterministic but Aperiodic.
Strange Attractors Exit Ticket Exit Ticket: Strange Attractors
NONLINEAR DYNAMICS :: SEQUENCE FINALE
Researcher ID
1 Distinguishing Chaos from Noise
A student observes the output of a Duffing oscillator and sees a signal that never repeats and appears random. How would a Poincaré map allow the student to distinguish between high-dimensional stochastic noise and low-dimensional deterministic chaos?
Mention geometric structure in your answer...
2 The Predictability Horizon
In the chaotic regime, two trajectories starting at \( x_0 \) and \( x_0 + \delta x_0 \) will diverge exponentially. If the Lyapunov exponent is \( \lambda > 0 \), estimate how the "time to loss of predictability" scales with the precision of your measurement \( \delta x_0 \).
Sequence Synthesis
Throughout this unit, we have moved from the "Simple" in Harmonic Motion to the "Complex" in Chaos. What is the most significant conceptual shift in your understanding of mechanical systems after studying nonlinearities?
