computational fourier...
TRANSCRIPT
![Page 1: Computational Fourier Analysissites.science.oregonstate.edu/~landaur/Books/CPbook/eBook/Lectur… · Most any periodic function “represented” by Fourier series. Most any nonperiodic](https://reader033.vdocuments.site/reader033/viewer/2022060607/605d45d229f98c37815bcf03/html5/thumbnails/1.jpg)
Computational Fourier AnalysisMathematics, Computing and Nonlinear Oscillations
Rubin H Landau
Sally Haerer, Producer-Director
Oregon State Universitywith National Science Foundation Support
Course: Computational Physics I
(Prerequisite: Intro Computational Science)
BMACC
Blended Multimodal Access to Computational Physics Curricula
1 / 1
![Page 2: Computational Fourier Analysissites.science.oregonstate.edu/~landaur/Books/CPbook/eBook/Lectur… · Most any periodic function “represented” by Fourier series. Most any nonperiodic](https://reader033.vdocuments.site/reader033/viewer/2022060607/605d45d229f98c37815bcf03/html5/thumbnails/2.jpg)
Outline
Text:Three Fourier Units
Unit I: Fourier series: Review & DFT(Only part of I today!)
Unit II: Signal filtering to reduce noiseUnit III: Fast Fourier transform (FFT)⇒
real time DFT
Today’s Lecture:
2 / 1
![Page 3: Computational Fourier Analysissites.science.oregonstate.edu/~landaur/Books/CPbook/eBook/Lectur… · Most any periodic function “represented” by Fourier series. Most any nonperiodic](https://reader033.vdocuments.site/reader033/viewer/2022060607/605d45d229f98c37815bcf03/html5/thumbnails/3.jpg)
Problem: Amount of sin ωt in Nonlinear Oscillation?
ExampleExample (in green): Sawtooth function
t
-1
0
1
y(t)
0 20-1
0
1
Y()
ω
What “frequencies” present? (What does this mean?)Meaning: “Fourier decomposition (right)”Periodic, many sinusoids; yet, one periodOnly steady state, let transients die out
3 / 1
![Page 4: Computational Fourier Analysissites.science.oregonstate.edu/~landaur/Books/CPbook/eBook/Lectur… · Most any periodic function “represented” by Fourier series. Most any nonperiodic](https://reader033.vdocuments.site/reader033/viewer/2022060607/605d45d229f98c37815bcf03/html5/thumbnails/4.jpg)
Physics Problem (see ODE): Nonharmonic Oscillator
Nonharmonic (nontranditional) oscillator
V (x) = k |x |p, p 6= 2
large p ⇒ square wellsquare well⇒ sawtooth
V
p = 2
xx
V
p = 6
Linear Nonlinear
Harmonic Anharmonic
Perturbed oscillator
V (x) = 12kx2 (1− 2
3αx)
Linear: 1st approximationAll Vs→ constant period TPeriodic not sinusoidal
xLinear
Nonlinear
Unbound
Harmonic
Anharmonic
V(x)
4 / 1
![Page 5: Computational Fourier Analysissites.science.oregonstate.edu/~landaur/Books/CPbook/eBook/Lectur… · Most any periodic function “represented” by Fourier series. Most any nonperiodic](https://reader033.vdocuments.site/reader033/viewer/2022060607/605d45d229f98c37815bcf03/html5/thumbnails/5.jpg)
Math (Theory): Fourier Series
Fourier’s Theorem: Function RepresentationSingle-valued periodic function ("signal")Finite number discontinuities
y(t) =a0
2+∞∑
n=1
(an cos nωt + bn sin nωt) (1)
an ↔ cos nωt “in” y(t)Intensity(ω), Power(ω) ∝ a2
n + b2n
True frequency ω ≡ ω1 = 2πT 6= HO frequency
Need know period T , y(t + T ) = y(t)Period may depend on amplitude
5 / 1
![Page 6: Computational Fourier Analysissites.science.oregonstate.edu/~landaur/Books/CPbook/eBook/Lectur… · Most any periodic function “represented” by Fourier series. Most any nonperiodic](https://reader033.vdocuments.site/reader033/viewer/2022060607/605d45d229f98c37815bcf03/html5/thumbnails/6.jpg)
Math: Fourier Series Fit
Fourier Series
y(t) =a0
2+∞∑
n=1
(an cos nωt + bn sin nωt) (2)
Theorem: Series = “Best Fit”
Least-squares sense: minimizes∑
i [y(ti)− yi ]2
⇒∑∞
i an cos nωt ' 〈 y(t) 〉⇒ miss discontinuitiesNeed infinite number of termsNot exact math fit (power series)Not good numerical solution (∞ terms)Not closed form analytic solution
6 / 1
![Page 7: Computational Fourier Analysissites.science.oregonstate.edu/~landaur/Books/CPbook/eBook/Lectur… · Most any periodic function “represented” by Fourier series. Most any nonperiodic](https://reader033.vdocuments.site/reader033/viewer/2022060607/605d45d229f98c37815bcf03/html5/thumbnails/7.jpg)
Fourier Integrals (Transforms)
Nonperiodic Functions
–1.0
–0.5
0.0
0.5
1.0
–6 –4 –2 0 2 4 6
t
s = 1, τ = 0
y(t) =
b0 + b1 cos ω1t + b2 cos 2ω1t + · · ·
Nonperiodic ≡ T →∞⇒ continuous ωMath:
∑→∫
Numerically:∫→∑
Numerically: the same!
Fourier Semantics
ω1 = 2πT = fundamental
nω1 = nth harmonicω > ω1: overtones
b0: DC component∑normal modes
Modes: see next slide
7 / 1
![Page 8: Computational Fourier Analysissites.science.oregonstate.edu/~landaur/Books/CPbook/eBook/Lectur… · Most any periodic function “represented” by Fourier series. Most any nonperiodic](https://reader033.vdocuments.site/reader033/viewer/2022060607/605d45d229f98c37815bcf03/html5/thumbnails/8.jpg)
Fourier Series for Nonlinear Oscillations?
y(t) =a0
2+∞∑
n=1
(an cos nωt + bn sin nωt) (3)
OK by Theorem if periodic, but
Nonlinear oscillators: ω1 = ω(A) = 2π/TIndividual terms 6= solutionLinear eqtn ↔ principle of linear superposition
y1(t), y2(t) = solutions (4)⇒ αy1(t) + βy2(t) = solution (5)
Useful if not too nonlinearBroadband spectrum⇒ chaosMany strong overtones
8 / 1
![Page 9: Computational Fourier Analysissites.science.oregonstate.edu/~landaur/Books/CPbook/eBook/Lectur… · Most any periodic function “represented” by Fourier series. Most any nonperiodic](https://reader033.vdocuments.site/reader033/viewer/2022060607/605d45d229f98c37815bcf03/html5/thumbnails/9.jpg)
Determine Fourier Coefficients an, bn
Project Out Components
〈ω|y〉 =
∫ 2π
0dtcos nωt
[a0
2+∞∑
n=1
(an cos nωt + bn sin nωt)
](6)
⇒(
an
bn
)=
2T
∫ T
0dt(
cos nωtsin nωt
)y(t) (7)
a0 = 2 〈y(t)〉Be smart: use SymmetryOdd y(−t) = −y(t) ⇒ an ≡ 0Even y(−t) = y(t) ⇒ bn ≡ 0Realistic calculations:
∫→∑
, y(t < 0) =? ⇒ small bn
9 / 1
![Page 10: Computational Fourier Analysissites.science.oregonstate.edu/~landaur/Books/CPbook/eBook/Lectur… · Most any periodic function “represented” by Fourier series. Most any nonperiodic](https://reader033.vdocuments.site/reader033/viewer/2022060607/605d45d229f98c37815bcf03/html5/thumbnails/10.jpg)
Determine Fourier Coefficients an, bn
Project Out Components
〈ω|y〉 =
∫ 2π
0dtcos nωt
[a0
2+∞∑
n=1
(an cos nωt + bn sin nωt)
](6)
⇒(
an
bn
)=
2T
∫ T
0dt(
cos nωtsin nωt
)y(t) (7)
a0 = 2 〈y(t)〉Be smart: use SymmetryOdd y(−t) = −y(t) ⇒ an ≡ 0Even y(−t) = y(t) ⇒ bn ≡ 0Realistic calculations:
∫→∑
, y(t < 0) =? ⇒ small bn
10 / 1
![Page 11: Computational Fourier Analysissites.science.oregonstate.edu/~landaur/Books/CPbook/eBook/Lectur… · Most any periodic function “represented” by Fourier series. Most any nonperiodic](https://reader033.vdocuments.site/reader033/viewer/2022060607/605d45d229f98c37815bcf03/html5/thumbnails/11.jpg)
Determine Fourier Coefficients an, bn
Project Out Components
〈ω|y〉 =
∫ 2π
0dtcos nωt
[a0
2+∞∑
n=1
(an cos nωt + bn sin nωt)
](6)
⇒(
an
bn
)=
2T
∫ T
0dt(
cos nωtsin nωt
)y(t) (7)
a0 = 2 〈y(t)〉Be smart: use SymmetryOdd y(−t) = −y(t) ⇒ an ≡ 0Even y(−t) = y(t) ⇒ bn ≡ 0Realistic calculations:
∫→∑
, y(t < 0) =? ⇒ small bn
11 / 1
![Page 12: Computational Fourier Analysissites.science.oregonstate.edu/~landaur/Books/CPbook/eBook/Lectur… · Most any periodic function “represented” by Fourier series. Most any nonperiodic](https://reader033.vdocuments.site/reader033/viewer/2022060607/605d45d229f98c37815bcf03/html5/thumbnails/12.jpg)
Determine Fourier Coefficients an, bn
Project Out Components
〈ω|y〉 =
∫ 2π
0dtcos nωt
[a0
2+∞∑
n=1
(an cos nωt + bn sin nωt)
](6)
⇒(
an
bn
)=
2T
∫ T
0dt(
cos nωtsin nωt
)y(t) (7)
a0 = 2 〈y(t)〉Be smart: use SymmetryOdd y(−t) = −y(t) ⇒ an ≡ 0Even y(−t) = y(t) ⇒ bn ≡ 0Realistic calculations:
∫→∑
, y(t < 0) =? ⇒ small bn
12 / 1
![Page 13: Computational Fourier Analysissites.science.oregonstate.edu/~landaur/Books/CPbook/eBook/Lectur… · Most any periodic function “represented” by Fourier series. Most any nonperiodic](https://reader033.vdocuments.site/reader033/viewer/2022060607/605d45d229f98c37815bcf03/html5/thumbnails/13.jpg)
Determine Fourier Coefficients an, bn
Project Out Components
〈ω|y〉 =
∫ 2π
0dtcos nωt
[a0
2+∞∑
n=1
(an cos nωt + bn sin nωt)
](6)
⇒(
an
bn
)=
2T
∫ T
0dt(
cos nωtsin nωt
)y(t) (7)
a0 = 2 〈y(t)〉Be smart: use SymmetryOdd y(−t) = −y(t) ⇒ an ≡ 0Even y(−t) = y(t) ⇒ bn ≡ 0Realistic calculations:
∫→∑
, y(t < 0) =? ⇒ small bn
13 / 1
![Page 14: Computational Fourier Analysissites.science.oregonstate.edu/~landaur/Books/CPbook/eBook/Lectur… · Most any periodic function “represented” by Fourier series. Most any nonperiodic](https://reader033.vdocuments.site/reader033/viewer/2022060607/605d45d229f98c37815bcf03/html5/thumbnails/14.jpg)
Example: Sawtooth Function, Analytic an, bn
Example
t
-1
0
1
y(t)
0 20-1
0
1
Y()
ω
y(t) =
{ tT/2 , 0 ≤ t ≤ T
2 (7)
t−TT/2 ,
T2 ≤ t ≤ T
Periodic, nonharmonic, discontinuous, sharp cornersOdd (⇒ sine series), obvious shift to left:
y(t) =t
T/2, −T
2≤ t ≤ T
2(8)
bn =2T
∫ +T/2
−T/2dt sin nωt
tT/2
(above) (9)
⇒ y(t) =2π
[sinωt − 1
2sin 2ωt +
13
sin 3ωt · · ·]
(10)
14 / 1
![Page 15: Computational Fourier Analysissites.science.oregonstate.edu/~landaur/Books/CPbook/eBook/Lectur… · Most any periodic function “represented” by Fourier series. Most any nonperiodic](https://reader033.vdocuments.site/reader033/viewer/2022060607/605d45d229f98c37815bcf03/html5/thumbnails/15.jpg)
Exercise: Sawtooth Function Synthesis
1 Sum Fourier series for n = 2,4,10,20 terms.2 Plot two periods & compare to signal.3 Check that series gives the mean value (0) at discontinuity4 Gibbs overshoot: Check that series overshoots
discontinuity by about 9%, and that this persists even whensumming over a large number of terms.
5 You deserve a break right now!
15 / 1
![Page 16: Computational Fourier Analysissites.science.oregonstate.edu/~landaur/Books/CPbook/eBook/Lectur… · Most any periodic function “represented” by Fourier series. Most any nonperiodic](https://reader033.vdocuments.site/reader033/viewer/2022060607/605d45d229f98c37815bcf03/html5/thumbnails/16.jpg)
Math (Theory): Fourier Transforms
Fourier series = right tool for periodic functions
Fourier Integral for Nonperiodic Functions
–1.0
–0.5
0.0
0.5
1.0
–6 –4 –2 0 2 4 6
t
s = 1, τ = 0
Wave packets, pulsesContinuous frequencies
Signal: y(t) =
∫ +∞
−∞dω Y (ω)
eiωt√
2π(11)
Think linear superposition: exp(iωt) = cosωt + i sinωtComplex transform Y (ω) ∼ (an,bn) ∼ coefficients1/√
2π, +iωt = physics conventions
16 / 1
![Page 17: Computational Fourier Analysissites.science.oregonstate.edu/~landaur/Books/CPbook/eBook/Lectur… · Most any periodic function “represented” by Fourier series. Most any nonperiodic](https://reader033.vdocuments.site/reader033/viewer/2022060607/605d45d229f98c37815bcf03/html5/thumbnails/17.jpg)
Fourier Transform: y(t)→ Y (ω)
t
-1
0
1
y(t)
0 20-1
0
1
Y()
ω
Fourier Amplitudes (an)→ Transform
Signal, function: y(t) =
∫ +∞
−∞dω Y (ω)
e+iωt√
2π(12)
Transform, spectral: Y (ω) =
∫ +∞
−∞dt y(t)
e−iωt√
2π(13)
Power Spectrum ∝ |Y (ω)|2, maybe log10 |Y |2
17 / 1
![Page 18: Computational Fourier Analysissites.science.oregonstate.edu/~landaur/Books/CPbook/eBook/Lectur… · Most any periodic function “represented” by Fourier series. Most any nonperiodic](https://reader033.vdocuments.site/reader033/viewer/2022060607/605d45d229f98c37815bcf03/html5/thumbnails/18.jpg)
Math Identity: Inversion of Transform
Substitute for y(t) and rearrange
Y (ω) =
∫ +∞
−∞dt
e−iωt√
2π
∫ +∞
−∞dω′
eiω′t√
2πY (ω′) (14)
Y (ω) =
∫ +∞
−∞dω′
{∫ +∞
−∞dt
ei(ω′−ω)t
2π
}Y (ω′) (15)
Depends on noncomputable Dirac delta function:∫ +∞
−∞dt ei(ω′−ω)t = 2πδ(ω′ − ω) (16)
Don’t try on computer!
18 / 1
![Page 19: Computational Fourier Analysissites.science.oregonstate.edu/~landaur/Books/CPbook/eBook/Lectur… · Most any periodic function “represented” by Fourier series. Most any nonperiodic](https://reader033.vdocuments.site/reader033/viewer/2022060607/605d45d229f98c37815bcf03/html5/thumbnails/19.jpg)
Summary
Most any periodic function “represented” by Fourier series.Most any nonperiodic function “represented” by Fourierintegral.Infinite series or integral not practical algorithm or inexperiment.Must know actual period.Period depends on amplitude in general.DFT (next): when done is simple, elegant and powerfulFast Fourier transform (FFT) is faster.That’s all folks!
19 / 1