1 / 53

Environmental Data Analysis with MatLab

Environmental Data Analysis with MatLab. Lecture 10: Complex Fourier Series. SYLLABUS.

hanley
Download Presentation

Environmental Data Analysis with MatLab

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Environmental Data Analysis with MatLab Lecture 10: • Complex Fourier Series

  2. SYLLABUS Lecture 01 Using MatLabLecture 02 Looking At DataLecture 03Probability and Measurement ErrorLecture 04 Multivariate DistributionsLecture 05Linear ModelsLecture 06 The Principle of Least SquaresLecture 07 Prior InformationLecture 08 Solving Generalized Least Squares ProblemsLecture 09 Fourier SeriesLecture 10 Complex Fourier SeriesLecture 11 Lessons Learned from the Fourier Transform Lecture 12 Power SpectraLecture 13 Filter Theory Lecture 14 Applications of Filters Lecture 15 Factor Analysis Lecture 16 Orthogonal functions Lecture 17 Covariance and AutocorrelationLecture 18 Cross-correlationLecture 19 Smoothing, Correlation and SpectraLecture 20 Coherence; Tapering and Spectral Analysis Lecture 21 InterpolationLecture 22 Hypothesis testing Lecture 23 Hypothesis Testing continued; F-TestsLecture 24 Confidence Limits of Spectra, Bootstraps

  3. purpose of the lecture switch from Fourier Series containing sines and cosines to Fourier Series containing complex exponentials

  4. purpose of the lecture • What we want to do is switch from Fourier Series containing • sin(ωt) and cos(ωt) • Toa Fourier series containing • exp(-iωt) and exp(+iωt) • Why would we do that? • Complex numbers are sometimes more complicated to comprehend • But in this case a single exponential can be manipulated more easily than a sum of sinusoids

  5. review of complex numbers

  6. imaginary unitisuch thati2 = -1

  7. complex numbera = ar + iai • This is a complex plane • Complex number can be viewed as a point or position vector in a 2D Cartesian coordinate system • Real part is horizontal component • Imaginary part is vertical real part imaginary part • If real part is 0, the complex number is purely imaginary • If imaginary part is 0, you have areal number

  8. adding complex numbersa = ar + iai b = br + i bic = a+b = (ar + iai)+ (br + i bi ) = (ar + br )+ i(ai+bi ) ci cr … just add real and imaginary parts, separately

  9. subtracting complex numbersa = ar + iai b = br + i bic = a-b = (ar + iai)-(br + i bi ) = (ar - br )+ i(ai- bi ) ci cr … just subtract real and imaginary parts, separately

  10. multiplying complex numbersa = ar + iai b = br + i bic = ab = (ar + iai)(br + i bi ) == arbr + iar bi + iaibr + i2aibi =(arbr - aibi )+ i(ar bi + aibr ) cr ci … like multiplying polynomials

  11. Complex Conjugates • A pair of complex numbers with the same real part, but with imaginary parts of equal magnitude and opposite signs • 3 + 4i and 3 – 4i are complex conjugates • Book uses notation a*, but might also see “a bar” complex conjugate, a*a = ar + iai a* = ar - iai

  12. absolute value, |a ||a|= [ ar2 + ai2 ]½note|a|2 = a* a

  13. Euler’s Formulaexp(iz) = cos(z) + i sin(z) … where does that come from ???

  14. start from the Taylor series expansionsexp(x) = 1 + x + x2/2 + x3/6 + x4/24 + …cos(x) = 1 + 0 - x2/2 + 0 + x4/24 + … sin(x) = 0 + x + 0 - x3/6 +0 + …

  15. exp(x)substitute in x=izexp(x) = 1 + x + x2/2 + x3/6 + x4/24 + …exp(iz) = 1 + iz - z2/2 - iz3/6 + z4/24 + …

  16. cos(x)substitute in x=zcos(x) = 1 + 0 - x2/2 + 0 + x4/24 + …cos(z) = 1 + 0 - z2/2 + 0 + z4/24 + …

  17. sin(x)substitute in x=z and multiply by i sin(x) = 0 + x + 0 - x3/6 +0 + … i sin(z) = 0 + iz + 0 - iz3/6 +0 + …

  18. add cos(z) and i sin(z)cos(z) = 1 + 0 - z2/2 + 0 + z4/24 + …+i sin(z) = 0 + iz + 0 - iz3/6 +0 + …=cos(z) + i sin(z) = 1 + iz - z2/2 - iz3/6 + z4/24 + …compare with results for exp(iz)exp(iz) = 1 + iz - z2/2 - iz3/6 + z4/24 + … they’re equal!

  19. note that sinceexp(iz) = cos(z) + i sin(z)then |exp(iz) |= cos2(z) + sin2 (z) = 1

  20. any complex number can be writtenz = r exp(iθ) where r = |z | and θ=tan-1(zi/zr)

  21. MatLabhandles complex numbers completely transparentlya = 2 + 3*i;b = 4 + 6*i;c = a+b;works just fine

  22. Warning!accidentally resetting i to somethingother than iis so easyi=100; (and then you get nonsense) so execute aclear i;at the top of your script if you plan to use i

  23. or use the alternate notationa = complex(2,3);b = complex(4,6);c = a+b;which is safer

  24. end of review

  25. Euler’s Formulascomplex exponentials can be written as sines and cosines

  26. or reverse themsine and cosines can be written as complex exponentials

  27. so a Fourier Seriesalternatively can be writtenas a sum of sines and cosinesor a sum of complex exponentials

  28. paired terms and sines and cosines of the same frequency,ω, are paired • positive and negative complex exponentials of the frequencies,ω, are paired • … so as to be able to represent a time-shifted sinusoid

  29. A and B and C- and C+must be related

  30. A B =0 implies CI- = -CI+ =0 implies CR = -CR+

  31. summary Cr is an even function of frequency • Ci is an odd function of frequency • coefficients at corresponding positive and negatives are complex conjugates • C(-ω) = C*(ω)

  32. illustration of symmetry of ω Cr Ci

  33. old-style Fourier Series non-negative frequencies only from 0 to ωny

  34. new-style Fourier Seriesor “Inverse Discrete Fourier Transform” added for compatibility with MatLab non-negative frequencies negative frequencies

  35. why the weird ordering of frequencies? • ωn = ( 0, Δω, 2Δω, …,½N Δω, -(½N-1) Δω, …, -2 Δω, -Δω ) • same as • ωn = ( 0, Δω, 2Δω, …,½N Δω, (½N+1) Δω, …, (N-1)Δω, NΔω ) non-negative frequencies negative frequencies non-negative frequencies up to Nyquist non-negativefrequencies above the Nyquist

  36. least-squares solution for the Fourier coefficients, Cnor “Discrete Fourier Transform” … derivation requires complex version of least-squares. See text of details …

  37. MatLabFourier Coefficients Cj from time series dnc = fft(d); vector of N data vector of N complex Fourier coefficients

  38. MatLabtime series dnfrom Fourier Coefficients Cjd = ifft(c); vector of N complex Fourier coefficients vector of N data

  39. standard setup M=N; tmax=Dt*(N-1); t=Dt*[0:N-1]'; fmax=1/(2.0*Dt); df=fmax/(N/2); f=df*[0:N/2,-N/2+1:-1]'; Nf=N/2+1; dw=2*pi*df; w=dw*[0:N/2,-N/2+1:-1]'; Nw=Nf;

  40. standard setup same number M of Fourier Coefficients as number N of data M=N; tmax=Dt*(N-1); t=Dt*[0:N-1]'; fmax=1/(2.0*Dt); df=fmax/(N/2); f=df*[0:N/2,-N/2+1:-1]'; Nf=N/2+1; dw=2*pi*df; w=dw*[0:N/2,-N/2+1:-1]'; Nw=Nf;

  41. standard setup maximum time, for N data sampled at Dt M=N; tmax=Dt*(N-1); t=Dt*[0:N-1]'; fmax=1/(2.0*Dt); df=fmax/(N/2); f=df*[0:N/2,-N/2+1:-1]'; Nf=N/2+1; dw=2*pi*df; w=dw*[0:N/2,-N/2+1:-1]'; Nw=Nf;

  42. standard setup M=N; tmax=Dt*(N-1); t=Dt*[0:N-1]'; fmax=1/(2.0*Dt); df=fmax/(N/2); f=df*[0:N/2,-N/2+1:-1]'; Nf=N/2+1; dw=2*pi*df; w=dw*[0:N/2,-N/2+1:-1]'; Nw=Nf; time column-vector

  43. standard setup M=N; tmax=Dt*(N-1); t=Dt*[0:N-1]'; fmax=1/(2.0*Dt); df=fmax/(N/2); f=df*[0:N/2,-N/2+1:-1]'; Nf=N/2+1; dw=2*pi*df; w=dw*[0:N/2,-N/2+1:-1]'; Nw=Nf; Nyquist frequency, fny

  44. standard setup M=N; tmax=Dt*(N-1); t=Dt*[0:N-1]'; fmax=1/(2.0*Dt); df=fmax/(N/2); f=df*[0:N/2,-N/2+1:-1]'; Nf=N/2+1; dw=2*pi*df; w=dw*[0:N/2,-N/2+1:-1]'; Nw=Nf; frequency sampling, Δf

  45. standard setup M=N; tmax=Dt*(N-1); t=Dt*[0:N-1]'; fmax=1/(2.0*Dt); df=fmax/(N/2); f=df*[0:N/2,-N/2+1:-1]'; Nf=N/2+1; dw=2*pi*df; w=dw*[0:N/2,-N/2+1:-1]'; Nw=Nf; frequency column- vector

  46. standard setup M=N; tmax=Dt*(N-1); t=Dt*[0:N-1]'; fmax=1/(2.0*Dt); df=fmax/(N/2); f=df*[0:N/2,-N/2+1:-1]'; Nf=N/2+1; dw=2*pi*df; w=dw*[0:N/2,-N/2+1:-1]'; Nw=Nf; number of frequencies between 0 and Nyquist

  47. standard setup M=N; tmax=Dt*(N-1); t=Dt*[0:N-1]'; fmax=1/(2.0*Dt); df=fmax/(N/2); f=df*[0:N/2,-N/2+1:-1]'; Nf=N/2+1; dw=2*pi*df; w=dw*[0:N/2,-N/2+1:-1]'; Nw=Nf; angular frequency sampling, Δω

  48. standard setup M=N; tmax=Dt*(N-1); t=Dt*[0:N-1]'; fmax=1/(2.0*Dt); df=fmax/(N/2); f=df*[0:N/2,-N/2+1:-1]'; Nf=N/2+1; dw=2*pi*df; w=dw*[0:N/2,-N/2+1:-1]'; Nw=Nf; angular frequency column- vector

  49. standard setup M=N; tmax=Dt*(N-1); t=Dt*[0:N-1]'; fmax=1/(2.0*Dt); df=fmax/(N/2); f=df*[0:N/2,-N/2+1:-1]'; Nf=N/2+1; dw=2*pi*df; w=dw*[0:N/2,-N/2+1:-1]'; Nw=Nf; number of angular frequencies between 0 and Nyquist

  50. Computing Power Spectral Density % compute Fourier coefficients mest = fft(d); % compute amplitude spectral density s=abs(mest(1:Nw)); % compute power spectral density s2=s^2;

More Related