Filtering of windspeed data

Hello,

I am searching for a filter to filter low frequencies out of my wind time series.

I tried some things out and ended up with negative wind speeds. Can you recommend a filter?

My time series is one year long with a time step of 10 min. I want to filter out all events with a cycle duration longer than 6h.

Thank you!

x-axis : Cycle duration y-axis : spectral energy density

Answer 1

[Star Strider]

I suggest a Chebyshev Type 1 bandpass filter for your application because it has a sharper cutoff than a Butterworth filter. I favour a bandpass design over a highpass filter because highpass filters have the unfortunate feature of also passing high-frequency noise. A bandpass filter will give you essentially your entire desired signal.

This is how I would design it (using the Signal Processing Toolbox):

Ts = 10;                    % Sampling interval 10 min
Fs = 1/Ts;                  % Sampling frequency (cycles/min)
Fn = 1/(2*Ts);              % Nyquist frequency

Fco1 = 1/(6*6);             % Low passband frequency 1/36 samples (6 hrs * 6 samples/hr)
Fco2 = 0.95*Fn;             % High passband frequency 

Fcn1 = Fco1/Fn;             % Normalised low passband frequency
Fcn2 = Fco2/Fn;             % Normalised high passband frequency

Ws = [Fcn1  Fcn2];          % Choose Chebyshev Type 1 Filter
Wp = Ws.*[1.1  1/1.1];
Rp = 1;                     % Passband ripple 1 dB
Rs = 40;                    % Stopband attenuation 40 dB
[n,Wp] = cheb1ord(Wp, Ws, Rp, Rs);
[b, a] = cheby1(n, Rp, Wp);

Consider using the second-order section implementation (see the Limitations discussion in the cheby1 documentation) with filtfilt.

See if this works for your application. I don’t have your data so I can’t test it.

[Anna]

Thanks a lot!

[Star Strider]

My pleasure!

(The sincerest expression of appreciation here on MATLAB Answers is to Accept the Answer that most closely solves your problem.)