So, you have time domain pressure data which you want to convert to frequency domain, in order to carry out a frequency domain vibration response calculation.

Due to periodicity of the rotor the pressure has to be periodic as well, thus frequency content of the pressure will concentrate on certain frequencies (rotational frequency harmonics). You can obtain this pressure amplitudes for those frequencies calculating the fourier series coefficients but also by FFT. FFT is just an algorithm to compute fourier transforms, so you can apply it if you make sure the frequencies you obtain are those you want, in this case the rotational frequeny harmonics I guess.

If you have extracted the pressure for one revolution, you have a time period that will give you a frequency resolution equal to the rotational frequency. Therefore, for that time period you'll get by FFT the amplitudes of the rotational frequency harmonics.

FFT and Fourier series are 2 manifestations of the same transform. Actually there are 4 'Fourier transformations':

1. Fourier transform or 'Continuous-Time Fourier Transform' (CTFT): Time domain and frequency domain are NOT periodic nor discrete;

2. Fourier series or  'Continuous-Time Fourier Series' (CTFS): Time representation is periodic, hence frequency representation is discrete;

3. Discrete-time Fourier Transform (DTFT): Time representation is discrete and frequency representation is periodic;

4. Discrete Fourier Transform (DFT/FFT): Time representation is discrete and periodic and frequency representation is also discrete and periodic.

See attached.

Thanks Alex McCloskey for the clear answer. I get it now. Performing FFT is the same as Fourier series decompistions since I extract the pressure values at each period.

Thanks very much.

regards.

Kossivi

Thanks Fernando Soares Schlindwein for the answer and the attached file.

regards.

Kossivi

Kossivi Gopki, take into account always the relationship between time period and frequency resolution, so be careful with the time period of your time data.  In purely periodic time data the frequency content is concentrated in certain exact frequencies, thus, small differences in the frequencies obtained by Fourier transform may give significantly different amplitudes.

Regards.

Thanks Ariadne Tsamban,

I stiil in confusion regarding the decomposition in series of fourier of the discrete pressure obtained with CFD.

Here is my concern :

I have 7 blades on my rotor. I performed the CFD. Here are my questions?

On the rotor since it's periodic should I extract the pressure at every period in order to perform the fourier decomposition?

Since I have 7 blades I guest to at least have 7 harmonics representing the effect of the blades. Is that right?

Can somebody help me find a fortran fourier decomposition code ? Because I have donne FFT and don't think FFT is the correct to use.

My regards to everybody.