Hello, we use sampling frequency 9 kHz and expect Strouhal number 0,18 for cylinder as found in literature. Re number is about 50 000. Now we have installed two single wire probes in horizontal position and try to calculate the frequency spectrum from difference of two signals.

We use fast Fourier transform and low pass filter too. I have succesfully measured the vortex frequency behind simple tall cylinder (dia. 55 mm, 600 mm high), and then I replace the cylinder with the cooling tower model (155 mm, 250 mm high) and see if there is any progress :-) Thank you anyway.

Well, I have measured the vortex frequency behind cooling tower and Strouhal number is lower than behind a cylinder. It works, but you have to find the right place for the probe :-)

I'm late to coming across this question, but I think it is over-simplifying the matter to look for a single shedding frequency given the dimensions of your cooling tower model. Unless I misunderstand the shape of the model and its placement in the tunnel, I would expect 3-D flow effects to be very important in determining the flow's topology in the model's wake. A clear shedding frequency would be expected for the tall cylinder you mentioned, but not for the cooling tower model, even if it is truly cylindrical.

Petr Michálek. You are working in a tunnel with developed boundary layer _, soon there is an interaction between the developed boundary layer and _ Kármán's wake as soon frequency is not necessarily the same. Something else, the traditional charts frequency vortex shedding are made for cylinders whith infinite length. In your case you have the boundary layer  influence _, neighborhoods (if you're simulating an actual case) and the top of the tower. The last observation of Chris's perfect, do not expect the same behavior for a cylinder with a whith infinite length with cylinder with the height / width ratio varies from 11:1 to 1.6:1, and more, with the influence _ of the soil!

You made ​​a previous analysis of the global flow with _ smoke tracer or otherwise? Perhaps better identify the region to put probes

I worked in a tunnel boundary layer _ in 1986 by measuring the characteristics of the turbulent developed boundary layer, so you can see that my opinion is not the most updated.

I would like to add few points with Chris answer. Certainly 3D effects present in the cooling tower wake. The positioning of the hot wire plays a major role (axial and lateral direction). I usually traverse the probe in the lateral direction and finds the frequency with maximum amplitude of FFT (power spectrum).  I do change the sampling frequency to twice and thrice to validate the obtained frequency. Since Re 50000, the wake will be turbulent for even 2D cylinder. The 3D effects (cooling tower), free stream turbulence further complicate the flow behind the cylinder. It will be appropriate to chose the most dominant frequency which can be obtained by traversing the probe in the lateral direction