No obvious maximum in power spectral density graph

I am extremely new to signal processing.

I have sampled the wind velocity downstream of a fan at 100 HZ and here is the time history:

I would like to plot the power density. Using pwelch I get the following plot which doesn't seem meaningful to me. I simply use pwelch(V,[],[],[],100). Based on Matlab documentations, I should get a curve with a peak! Can someone help me figuring out what the problem is? Or if this odd-shaped pwelch plot is actually meaningful and I can get it?

Answer 1

[Star Strider]

You are seeing a peak, just not the one you’re anticipating. The reason is that you have a significant d-c (constant) offset in your signal.

To Illustrate —

t = linspace(0, 1);
V1 = sin(2*pi*t) + randn(size(t));
V2 = sin(2*pi*t) + randn(size(t)) + 10;

figure
pwelch(V1,[],[],[],100)

figure
pwelch(V2,[],[],[],100)

See if subtracting the mean of your signal produces the result you want:

pwelch(V-mean(V),[],[],[],100)

I¹m thinking it will.

[sarah Abdellahi]

Thank you for your response. I see your point but When I subtract the mean, I still see the same graph. Here is the graph: I mean I dont see an obvious peak at the middle of the spectrum!

I can also attach the vector if that would help to solve the problem.

[Star Strider]

The peak is there. It is at very low frequencies. The rest is noise.

If you want to see the amplitude spectrum, use the fft (link) function. That may actually be what you want.

Remember to subtract the mean to calculate it as well, so the d-c offset does not obscure the peaks you are interested in.