Compare PSD to db 1

[colorforadollar]

I need to compare the far field acoustic pressure that is predicted in a simulation against experimental measurements. However, the experimental results are given to me in PSD [dB/sqrt(Hz) vs Hz] and I need to compare to a simulation result of a db at a single frequency. I cannot get the experimental results in dB for some reason. Is there a way to compare these two pieces of data? I'm a hydro guy who put all the work into the simulation but the bottom line comparison of all my work is this one comparison.

 

[BrianE22]

I've not worked with dBs so your units of dB/sqrt(Hz) are a bit strange to me. I'm used to PSD in units of (g's squared)/Hz. (for accelerations).

For (EU squared)/hz you would multiply the value of the PSD by the frequency resolution ("bin width") used to generate the PSD. For instance, if the value of the PSD was 10 g's squared/Hz at a value of 20 hz and the frequency resolution was .2 Hz then the average power at 20 hz would be 10 x .2 = 2 g's squared.

 

[colorforadollar]

Can I determine the frequency from the PSD plot or the x,y used to plot it? The PSD data is plotted in 200 Hz increments. The range of the PSD is 0 to 100 kHz.

 

[BrianE22]

Whoever digitized the data would have to tell you the frequency resolution (or, given sample rate and FFT size, you could calculate it: DeltaF = sample rate/FFT length). You're probably pretty safe though in assuming the resolution in your case is 200 hertz.

 

[GregLocock]

" The PSD data is plotted in 200 Hz increments. The range of the PSD is 0 to 100 kHz. "

Might want to check that, it sounds absurd.

first convert your psd to Pa/sqrt(hz) using Pa=10^(dB/20).

then multiply it by sqrt(hz), whatever the frequency step is.

Then square it

then add all the lines, then square root to give rms Pa, then back into dB=20*log(Pa)

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?