PSD (Power Spectral Density) Interpertation/Understanding

[struclearner]

Dears,
I need help to understand PSD values interpretation.
When a PSD is created from acceleration values of a Time domain data from a random vibration scenario.
The acceleration values might occur more than once, or the data values might fall within a certain range/width of values.
A Histogram of the data will show the frequency content of the values.
A PSD will convert the Time domain data (acceleration) values to frequency domain data ((accelerationRMS)^2/Hz) values.
What the PSD data tell us?
Does PSD tells that Ordinate (y-axis) value occurrence is this many times, which is the frequency value on the abscissa (x-axis).
Can the PSD values be calculated from a Histogram of the Time domain data?

Thank you very much for your help of explanation and insight.

 

[IRstuff]

The PSD is the square of the Fourier transform of the time domain data, so no, you cannot get it from a histogram.

I suggest you read up on Fourier transforms.

TTFN (ta ta for now)
I can do absolutely anything. I'm an expert!

https://www.youtube.com/watch?v=BKorP55Aqvg

faq731-376 forum1529 Entire Forum list

http://www.eng-tips.com/forumlist.cfm

 

[electricpete]

I agree with IRstuff.

A histogram tabulates values into bins/categories, and you might loosely call the number of items in a given bin the "frequency" of occurrence of that value range.
But that type of "frequency" has absolutely nothing to do with the type of frequency studied in fourier analysis or psd analysis.

Does PSD tells that Ordinate (y-axis) value occurrence is this many times, which is the frequency value on the abscissa (x-axis).

No.

Can the PSD values be calculated from a Histogram of the Time domain data?

No.

What the PSD data tell us?

I think it deserves a bit of googling on your part to get an idea of the complexity of the topic and various definitions and applications of PSD.
I don't work with it much (work with FFT's a lot more). My two cents is that PSD is typically useful for studying and characterizing random signals, whereas as fft are more often more useful for periodic signals.






=====================================
(2B)+(2B)' ?

 

[Strong]

PSD = Power Spectral Density
You need a spectrum (amplitude versus frequency)and cannot get it from a histogram.
Amplitude values are squared

https://www.bing.com/search?q=power...Sekyl*MHpeSsnXbUv47*ssSVqm09g5dz6ToJg&plvar=0

https://community.sw.siemens.com/s/article/what-is-a-power-spectral-density-psd

https://blog.endaq.com/vibration-analysis-fft-psd-and-spectrogram

Walt

 

[IRstuff]

The only thing of note is energy is conserved across the Fourier transform, which is a good thing, so the integral of the PSD results in the total power of the measurand.

TTFN (ta ta for now)
I can do absolutely anything. I'm an expert!

https://www.youtube.com/watch?v=BKorP55Aqvg

faq731-376 forum1529 Entire Forum list

http://www.eng-tips.com/forumlist.cfm

 

[hacksaw]

struclearner

when dealing with "discrete" time series data, you use discrete transforms. They work the same way as traditional PSD methods, but with discrete time series data.

Electronic measurements use this too, as the data is digitized.