Need an intro to Bode plots & Campbell diagrams w/ theory behind them

[jean15paul]

A little background: I have about 8 years of experience, but all of my experience has been with static analysis of structures. I recently started a new job where I'm responsible for both static and dynamic analysis including analysis of drive trains.

I've been doing a lot of normal modes analysis to find resonant frequencies, and yesterday my manager said that I should probably create a Bode plot to keep track of and present my results. He also said that he likes Bode plots more than Campbell diagrams.

I have absolutely no experience with either of these, and I don't even know where to start. I've been searching for info online, but haven't yet found a good introduction that included the necessary theoretical background. Can anyone help? Where can I get more information?

He also asked in Nastran can create Bode plots. Any insight?

(When it comes to dynamic stuff, talk to me like I'm stupid Thanks in advance.)

 

[IRstuff]

There is a NASTRAN forum on this site, where your second question is more likely to be answered.

Bode plots display transfer function and phase response as a function of frequency. IF you imagine a spot in your nodal analysis where there is resonance, and graph its input normalized amplitude and phase relative the input stimulus as a function of frequency, you'd have Bode plot.

TTFN
faq731-376

 

[jean15paul]

Thanks IRstuff. Yeah, the Nastran forum will probably be more help in how to actually create these plots, but before I get into that, I'm trying to understand what they are. I thought this was a better forum to ask. I'm really looking for background theory.

"IF you imagine a spot in your nodal" (modal?) "analysis where there is resonance",
I'm with you through here

"and graph its input normalized amplitude and phase relative the input stimulus as a function of frequency, you'd have Bode plot."
What is "input normalized amplitude and phase"? What input are you referring to? The forcing function?

 

[IRstuff]

Somewhere, your system is being shaken by some sort of forcing function, The amplitude at your node of interest divided by the amplitude of the forcing function is "normalized" to the input.

Generally, when something is resonating, it is no longer moving in concert with the rest of the system; this is variously described as "decoupling" or "out of phase." As a general rule, if something is resonating, it's pretty much moving out of phase relative to the rest of the structure. This is generally reflected in the modal plots as places of high stress, since the phase difference manifests as bending, etc.

TTFN
faq731-376

 

[GregLocock]

In typical NVH usage -

A Bode plot is a pair of graphs with the upper graph showing wrapped phase vs frequency on a log scale, and the lower graph shows the gain in dB vs frequency on a log scale. If you are doing a modal you will often use a linear scale for frequency, strictly speaking it isn't a bode plot any longer, I was told.

A Campbell's plot is a graph with rpm on the x axis, frequency on the y axis, and some representation of amplitude, either colour or blobs on the graph itself. Spectograms and waterfall plots are very similar concepts.

They aren't really interchangeable, a Bode plot is primarily used for transfer functions, a Campbell's diagram is for vibration, etc, during a speed sweep.




Cheers

Greg Locock


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

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