EMA of air in a duct network
EMA of air in a duct network
(OP)
Does anyone have experience or pointers to EMA of a duct (e.g. intake system) network? I'm assuming that the system can be excited by a pressure pulse (or randomly) and then FRFs around the system could be measured and processed like classical EMA of structures. Anyone done this or read about it?





RE: EMA of air in a duct network
Use a reference pressure near, or an accelerometer on the loudspeaker, and integrate that to give velocity. If you were feeling dead posh you could even measure the intensity, but that is gilding the lily.
Then measure the pressure down the duct with a mic. Process as normal.
No real gotchas, visualising the results is a bit hard, but not vital. You could probably import them into an FE post processor to get pretty pictures of the modes.
Cheers
Greg Locock
Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.
RE: EMA of air in a duct network
That might seem a bizarre thing to do, but if you are primarily interested in the energy radiated from the end of the duct it might help. It is bizarre because of course a standing wave has zero intensity. The intensity of a pressure wave in a duct with moving fluid and changing temperature is fairly easy to derive from first principles, I seem to remember there wasn't that much literature to help.
I did this in an exhaust, the results were pretty neat.
Cheers
Greg Locock
Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.
RE: EMA of air in a duct network
I'm not sure what an intensity mode shape means though, or how to derive it. Do I just use instantaneous p(t).v(t) as my measure instead of p(t)? And then compute an FRF of that with the reference (speaker) velocity?
RE: EMA of air in a duct network
Cheers
Greg Locock
Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.
RE: EMA of air in a duct network
Thanks for the comments and assurance that this is worth pursuing.
I do seem to be getting some sensible mode shapes out now. I'm using piston mass flow (proportional to velocity) as my reference and pressure as my outputs (proportional to force and therefore 90deg out of phase with ref?). So I'm taking the real part of the FRF for my shapes. Something tells me that I may get better results if I use mass flow throughout and then take the imaginary part of the FRF.