I am trying to use ABAQUS/ANSY to calculate the radiated acoustic pressure (outside the muffler)from a mufflter (due to resonance of muffler shell). I want to INPUT acoustic pressure at the inlet of the muffler. Acoustic pressure is the difference between instantaneous pr. and static pr. If that is the case, IS DYNAMIC PRESSURE obtained from CFD of muffler can be used as ACOUSTIC PR??? Thanks. If not how to get this info.

Larry,

We are doing many exhaust aero-acoustic numercial simulations with our code, RADIOSS-CFD. We believe that it is fairly difficult to get a good answer in a muffler by using a decoupled CFD/Structure approach. For instance, the following factors have a great influence on the system response: (i) There is a good likelyhood that there will be noise generation through turbulence INSIDE the muffler. Therefore, you cannot derive the structure excitation from outside BCS. (ii)You need to take into account the external air acoustic impedance. We have had some problems last spring with bad exterior air impedance boundary condition on the sides of a muffler yielding a unphysical structure resonance overexcitation and a spurious sound peak.

If you want more details on our fully coupled approach, you are welcome to e-mail me and I'll try to answer your questions. You can also have a look at our website www.mcube.fr under the M-Explicit section.

Dimitri

No, You did not ANSWER MY question.

My question was: Acoustic pressure is the difference between instantaneous pr. and static pr. If that is the case, IS DYNAMIC PRESSURE obtained from CFD of muffler can be used as ACOUSTIC PR???

FOR YOUR INFO: Regarding FEA of muffler, ABAQUS can consider fluid & other material impedances on both inside and outside of the muffler and NO SPECIAL ADDITIONAL SOFTWARE is necessary.

Let me give it a shot, since I am currently struggling with the same topic (CAA):

Acoustic pressure are very small amplitude (less than 1 Pa) static pressure fluctuations felt in the far-field of unsteady flow. The frequency of these fluctuations are usually in the audible range (20Hz to 20kHz). Computational aeroacoustics (CAA) generally involves a two step process.

Let me use the example of 2-D unsteady flow over a circular cylinder with alternating vortex shedding (subcritical Re#). In CAA, one first computes the unsteady flowfield around the cylinder using general CFD code (NS solver). Once the pressure fluctuations at the surface of the cylinder are known for a large number of vortex shedding periods, the acoustic pressure fluctuations at the far-field (say r/D = 30) are then computed using a partial differential equation and boundary conditions known as the Lighthill aeroacoustic analogy (or Lighthill-Curl). The two-step process is used because most flow solvers are too dissipative to preserve the far-field acoustic pressure fluctuations, which is orders of magnitude smaller compared to the near-field.

I hope this helps. Also, for a better understanding on acoustics, check out the following links:

www.vibrationdata.com

www.campanellaacoustics.com/faq.htm

Fantastic! That is exactly what I wanted to know. Now, I have another question: I am trying to model an automotive MUFFLER STRUCTURAL SHELL NOISE. And I need Acoustic pressure as an INPUT or Boundary condition at the INLET of the muffler so that I can predict the ACOUSTIC PRESSURE generated at the outside of muffler due to resonance of muffler structural shell. HOW TO GET THIS INPUT ACOUTIC PRESSURE at the inlet on muffler? We have 1 dimensional GT-POWER for wave calculation? Assume that I know the frequency of this acoutic pr. input. Thanks for the correct definition. Lot of people just use 1 Pa as an input and get the result and modify the muffler (design modification) and use 1 pa again and compare the results. This is great if we want to compare the designs. But if we want to actually simulate, we can not use 1 pa. we have to use the actual ACOUTIC PR. value??

I used the cylinder as an example, because I really don't know anything about mufflers. Also, I can't speak for fluid-solid interaction. Thus I did not exactly answer your questions, and I am not sure if I can be of help much further on this topic.

Can you somehow measure or obtain data for inlet conditions such as frequency, amplitude and wave form of these pressure waves and then recreate them numerically? I would think your pressure amplitude at the muffler inlet is larger than 1 Pa. What is GT power?

Thanks again Axel. GT power is a 1 diminsional acoustic software used in auto world for acoustic simulation. I just now asked my colleague who uses the GT power. She is telling me that yes, it is possible to get ACOUSTIC PR near the inlet of the muffler and I can use that as an input for my calculations. Let me see if she can REALLY give me the data I need.

Good luck! Sorry I can't assist you further.

Actually,

When applied to a muffler or more generally a cavity, there is a big problem in this approach since the solution that the incompressible NS solver that are usually use will NEVER allow the cavity modes to be computed. Let me give you a simple example: let say you want to compute the behavior of a rotating simplified tire excited by the ground. You should find a shift and split of the cavity modes due to propagation speed difference for the noise going upstream and downstream. I doubt that you can solve such a problem with Incompressible NS + Lighthill ( http://www.mcube.fr/M-Explicit/appli...ire/frame.html )

Last, in a muffler, there will be some coupled effect of the turbulence generating noise AND sound pressure waves interacting with turbulence.

Good luck anyway!

Dimitri