# Gunshot simulation with FEMLAB

1. Apr 27, 2006

### kaabdalis

Hi!

I'm about to start an acoustic analysis of the local shooting range.
The range is (as every other range in sweden) about to be forced to shut down since it's not PC to practice shooting in sweden.

However, the official reason is that the sound pollution is to great.

Im interested to simulate the shooting range and se where/how the sound can be reduced (i.e walls, ditches etc..)

I'm using FEMLAB 3.0 and the acoustics module to construct the model.

sofar i have built a transient model of a range measuring 200 x 500 meters.

My questions:

1. How do i simulate the shot fired?
-to ease calculations i have replaced the shoother with a 5 m radius sphere where the preassure spikes at 1 MPa for a very short time.
Is this a good approximation? What can be done to improve the model and still be able to keep the calculation time in minutes.

2. How about the boundary conditions? Since the sound propagates outside my 200x500 model there should be no interaction with the boundarys. However, it seems like the preassure wave do "bounce" at the boundarys. What boundary condition should I set?

3.I will be introducing buildings as well. How do change the ( ) "bouncability" of the different wall structures (wood, concrete, plastic, and so on..)

Best regards
Rickard

2. May 4, 2006

### PerennialII

Hi kaabdalis,

you've probably solved your problems but I'll return to the topic if there is something would like to discuss or could help.

1. Sounds in principle reasonable, although since haven't ran an analysis for gunshots can't really say much about the validity of the assumptions regarding the size of the sphere and the pressure spike (how did you arrive to these?). Know a few references can find out anyways if you haven't already solved this?

2. If the model is "large enough" (on the basis of the first results) something like a soft boundary condition might do it, getting the domain "to continue" can be accomplished with a radiation bc.

3. Might be done via a "transmission analysis", different acoustic properties of different domains. Probably don't need to do a fluid-structure analysis.

3. May 7, 2006

### kaabdalis

Hi PerennialII,

I havent solved everything yet.
As for the sound source i just made an estimated guess. The 5 m sphere was made in order to keep the mesh cleaner. Any suggenstions on how to make the gunshot more "real" is much appreciated.

I did try the "soft boundary" condition, but when I play the movie it still seems like the pressurewave bounces.
I also get som odd negative pressure reaings. This confuses me. Pressure cant be negative. (might be the reason for the starnge bounce)

Maybe the mesh isn't refined enough. ?

I will add buildings and walls in order to se how the sound behaves.

regards
Rickard

4. May 7, 2006

### PerennialII

.... I'll take a look at some references about how they've introduced the pressure spike. Instead of using the soft boundary condition, you might give it the radiation one (which you can also found 'by default' in femlab, probably a plane wave one depending on how you've constructed your model), it should eliminate the problems associated with soft/hard bcs (usually, naturally for example model size is a factor, but it's pretty easy to add excess domain to get it working). I'd probably first get the bcs ok and then work on the mesh, although in femlab creating an 'improved' mesh is but a click away .