AFMG Network Forum

ambupe · Member

Hi all,

I'm trying to model four differents machine tools with EASE. The only information available of each one is the sound power level emitted in each octave band.

What is the way for this? I tried to do with the SpeakerLab module but can not find any way to introduce the sound power level ...

What else can I do??In the Edit Project Module->insert loudspeaker, is there the SPL(1m) table. But how can I use the sound power levels??

Thanks!

Tim · Member

Sound Power Data is common for acoustic measurements of mechanical equipment and other large plant equipment.
It is possible to convert the SWL data to SPL referenced to 1m. The 'real' value may vary due to the measurement conditions. 360 degree polar data is rarely obtained for mechanical equipment and it is typical for measurements to be obtained in an anechoic environment at 1.5m or 3m distance.

Please refer to your trusty technical/reference literature, but a formula for SWL to SPL assuming free field conditions should be something like:
SPL=SWL – 10 log x 4 x (pye) x r2
where r = the measurement distance from noise source.

Hope this gets you started! Smile

ambupe · Member

The sound power values I have are according the ISO 230-5, using the standard ISO 3746, because is the most similar test code for the machine tools i have found. I have used a imaginary measurement surface (rectangular parallelepiped) to envelope the equipment with microphone positions located over that surface.

The test code recommend a distance d=1m, which i have used.

So that was the possibility to convert SWL->direct SPL i had thought , but i not was completely sure.

Thank you so much!!

Lindsay Smith · Member

At one meter the formula gives you a value of approximately -11 dB. I made an omnidirectional speaker with the output values of the one octave bands. Don't interpolate in-between values or you will add approximately 4 dB to the output.

My values were for the sound power level at a louver so I added 3 dB to the levels and placed my speaker next to an absorptive face I coated that section of the wall with.

The results look very odd because the frequency response looks like a square wave but the values are consistent between the EASE model and calculations made by others in our office.

Lindsay Smith
Seattle

Peter Patrick · Member

If you subtract 4.77 dB from the Octave values you will get values representing the same level in three 1/3 octave bands to produce the same octave band level.
Use that value in all the sphere's 1/3 octave bands for each octave.
Then you won't get the square wave frequency response.

All this assumes that the machine is small.
1.0m extrapolations are meaningless for large machines with a large near field.
That means you can use the 1.0m extrapolation provided you use it in the context of a far field outcome.

Best
P.J.P.