Approach to predict noise level and how to communicate it

[hirschaplin]

Hello,

Over the years, I have participated in a number of tenders for the supply of rotating machinery to various EPC contractors pretty much all over the world.

Nowadays, machinery noise level seems more important than ever. I guess mainly due to more stringent regulatory requirements pushing operators to ensure better field working conditions for their employees.

This is a bit problematic as we are required to guarantee machinery sound level @ 1 meter, typically the value has to be below 85 dBA or even less. More often than not we are also requested to accurately provide complete noise data for the product we offer. This is done by filling of the EPC company's standard noise data sheet table, something like this:

The product typically consist of a skid with a compressor driven by a motor either DOL or through a VFD. The compressor normally requires several utility systems with several small pumps and motors as well.

The problem is that the supply is 100% bespoke to each and every project. Even if it would be the same compressor model as supplied before, there is a big chance that there is a different material or that the process is different. Or that piping has another material or schedule (wall thickness) etc. There are so many factors that would affect the noise data...

Typically vendors of motors, compressors and pumps only provide simple noise data such as Sound pressure level Lp(A) xxx dB(A) and especially during bid phase. Let say we are able to obtain the following data:
Motor Lp(A) 82 dB(A)
Compressor Lp(A) 99 dB(A)
Utility pump Lp(A) 75 dB(A)
Utility motor Lp(A) 70 dB(A)

How can I use this simple data to calculate the mean/average sound of the complete skid with those items operating at the same time and how can I use that simple information to fill in the datasheet?

How are we expected to fill out such a data sheet for a bespoke skid during the bid phase and without possibility to measure until completion at final testing? What is the correct approach to handle / reply to such a request from a major EPC company?

 

[GregLocock]

Lw = Sound Power Level re 1E-12W (dB) Lp = Sound Pressure Level re 20mPa (dB)

The sound power is the total power of the source, whereas the pressure is specific to a particular location. In the far field you can convert from one to the other to some extent.

You can add the dBA up for a rough estimate of the overall sound power. The best thing to do is turn all the dB into power, add the power up then convert back to dB. dBskid=10*log10(10^8.2+10^9.9+10^7.5+10^7.0)= 99.1 dBA, because the compressor is so much louder than everything else. This is less reliable for pressures as they are directional.

You are being asked an unrealistic question for a bespoke system, I suspect what they really mean is if it is too loud you'll have to fix it.

Cheers

Greg Locock


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

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

 

[GregLocock]

Incidentally the sound pressure generated from a given sound power will depend on where it is installed, in the free field or a semi anechoic chamber it will be quieter than in a concrete room.

Cheers

Greg Locock


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

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

 

[geesaman.d]

I face the same requests (probably from some of the same EPCs) and like you, I have no way to provide noise predictions. The equipment I supply is driven by a gearbox and electric motor and we have for decades provided an overall limit of 85dBa. We use a wide variety of our gearbox designs with various reductions and various electric motors, so in this sense most every unit is bespoke. Most end up testing less than 80 but there is variability. We state at the very beginning that octal-level noise predictions are not available and will not be available*. They may do their own sound survey after the equipment is installed.

*We have performed the noise level testing for a couple of customers but it's a huge undertaking here and the personnel who last figured out how to cobble together a compliant testing arrangement no longer work for us. So it's generally never quoted as an option either.

Most customers accept the 85dBa overall because in most cases, the equipment is not enclosed, next to personnel working persistently in that location, and not next to other very loud equipment. With a little common sense the octave band noise data can be rationalized as unimportant.

The biggest issue I face with this is the EPC already agreed to provide this to the end user. And there is nothing an EPC hates more than to go back to their customer with what sub-suppliers can actually offer (vs. whatever generic equipment requirements that got shoved into the specs) or take on a risk for themselves. For an item like these octave-band sound surveys we typically need to take exception to it very clearly at the start and then keep the receipts to confirm it is not in scope numerous times throughout the project.