Do you have access to accelerometers and the data acquisition system to record the vibration on the surface of the compressor?  To simulate sound pressure level (SPL) in the air around an object that is vibrating, you need to know the vibration magnitude and frequency of the surfaces on the object.

Bolts that fasten a compressor to a large flat plate can incease the noise level. Vibration is transfered to the plate which will have different natural frequencies than the compressor and a large surface area to create noise in the air.  Vibration isolation supports are often added to a design to limit how much vibration transfers into the plate the compressor is mounted to. Even without vibration isolation supports, changing the thickness of the plate could have an impact on the noise sent into the surroundings.

Bolts in the presence of vibration tend to become loose if they are not torqued properly. I doubt lower torque will have any effect on the noise, but it could result in a fatigue failure of one of the bolts in the lifetime of the compressor.

You could separate the contribution of the mounting plate to the total environmental noise from the noise coming from the compressor itself if you were to hang the compressor from wires and measure the noise, or at least put a large thick foam pad on the floor under the compressor that has no large plate bolted to it.

You could put an enclosure around the compressor to reduce the noise being sent into the environment.

If your goal is to reduce the noise, an experimental approach is best. 

If your goal is to learn to use acoustic simulation software, choose a simple problem where you can simulate the vibration of the object creating the sound. That is what I did 4 years ago when I made this video.