Three people so far have suggested in the comments here it should be "easy" to triangulate it. Note that according to Zug Island's Wikipedia page, "As of April 2013, a Canadian scientist is using sound-level meters and a portable "pentangular array" of cameras and microphones to try to precisely identify the source of the sound, in order to know whom exactly to ask to fix it" and that "the City Council had already sp…
typical triangulation would depend on the sound being heterogeneous in waveform/volume right? If it's a constant hum then you'd only have volume dissipate as you move away which would get a lot harder as sound has different loss functions as it moves through different materials.
The sound pressure wave is picked up by one, then some few microseconds later, it appears on the other(s) a few centimeters away.
This is pretty straightforward with ordinary sounds that dominate the signal. You could pick out the wire which was struck in a piano, and even the nodes and harmonics along its length.
It gets more difficult at lower frequencies and with quieter signals, which precisely describes this exact situation.