The cause of low-frequency hum noise complaints from residents had been misdiagnosed in multiple reports. Consequently, no solution had been found despite extensive resources invested by both the company to whom the problem tonal noise had been ascribed and by the Environment Agency (EA). When contacted, we provided both our client company and the EA with the list of smartphone video recordings needed for the diagnosis, namely a recording close to each potential source plus off-site samples. This allowed us to frequency match the tones at the resident with those from the various items of plant.
Diagnosis showed the noise complaints were due to a pair of tones at 71Hz and at exactly 75Hz. All previous evaluations had mistakenly used 1/3 octave frequency analysis and assumed there was a single tone (despite the "beating" between the 2 tones - see audio and below). We diagnosed that the 75Hz was from a CHP generator on the company site (75Hz is a classic generator tone), but the other 71Hz tone was from a chiller on another site. We recommended a tuned exhaust silencer that eliminated the generator tone. The 71Hz can easily be resolved using aerodynamic fan noise control technology.
Poor noise diagnosis and reporting
This plot shows an overlay of a 1/3 octave analysis of the noise at the complainant and the (correct) narrowband analysis. This illustrates an all too common mistake by noise consultants using inadequate frequency analysis to investigate tonal noise problems. In this case, both problem tones are more or less in the same band (the 71Hz also contributes to the 63Hz band), hence the mistaken diagnosis that wasted considerable time and resources for both the company and the EA. As capable (free) FFT analysis phone apps are available, you have to question the competency of consultants using 1/3 octave frequency analyses to investigate tonal noise sources.
The correct noise diagnosis process
Narrowband frequency analysis of the tones that are the cause of complaints provides objective evidence of the subjective problem in a form that closely matches the human response making it the ideal metric to assess tonal noise nuisance.
In addition, it also provides very accurate and precise diagnostic information about the noise source. In this case, for example, we know both from calculation and experience that exactly 75Hz is a common tone from generator sets connected to 50Hz mains supplies and that 71Hz is in the common range for chiller fan tones.
A simple off-the-shelf tuned exhaust silencer fitted to the CHP engine reduced the problem 75Hz tone by c 19dB, eliminating this element of the nuisance noise.