Rapid Resolution of a Complex Noise Complaint Case

A joint approach using The Noise App and analysis techniques provided by INVC proved to be an elegant and powerful approach to resolve the case.

Summary                                                     

The case concerned noise complaints about a low-frequency-hum that was thought to be emanating from a restaurant fan. As the fan was within a gated lightwell, council officers could not readily witness the noise, although an officer did manage to record a smartphone video clip near the source, meaning that the council was able to serve an abatement notice. The restaurant operator then stated that the issue had been resolved as the fan no longer operated at the complaint times and access to the lightwell had to be arranged in advance.

The complaints, however, continued. This was rapidly resolved by issuing the complainant The Noise App to log occurrences of the noise and record audio. Both the audio and video clips were then forwarded to the INVC for analysis that showed a tonal frequency match between the source and complainant recordings. This diagnostic analysis not only proved that the source was the restaurant fan, but it also explained why the noise was so difficult to witness. It could also pinpoint and cost the potential solutions.

The noise nuisance was resolved quickly once this evidence was sent to the restaurant operator.

Noise complaint investigation

On a cold morning, at the start of 2023, a complaint from a resident about a loud tonal noise before 6 a.m. triggered an inspection visit from Council Officer Marco Ditano. Witnessing the amplitude and irritating tonality of the sound, he was surprised that Westminster Council hadn’t received more complaints. There were no obvious sources, but just before the sound faded, he was able to determine that the noise seemed to be coming from the direction of a nearby restaurant.

He discovered a lightwell at the rear of a restaurant that included a large kitchen extract duct and fan plus a number of condensers. Whilst there, the hum returned so he recorded a short smartphone video and was convinced the source of the noise was the kitchen fan, not the condensers.

Abatement notice

He served a noise abatement notice under EPA section 80 on the restaurant operator, giving them 8 weeks to resolve the situation.  

The resident continued to report the noise, but when officers visited, they were unable to witness the sound as it seemed to be intermittent. Nonetheless, Marco contacted the restaurant operator to ask them to remedy the nuisance. He was informed that as the fan had been reprogrammed to run after 7 a.m., it could no longer be the problem.

Frustration resolved by The Noise App evidence

Whilst the resident continued to report a noise issue, officers were unable to experience it first hand, despite multiple time-consuming visits.

On seeking advice from the Environmental Sciences (ES) Team, Marco learned that deploying The Noise App would allow the complainant to provide evidence of the frequency, time and duration of the hum, without wasting further council resources.

After the resident had downloaded and used The Noise App, the recordings made it clear that the intrusive noise occurred regularly and that the tonal character constituted a statutory nuisance.

Marco forwarded the evidence to the restaurant operator who insisted that as the fan did not operate before 7 a.m. it could not be the cause of the noise problem. The ES Team reviewed The Noise App and video recordings, but whilst the noise from the lightwell (with fan and condensers) and the resident recordings sounded similar, it could not be construed as hard evidence of cause and effect without further analysis.

As the noise was so intermittent and the restaurant operator would not allow easy access to the lightwell, it seemed unlikely that officers would be able to make attended measurements for analysis.

Initial remote analysis: INVC interpretation and resolution

The ES Team emailed sound samples recorded on The Noise App by the complainant and the lightwell video clip video to the INVC for detailed analysis.

The result of the initial narrow band frequency analysis, shown in Figure 1, proved that the tonal noise recorded by the complainant is from the fan in the lightwell.

Figure 1: Overlay of source and complainant noise recording analyses

The cause of the nuisance is a classic fan noise signature tone at 384–385 Hz (a slight variation in fan noise frequency is very common as the load on the fan changes over time).

The ES Team wrote a technical report that included the INVC analysis, sending a copy to the restaurant operator. This was the key to enforcement of the section 80 notice. 

As a result, the restaurant operator resolved the noise problem within 10 days.

The ES Team subsequently returned to the site to make measurements of the kitchen extract equipment to confirm that it complied with Westminster Council’s plant noise criteria. The fan was found to be operating normally, with no tonality detected.

Forensic source diagnosis: INVC analysis

In addition to the simple analysis undertaken initially, a more detailed analysis of the recordings indicated that the tonal noise was caused by the blade-pass frequency of a belt-driven 12-bladed centrifugal fan running at 1920 rpm. It also showed that the tonal amplitude varied considerably over time. Moreover, the fan also had either damaged bearings and/or insufficient structural stiffness between the motor and the fan.

This diagnosis meant that the mitigation options (in priority order) were:

1. Check and replace bearings

2. Check and, if necessary, stiffen the fan/motor-mounting structure

3. If 1 and 2 prove unsuccessful, install a low-cost aerodynamic insert inside the fan to eliminate the tone at source. Conventional silencers are not an option as they would be inefficient and clog over time.

Analysis data

The above diagnosis was based on the following analyses:

1   Tonal amplitude variation

The Noise App’s 26-second recording analysis shows that the fan’s tone amplitude varied by 20 dB over that period, as shown by the following.

This shows that the fan blade-pass tone was very sensitive to back-pressure (quite common for some fan designs) which explains why it was so difficult to witness. When running with low backpressure, there was no significant tone.

2   Centrifugal fan fault diagnosis

Figure 4 shows a more detailed analysis of the lightwell video audio sample at low frequencies. There is a series of tones at harmonics of the 32 Hz fan speed (of which the fan blade-pass is the 12th harmonic), corresponding to a fan speed of 1920 rpm. As this is not a standard synchronous motor speed, the fan is either direct variable speed drive or belt driven.

The rotational speed harmonics in the signature indicate that there are worn bearings or that there are loose structural components.

The Noise App and Smartphone Microphones 

This case study will be of interest to professionals investigating tonal noise, as The Noise App is proven to be efficient at collecting and managing data in a secure environment accepted by legal professionals and the UK courts as admissible evidence in legal cases.

The Noise App is available on all smartphones operating on iOS or Android (see Device Support on the TNA website https://www.thenoiseapp.com/technical).

It is important to note that the quality of audio data collected is dependent on the quality of the smartphone. Anecdotally, 10 years of noise app data collection has shown that iPhones and most high-end Android devices will provide good quality audio data capture. Provided the limitations of smartphone microphones are taken into account The Noise App is a reliable tool.

Most smartphone manufacturers use microelectromechanical systems (MEMS) microphones in their devices. MEMS microphones typically have a sensitivity between 5 and 17.8 mV/Pa and can capture signals as low as 30 dB SPL and as high as 120 to 130 dB SPL (signal-to-noise ratio >60 dB). MEMS microphones have a flat frequency response similar to ceramic and condenser microphones used in type 2 noise dosimeters. There are limitations with the recording quality of low frequency noise (<200 Hertz) and noise of a very low level (<25dB(A)).

When recording noise on a smartphone, The Noise App bypasses the default software that filters the audio signals from the microphone. The app achieves this for all iOS and Android smartphones; it also disables any other noise suppression and noise cancellation software to ensure that the audio captured is as representative as possible and of sufficient quality to support professionals investigating noise. The Noise App captures over 1 million noise reports every year and is proven to be very efficient at taking good quality recordings of noise issues such as animal nuisance, loud music, anti-social behaviour, construction noise, industrial machinery noise and as in this case, tonal noise. 

More Information

The Noise App from RHE Global

Contact us for further information or support around noise nuisance issues.

Industrial Noise & Vibration Centre

Contact us via the web form or alternatively contact [email protected] or call 01753 698800