$5 million wasted on predictably ineffective Williston data center noise control
The Atlas Power data center built in Williston, Williams County, N Dakota caused a torrent of environmental noise complaints that were widely reported on the local TV news. The company promised to mitigate the problem by commissioning a truly massive noise wall from the Urban Solution Group that wasted $5 million. We analysed the TV news video soundtrack to show why...
As is typical of data centers, the problem noise is a hum generated by the host of cooling fans you can clearly hear in this KFYR-TV video...
This was a very costly mistake that could have easily been avoided as it was obvious from even the simplest diagnosis that the noise wall would have negligible effect. $5 million was wasted on the wrong solution, not to mention the $232,000 fines. See below for the simple, effective alternative solution.
Why did Atlas Power choose the wrong noise control solution?
Poor and incompetent advice. Presumably, acoustic advice was sought before spending $5 million on what turned out to be a bad guess. However, even the most cursory noise analysis would have shown that the noise wall would not and could not solve the noise complaint problem. Coming across the story online, analysis of a few seconds of the TV video clip proved that the problem was caused by a low frequency hum that could not possibly be mitigated by a noise wall. This analysis literally took less than one minute. One minute of diagnosis that would have saved a fortune...
The frequency signatures shown here overlay the noise close-in to the site and at a considerable distance taken from the video shown above. These show a high amplitude tone at 118Hz plus harmonics.
The results predicted that the tones are caused by 6 bladed fans running at 1180rpm - later confirmed (fan blade pass frequency, standard synchronous direct drive motors) with a simple, low cost solution...
Unfortunately, this kind of mistake (a lack of diagnosis leading to wasted expenditure on the wrong noise reduction measures) is commonplace across the world. In the UK, a 2023 survey of acoustic reports from noise consultants by the Environment Agency showed that 95% of them were inadequate. The same story, the same lack of diagnosis is repeated worldwide...
How does the wrong solution recommended to Atlas Power look and perform?
The installed noise walls are truly massive structures - not to mention an eyesore. They have also been completely ineffective as a means to stop the noise complaints.
This failure was completely and easily predictable.
Any competent noise consultant or company involved in noise control should have known that to be the case prior to installation.
Reportedly (as in the video), even the overall noise level remained unchanged after installation of the wall. Based on the news stories, it appears that no one carried out even the simplest noise problem diagnosis before recommending a high-cost solution.
Free Remote Diagnosis
Email us a video from your smartphone for an evaluation of best practice
This is a narrow band noise signature analysis taken from a video clip recorded after the noise wall had been installed. The low frequency fan hum at 118Hz remains at around a factor of x30 above the background noise.
As was easy to predict, there has been with no change in the typical tonal noise levels that have been the cause of the complaints.
It is a very basic acoustic fact that, due to the long wavelengths involved, it is almost never practical to attenuate low-frequency noise using acoustic barriers.
The low-cost and effective noise control alternative
There is an elegant engineering solution to this fan noise problem using aerodynamic fan noise control technology.
We use Computational Fluid Dynamics (CFD) to model airflow through the fan and then design simple retrofit aero modifications that have been applied to fans across the planet. The process is fast as it is based on photos, dimensions and smartphone recordings and does not require a site visit.
Not only would this approach eliminate the complaints by reducing the fan hum by around 99%, but would also be likely to increase the efficiency of the fans (in one case by 20%) and therefore improve cooling and reduce running costs. This could make the noise control project potentially self-financing and even profitable in the longer term.
More information on optimum data center noise reduction >
What conclusions can be drawn from this example?
Worldwide, the approaches used to reduce the increasing impact on health and lives associated with environmental noise are often very ineffective. Whether the cause of complaints is the extract system for a fast food restaurant or a massive data center, the problem is the same. There are serious consequences:-
- Health: many people are needlessly exposed to stressful environmental noise for prolonged periods, a serious health hazard.
- Industry: wastes a fortune on the wrong noise mitigation measures.
- Regulators: waste time and resources dealing with problems that should be resolved in weeks or months rather than years.
Why is environmental noise so often addressed inadequately?
- Inadequate regulations or ordinances: simplistic noise requirements that do not address common causes of complaints. Organisations can meet those requirements whilst still generating noise that is unacceptable to everyone in the area. The most common failing is the use of dB(A) without considering low-frequency noise (tonal hum in particular) that does not contribute to the dB(A) level (the "A" weighting progressively filters-out low frequencies) or tonal noise in general.
- Inadequate or ineffective use of regulations: even in countries that have environmental noise standards that potentially reflect the impact on people, the quality of noise reports is often very poor, resulting in bad advice.
- Lack of any diagnostic process: it is a simple premise; before making recommendations, you should diagnose the precise cause of the problem. Sadly, many recommendations are based on guesswork as there has been little or no diagnosis.
- Ignorance of modern noise control methods: even amongst noise consultants there is inadequate knowledge of the options leading to high costs, delays and failures.
Potential solutions?
Perhaps...
- Update noise ordinances or regulations? In principle, across much of the world, all that's needed to take a huge step forward in the ability to assess environmental noise accurately is a simple sound level meter, a smartphone and a free app. It's not rocket science, anyone could do this with an hour of training. Whilst we've helped update sophisticated Environment Agency noise guidance, the basics can be covered on 2 sides of A4. Contact us if you'd like to discuss a noise ordinance review or update.
- Put pressure on consultants and suppliers? Should regulators take action against incompetent consultants or suppliers and should professional bodies such as the Institute of Acoustics introduce and police strict quality controls with sanctions?
- Diagnosis, diagnosis, diagnosis: insist on proper diagnosis of the noise features causing complaints and the culprit noise sources. This should include narrow band frequency analysis.
- Get a second opinion: we provide remote noise diagnosis and control services based on simple noise data and smartphone recordings. This allows us to diagnose the precise causes of noise issues and to determine would would constitute the best practice solution. For this data center example, one minute of our time would have prevented the $5 million wasted expenditure...
We also have extensive experience in generating acoustic models of both existing and planned data centers. These can be used both to predict noise problems before they occur and to aid in defining the optimum noise mitigation measures for a site.
