Detect errors in the buildings technical infrastructure – before they occur

You've been through it before: You’re in a technical room, and everything  seems to be in order, but at the same time, something seems a bit off. A slightly different humming sound, or a change in pitch-frequency. It doesn’t sound like it usually does, which often warns of problems ahead. 

It’s a well-known fact that the components of a technical installation change their sound and vibration signature before a technical failure. Experienced operating technicians and janitors — maybe you are one yourself — are often able to detect these changes through experience alone.. 

Which is great: Detecting these changes in sound and vibration signature provides the opportunity for preventive measures, planned maintenance and thus lower costs and a better experience for the tenants.

The challenge is that we cannot rely blindly on manually detecting these changes. The expertise required is rare and fleeting, and one cannot be everywhere at all times to catch every change of pitch and vibration.

The result is that building operation is mostly done in a reactive manner, and janitors usually deal with errors in the building’s technical infrastructure after they occur. 

Until now.

This is Soundsensing

Soundsensing is a Norwegian company that specializes in sound and vibration-based automatic predictive anomaly detection in technical facilities in commercial buildings.

Together with major Norwegian property companies such as Thon Eiendom, Malling & Co, Eiendomsspar, AKA and Aspelin Ramm, we have developed a solution to catch errors in  technical installations and infrastructure, before they actually occur. Like a digital, always-present technician's ear.

How Soundsensing works

Soundsensing works in three steps:

1) We install sound and vibration sensors on the technical components: The sound sensor captures the normal soundscape of the room, while the vibration sensors are mounted on key components such as fans, circulation pumps, motors and compressors. The sensors and the system are device agnostic, which means that it works on all manufacturers, all models, new and old, and with or without a Control System.

2) The data is analyzed by Soundsensing using machine learning: The data is sent via wired internet or cellular 4/5G to Soundsensing's cloud. It takes approx. 4 weeks from installation until Soundsensings machine learning algorithms have learned to know the components and technical rooms' varied sound and vibration signatures. The algorithms will then continuously search for unwanted changes and deviations that we know are associated with future errors and failures.

3) The operations service gets control: In the event of a detected upcoming technical failure, the operations service is notified by e-mail, and proactive measures can be taken. The operations service also has an overview of the condition of all its monitored components, in all its technical rooms.

We call it true predictive and data-driven technical supervision. Simple, reliable and proactive.

More on the effects of implementing predictive anomaly detection:

How to get started:

Contact Soundsensing by filling in the form below. We help you with sound and vibration-based predictive deviation detection that will help realize your ambitions for future-oriented, data-driven and proactive maintenance.