It is no secret that FM tasks and operational routines are characterised by a great deal of routine work. The routine is the result of a risk assessment. We choose to check the ventilation system every single morning because the consequences of it not starting can quickly be critical.
But if the building has 9 ventilation systems located in various places — where you often have to disturb tenants by walking through the premises, and where each round takes on average 10 minutes per system including travel time between units — does this justify 90 minutes? 90 minutes normally equates to NOK 900 if you have a contracted building operator taking care of this. Calculated on an annual basis, this quickly adds up: 5 days a week over 52 weeks a year comes to an extraordinary NOK 234,000 (!!)
For NOK 234,000 you can take quite a number of measures, but it is clear that we cannot simply cut it out entirely. However, given that we use the BMS correctly, or that we use sensors on the building, we can often reduce the round to once a week, or once a month.
In larger buildings over 4,000 sq m there is typically a BMS. This can vary in level, size and complexity, and for the BMS to be used as a solution towards data-driven operations it must support API integration. Otherwise you fall into the "SMS trap". An SMS is dumb. It works fine at the time, but no one knows whether the issue was fixed, we keep no statistics on SMS messages, and the SMS does not communicate with the FM system. That is one reason why you cannot use this in a risk assessment to revise the maintenance need. If you are on holiday or sick, the SMS generally does not go to the person covering the building. And if the message does not immediately reach the right person, it is critical for tenants and their productivity.
If you do not have a BMS, or if the data from the BMS is not good enough, you can easily supplement with sensors. This can be a very profitable investment, as sensors are often priced as a monthly fee or "as a service". This means you avoid the capital investment costs and can instead invest it as a common cost, using the savings from that cost.
Let us look at two simple sensors that can help us verify that the ventilation system is running.
VOC stands for Volatile Organic Compounds. Large correlations have been observed in buildings where VOC rises when the ventilation system stops, and you can therefore use simple machine learning algorithms, or just simple rules with threshold values.
Here you can see that — with the exception of a lunch meeting that took place on Wednesday at 13:00 — it correlates very well with the system's operating status indicated by the dotted black squares. At Proptech Bergen, the VOC sensor is used as one of many parameters that indicate whether the ventilation system needs to be checked.
VOC can of course also come from people, food, perfume and similar, so this must be used consciously — and this is where data and machine learning come in. One can assume that this is a working meeting with food based on the time of day and CO2 level, along with other parameters indicating the ventilation system is running. Another example of this is decibel meters.
The same type of correlation has been observed with decibel meters, as background ventilation produces a certain noise level. It is important to place the sensor somewhere that does not constantly have noise, in order to detect faults related to the ventilation system stopping, as opposed to common areas that constantly have noise.
There are of course many more ways to do this, but in this way you have simple and affordable alternatives for verifying that something is running. It is essential to be confident in the data in order to be able to defer traditional routines.
The next step is that there must be communication between sensors and people. What we see today is a jungle of sensors and more and more web interfaces or apps that we humans have to relate to. For real estate technology, I believe the sensors must speak to the tool that operations uses — in many cases, that is the FM system.
I previously published a post where we announced that Proptech Bergen became the first building to go from calendar-based to demand-based control with a fully integrated FM system, linked to below.
https://www.linkedin.com/feed/update/urn:li:activity:6539518680789725184/
A key to succeeding with data-driven operations is, of course, using the data you get. I believe that if we can simplify rather than create more systems, the threshold for getting colleagues on board with the digitalisation wave is lower. Start simply with a few parameters, learn what the data means, and use it in existing systems.
A final tip: it is a jungle of sensors out there and there are many pitfalls. Use recognised suppliers to ensure you have confidence in battery capacity, coverage and lifespan.
Fill in some info and we’ll get in touch by email or phone.
You can also call us on +47 516 10 170
Fyll ut litt info, så tar vi kontakt på epost
eller telefon.
Du kan også ringe oss på telefon
+47 516 10 170
