Noova Energy Control

How do you keep control?

Article written in collaboration between Tommy Hagenes at Energy Control/Airthings/Proptech Bergen and Carsten Lehbrink at qbee AS

Many people know the feeling of having too many balls in the air. What if you had 31 billion devices in the cloud? How do you keep control?

During 2020 there are 31 billion devices, or IoT devices (internet of things), according to Leftronic.com. Every single second this grows by 127 new devices, and by 2025 it is estimated that over 75 billion devices will be connected to the internet.

Data-driven buildings

This requires that control over data must be systematic. For commercial buildings in particular, we need to start thinking about how to handle the fact that everything in the building is a data unit. We have previously talked about the future, but the future is here. Already 90% of cars will be online within the next year, and 80% of industrial production uses or will actively use IoT in the same period.

But a successful IoT project requires that a whole technological and commercial value chain works together. The exciting and challenging aspect of somewhat larger IoT projects for buildings is that they usually don't start life as new builds, as only 1% of buildings are new builds each year.

Often you therefore have to deal with many different suppliers and already existing technology that would be too expensive to replace. Here we are talking about technology that can be more than ten years old – just as relevant for smart buildings as for industry 4.0.

The art is to design a flexible technological architecture that makes it possible to connect existing traditional systems, such as ventilation systems, heating systems, etc., and get these to work together with new available technology, so that you achieve effective data capture from both new and existing devices.

You also need secure and reliable communication to a database that can collect and organise data. But it doesn't stop there. With so many data points being retrieved every minute or second today, you need smart systems that can visualise and analyse data.

Now we're talking about several thousand data points in a building, and at that scale humans simply cannot see what machines can. And this is usually where costs arise. If the system is designed correctly, value can come out instead of just raw data. This can happen through threshold values and ordinary statistical analysis, or through machine learning and artificial intelligence.

How to go from insight to control

The next step is to illuminate the system in the other direction. Now I have control information and this must be fed back into the system to change set points or achieve some form of control. Perhaps only when data analysis can be used to control systems do you achieve the value that gives a project a good ROI (payback period).

Before each project it is important to understand some basic things:

Where should the control be located? Can it happen in the cloud or should it be local? Or perhaps a hybrid solution? What happens if the network goes down?
Who will need access to the system? How do customers use the system and who is responsible for operations?
Machine learning and artificial intelligence have come a very long way and a lot can be done with this technology. Nevertheless, it is experienced that much can still be solved with threshold values, hysteresis and regular statistics. The result is often better since more people can understand the control logic and ask critical questions about it. Advanced machine learning is often not necessary in these projects. It's a bit like using a sledgehammer to crack a nut.
Many large IT projects fail because a large monolithic solution is built. In the IoT world, the art is to put together many smaller systems. Especially if there are already some "legacy devices". But then you need to find good technology to be able to control, monitor and update all the building blocks in a simple, reliable and cost-effective way.
Traditionally, building control systems have been purchased as a one-off cost. That is, you make an investment and then the system is expected to work for the next 12 to 20 years. Since modern systems are much more software-based and must be operated both in the cloud and locally, suppliers often need a model where monthly operating costs are calculated, typically shown as a subscription solution. This must change in order to succeed with data-driven buildings.

Summary

The future is closer than ever. We already have enormous amounts of data in buildings, and there will be very exponential growth going forward. The importance of thinking about "fleet management" – where all devices in the building must update themselves – is something all organisations must consider.

Just imagine you install a data unit in a building that gets a security vulnerability, or has a software bug that needs to be fixed. What if that unit exists in 1,000 buildings? How do you proceed?

The IoT challenge that must be solved is to gain control – we must have fleet management from day one.

Experiences from Proptech Bergen with Matterport 3D scanning

‍