Self-learning algorithms make office buildings greener

In a time when energy prices are reaching record highs and sustainability targets are becoming increasingly stringent, building owners and organizations are looking for solutions. Qien software, formerly known as Cloud Energy Optimizer, can help with that. Niels de Jong, product and innovation manager, explains during the conference Digital Building of the Future more about it.

By: Dimitri Reijerman

With its platform, based on smart algorithms, Qien promises to achieve significant energy savings in existing buildings. De Jong: “We have developed self-learning algorithms that take over the building control in order to optimally manage energy consumption. This allows offices to follow the KPIs of the 'Paris-proof route'.”

De Jong mentions a project at ABN-AMRO as an example. “We show how an office building has performed in recent years. We guarantee a saving of 15 percent on climate control costs, but that is often much more. With the old energy prices there was a payback period of about 5 years, now that is sometimes only 1.5 years or even shorter.”

Qien's software is constantly thinking ahead, says De Jong: "Things like wind direction are also included in the models. The installation has to act differently in a certain wind direction. In traditional building management systems, the heater would be turned up all day, even if the wind comes from a more favorable (more favorable ??) direction later in the day. We calculate that at least 24 hours in advance."

Improving the algorithms that control the customer's building management system is a continuous process, the manager explains. "That's why we use different types of algorithms, such as supervised, unsupervised and reinforcement learning. Supervised is the most important, because you already know some of the data. For example, what time the staff arrives at the office and leaves again. With these models, you work further with so-called energy flows, with which you can control certain zones. Because buildings are quite slow: you have to think of them as an oil tanker. You can want all sorts of things, but that doesn't happen immediately. With our platform, you control the heating in a different, smarter way."

Qien now has more and more customers: “We supply banks, distribution centers, real estate companies and many municipalities and provincial government buildings, among others. Installers also use our platform in old buildings that will eventually get a new installation. This way, they can already see what a building requires in terms of energy.”

Qien is also involved in the Brains4Buildings program and contributes knowledge to it. De Jong: “We see that we are increasingly moving towards predictive maintenance. What you always have to keep in mind is that algorithms are just like teenagers. By that I mean that they experiment. You have to give them access to the right information, otherwise the wrong conclusions are drawn. That is why we also apply multiple layers to our algorithms.”

Would you like to attend this lecture? Register for free for a visit to the conference Digital Building of the Future.