Author profile picture

How can we keep our gas consumption as low as possible? This is a question that has preoccupied the Netherlands since the start of the war in Ukraine. Besides the obvious ways, such as insulating homes, there are also innovative, data-driven solutions: a self-learning office building that takes the outdoor temperature into account, or an AI-controlled hybrid heat pump that only switches on the boiler when it is really needed.

The Netherlands wants to get rid of Russian gas. But this is easier said than done. We import about eleven billion euros worth of oil from Russia to the Port of Rotterdam every year. Innovative techniques to cut down on gas consumption are therefore more than welcome. The data-driven solutions offered by the research organisation TNO and start-up Quatt have the potential to be part of the solution.

Office building improves energy conservation

For twenty years, TNO has been working on building models that provide insight into energy consumption: insights that come in very handy today. In 2017, TNO developed a so-called self-learning office building in collaboration with the energy company Eneco. A building that provides insight and recommendations on energy consumption based on a mathematical model and artificial intelligence. The model works with sensors and data from the building and collects daily data on the energy performance of the systems in the building as well as the weather conditions. In the long run, the buildings that use the system will not only be energy-neutral, but will even generate energy.

“Data-driven buildings can accomplish a huge reduction in energy consumption,” says Huub Keizers, MT member of the BTIC and coordinator of the Working Group on the Built Environment within the Dutch AI Coalition. “We have shown that the system can save between 20 and 25 per cent on energy, among other things, by not cooling down when heating is in progress and by taking weather conditions into account.” The AI model was tested in actual practice in a number of office buildings, including at the headquarters of Eneco. “Over the years, the system gets smarter and smarter, enabling you to save more and more energy.”

The project has in the meantime been completed. Last year, TU Delft started a new project in collaboration with TNO and others: Brains for Buildings. “It is a fine successor. Together with forty partners, we will use models and methods to guarantee lower energy consumption.” One of the themes revolves around achieving energy positivity. The project is also concerned with the integration of solar cells into building elements such as walls, bricks, paving, coatings, roofing and acoustic screens.


The beauty of these data-driven solutions is that they can also make predictions about the future when an office or home is modified, ideal for times when gas consumption needs to be monitored, Keizers goes on to say. “What will the impact of new glazing be? And what will happen to the heating costs if you install insulation? Models can help to make the best choice based on personal circumstances. For example, you can gain insight into your energy consumption when the family grows, and then adapt accordingly.”

Over the coming years, innovative, data-driven solutions need to be made more widely known and applied in society as soon as possible, Keizers contends. “Unfortunately, the price of gas is sky-high. But what you see at the same time is that the payback period of innovative solutions is actually shortened in times of crisis. The costs of these will slowly come down.”

A smart hybrid pump

For many Dutch people, it is not yet realistic to radically switch off gas in the near future. Investment costs are simply too high. It costs an average of 4,000 to 20,000 euros to convert a house into a gas-free home. Quatt thinks it has the solution. The hybrid heat pump that is based on artificial intelligence offers an affordable solution for reducing gas consumption, with a price of 1,599 euros (after deduction of the subsidy). In the hybrid system, a pump works together with an existing central heating boiler. Based on an AI-controlled system, the boiler is only activated when absolutely necessary.

“Before the war broke out, we had already developed the whole concept, but we have definitely stepped it up a gear recently,” says Bas Flipse, COO & co-founder of the company. Quatt – the start-up has only been in existence for eight months – has in the past few months already received more than a thousand pre-orders.

Self-learning software

Hybrid heat pumps have been available on the market for a long time already, but what is innovative about the AI-controlled pump is the precision with which it works. Over time, this leads to a significant reduction in gas consumption. “With a regular hybrid pump, the mechanic sets the time when the heat pump should switch on and when the boiler should do its work. This is done on the basis of experience and ‘fingerspitzengefühl’,” Flipse explains. “Our pump, on the other hand, has self-learning software and can regulate itself predictively. The algorithms calculate the cheapest way of heating, but you can also set it to use as little gas as possible. The model takes the weather forecast into account and learns how quickly the home will cool down, for example when the sun disappears or the heat pump is switched off.”

Those who buy the hybrid pump but still want to completely dispense with gas are in for a treat: “We work with modules. If you also purchase the second module, which is still under development, you will have enough power and the right level of water temperature to replace your boiler in its entirety.”

Making a difference

In October, the first pumps will be installed in the homes of consumers. “Then we can really get stuck into working on the energy transition. The heat pump market is pretty traditional and not very innovative. I’m thrilled that we can already make a difference with our pumps in the short term.”