Danes are currently experiencing skyrocketing energy prices – and although many homeowners install heat pumps as a replacement for a wood or a gas furnace, preliminary results from the Department of Computer Science at Aalborg University show that there is also money to be saved simply by adjusting the control of the pump.
A group of researchers have developed an intelligent control system for single-family homes that saves up to 30 percent of energy costs compared to a traditional heat pump setup.
Assistant professor Peter Gjøl Jensen from the Department of Computer Science is one of the main forces behind the research project. He explains that a classic heat pump works like a refrigerator, where the compressor switches on and off depending on the temperature in the house. But the many interruptions require a lot of power, and factors like whether the energy price is high or low is not taken into account.
An accumulation tank increases the reduction considerably
In collaboration with his colleagues, Peter Gjøl Jensen has done just that, i.e., incorporated electricity spot prices and weather data in their work on models that have been continuously tested and adjusted to the conditions in a virtual family home.
- With this kind of models, we can control how long and when the heat pump has to run to obtain, for example, a temperature of 22 degrees in the house when you get up in the morning. It is a question of starting and stopping the pump at the right time, meaning when it is windy, and the wind turbines deliver cheap energy. Just as significantly, our results show that we can obtain exactly the same experience of comfort for the residents, says Peter Gjøl Jensen.
He adds that if you also install an accumulation tank, the reduction is expected to improve even more.
- It is especially the combination of the accumulation tank and the smart control that makes the difference; we can heat up the accumulation tank at night and "save" the cheap night power for the daytime. It requires artificial intelligence to take over control and find an optimal strategy for switching the heat pump on and off.
Cheapest is often also greenest
It is not only the home owners' wallets that benefit from intelligent heating management. From a climate perspective, it is also far better that we use electricity when it is cheapest, which is also typically the time when it is greenest. Peter Gjøl Jensen explains:
-Saving electricity for later use often results in an energy loss. We cannot avoid that. But even if we use a little more energy overall, it will both be much cheaper and have a lower CO2 emission, because we do it at the right times. We actually also observe cases where we can make it all more energy efficient. This requires that we can get the pump to run when it is most efficient – and that depends, among other things, on outside temperatures.
Must be tested in practice
As already mentioned, the models have been tested in a virtual family home, and the researchers are now looking forward to start testing the controls in practice. According to Peter Gjøl Jensen, the preliminary calculations indicate that the models correspond quite precisely to reality.
A decisive factor is that the control becomes so easy to install that the ordinary home owner does not have to set aside large sums of money for the installation. Therefore, the researchers are working on a simple solution, where you buy and attach a box that communicates with and controls the pump. In addition, you set up an inexpensive temperature sensor in every room in the house
- It is not realistic that all home owners, for example, replace their old thermostats to get an active heat pump control. So, the big question is how much we can actually reduce energy costs just by changing the control in a house with old-fashioned mechanical thermostats. There are a lot of challenges in that. But we hope and believe that we can achieve somewhere between 10 and 30 percent reduction in costs by simply moving around when we send heat into the house. We will now demonstrate this in practice.
This is how the researchers did it:
Over the past 25 years, researchers at the Department of Computer Science at Aalborg University have developed the software tool UPPAAL, which can be used to improve, validate and construct systems that we can depend on - for example in connection with traffic management, heating regulation and rainwater management.
They combine mathematical models with machine learning and artificial intelligence. The tool uses digital models of the real world (so-called digital twins) to learn optimal control, taking, for instance, uncertain weather conditions into account.
These results are published in “International Symposium on Theoretical Aspects of Software Engineering”.
The research team:
Imran Riaz Hasrat, Peter Gjøl Jensen, Kim Guldstrand Larsen & Jiří Srba.