Photo: Colourbox

Indoor climate controlled with mobile phones and artificial intelligence

Tuesday 09 Oct 18


Jørn Toftum
DTU Civil Engineering
+45 45 25 40 28

Log in to read the master's thesis 'Implementation of demand-driven HVAC control based on personal comfort modeling' by José Joaquin and Aguilera Prado.

User feedback and control systems based on IoT and artificial intelligence improve indoor air quality and make buildings more flexible.

A study by DTU Civil Engineering shows that intelligent heating systems in buildings can be controlled with the use of data from mobile phones and can improve the indoor climate for 28 per cent of users. The test results confirm that building automation based on personal preferences as input and utilization of internet technology holds huge potential and makes buildings more flexible in terms of being used for new purposes.

“The technology takes into account that users need to adjust the indoor climate on an ongoing basis. This provides a huge potential for quality improvement, one reason being that earlier studies point to user influence as one of the most important factors for user satisfaction. There’s also a number of benefits to the technologies, making the buildings easier to adjust to new tenants and new needs—a request we often get from clients and contractors,” says Jørn Toftum, Professor at DTU Civil Engineering.

In the study, which was carried out in Spring 2018, researchers used artificial intelligence to control a building’s heating system. User feedback on the indoor climate in the building was entered into so-called comfort models on a mobile phone, which was then used for optimizing the building’s temperature. The personal comfort models were combined and used to establish the temperature level in an open-plan office that would cause the least requests for changes in temperature.

Testing took place in three buildings over periods of up to three weeks where the indoor climate was measured. Using their mobiles, users then gave feedback as often as possible, describing how they experienced the indoor climate. After a pilot study at DTU, the experiment was repeated in a department at the engineering firm Cowi.

Most active users found that the indoor climate was better, while a small number had the opposite experience. Simulations showed a rise in energy consumption in the case studied because more users wanted a higher temperature, which, among others, was due to the fact that the study was carried out during the heating season.

“One of the challenges in basing heating systems on user control is that there’s a need for more frequent user input. It’s therefore very important that users receive instructions on how to use the technology. However, the experiment demonstrates that controlling the indoor climate using a combination of wireless sensor networks, user feedback, and artificial intelligence can be a good approach to improving the operation in new buildings,” says Professor Toftum.

The results from the study will be presented at DTU High Tech Summit, which will be held on 10-11 October. The study is part of a theme on how new technologies can be used to improve indoor climate and health.