AI will facilitate maintenance
By using artificial intelligence (AI) to spot deviating patterns in the district heating network, energy companies can make better decisions regarding preventive maintenance. Developing such a digital tool is the goal of a new research project FVB is part of.
A computer makes analyses and spots deviating patterns much quicker than people. This is what we want to take advantage of in this research project.
Many district heating networks are in need of restoration and renewal, mainly due to age. It is important – both financially and technically – that these measures are taken at the right time. To understand when it is most appropriate to perform preventive maintenance, energy companies are primarily using the information coming from the ongoing supervision of the district heating network. This is based on readings from moisture alarms, the presence of water, humidity, and the wall thickness of the pipes.
“There is a lot of data available about the district heating system, and a computer can make analyses and spot deviating patterns a lot quicker than people can. This is the advantage we hope to be able to benefit from,” says Kristin Åkerlund at FVB, who is participating in the project.
RISE, Energiforsk and seven energy companies are also part of the project. The overall goal of the project is to explore the possibilities of using AI for preventive maintenance in the district heating network and to develop a digital platform for this. The joint platform allows for analysis of the companies’ aggregated data and returns results to the companies. These results can then be used as decision-making information in the companies’ maintenance planning. The platform will be generic, so it can be used by different energy companies. An important part of the project is to establish a “proof of concept” for the platform. This means showing that the platform is technically and functionally feasible. The data that will be used is static information from GIS/NIS, such as information about district heating networks and historical data on errors and, if possible, readings from various sensors in the systems.
“To train the platform, we will initially use data from some of the energy companies who are part of the project. The participating energy companies will also contribute to validating the platform,” says Kristin Åkerlund, who continues:
“If we get the platform to work as intended, there are major benefits for energy companies – in terms of both time and money.”
Kristin Åkerlund’s role in the project is to compile the risk factors that affect the status of the pipes and to develop the requirements specification for the platform along with the participating energy companies.
For more information:
Kristin Åkerlund, 026-14 16 20