Our challenge

Reasons for increased energy use can be e.g. malfunctioning control units in the ventilation systems, district heating systems, or that the use of the building has changed. There is also a need to compare buildings to identify and decide on upgrading equipment and renovation measures for reduced energy use and improved function. Large property owners such as municipalities have a large number of buildings with tens of thousands of values that are monitored and can generate alarms. There it is very time-consuming to analyze, sort and manage deviations.

The current situation

We have a lack of systems/products/services that enable optimization of energy use in society. Large property portfolios have an unnecessarily high energy consumption, there are also no good tools to balance the consumption between properties, which leads to the power requirement being higher than it could be with efficient tools.

What do we want to achieve?

The project aims to develop innovations to analyse the energy demand of buildings under different circumstances by using historical data taking into account weather conditions, building information and the use of the building. This data is used to predict future energy needs in buildings, to balance energy needs between buildings, and to reduce total energy consumption.

The project will contribute to new knowledge, solutions, processes, working methods and methods that lead to efficient energy and resource use in buildings, interaction with neighboring sectors, functionality and well-being.

  • Energy optimization

    We optimize energy consumption in large property portfolios by including data sets on construction, analyzing energy usage patterns, activities in the buildings, microclimate, etc. With the help of this combination of data sets that are not included in today’s property controls, more informed decisions about energy measures can be made.

  • Cyber security

    Real estate systems can be attractive targets for attackers, which is why we are also working to build cybersecurity into our systems.

  • Machine learning

    We work with machine learning to automatically recognize what are normal patterns in data from the properties, and what are deviations. To train our models faster, we use transfer learning, i.e. we use already trained models from similar properties and supplement the training data with a smaller amount from the current property.

  • Augmented reality and 3D visualization

    We use augmented reality and 3D visualization to inform energy technicians about the condition of the properties, and suggest measures.

Goal

The overall goal of the project is to, through data collection and the use of machine learning methods/artificial intelligence, reduce the total energy use by up to 15% within large property holdings. This will be achieved through increased understanding and balancing energy based on property operations and use.

Target groups

The project’s main target group is owners of, and companies managing, large building stocks. Primarily, properties in Skellefteå, Piteå and Kristianstad are included, but we will take in needs from, and spread the results to other owners of large building stocks in Sweden.