AI-based recognition of heat-leaks for improved thermal retrofitting

Heat leaks and poor insulation are two important contributors to energy waste in buildings, and therefore important targets on the path towards climate neutrality.

Among the different existing techniques to measure and assess insulation efficiency, thermal vision is the only one capable of locating individual deficiencies.

However, inspecting buildings with thermal vision detectors is often a time consuming and expensive technique, as it typically involves an operator to manually scan and evaluate the data of each building, case by case.

In collaboration of Umeå Kommun and under the guidance of academic experts, we are developing a new automated method capable of locating heat leaks of most buildings across large urban areas.

The method is based on an artificial intelligence (AI) that collects and evaluates data from several sensors in a similar way as a person would do. The result is a map with the thermal efficiency and retrofitting potential of most insulation elements in a city.