Innovative inspection of district heating networks with artificial intelligence

This project involves the inspection of around 130 district heating networks with a total length of 800 km. The inspection of these networks is time-consuming and complex due to the infrared thermography used and the associated requirements. A sufficiently cold ambient temperature, which is only reached on a few days a year, is required for sufficient contrast in the IR spectroscopy evaluated by humans. In addition, the inspection is currently carried out using an IR camera mounted on a box truck, which has to be operated in parallel with a laptop. Due to the circumstances mentioned, an inspection of the entire district heating network is only possible over a longer period of time.

The aim of our project at aku is to develop, design and test an innovative solution that largely automates the current manual process, thereby improving it in terms of usability, accuracy and speed. A thermographic camera installed on a vehicle is used to scan the entire pipe network. The inspection is recorded and made available as a video file. Noticeable areas can be analyzed in more detail by the AI and recorded as separate file formats and attached to a report as image documentation.

Our solution with the aku.visionManager® and the neural evaluation of image data significantly simplifies and accelerates the inspection of district heating networks. Automated image evaluation using artificial intelligence enables anomalies to be identified quickly and precisely, reducing inspection times and increasing accuracy. The integration of the camera and machine vision software into the aku.visionManager® eliminates the need for additional software conversions. Image acquisition parameters can be set directly, overlaid with geographical data and damaged areas can be localized. This makes the entire inspection process more efficient and user-friendly, which ultimately leads to considerable cost savings and better use of resources.