Press roll barring: online tracking and systematic cause analysis

As part of a student project, a pilot project was launched at Munksjö Dettingen 2022 to monitor rolling bearings using an IoT sensor platform. Initially, 26 sensor units were installed and commissioned on the two paper machines located at the plant.

The aim of the project was to expand the maintenance process with a predictive component. The installed sensor units, which were specifically developed for retrofitting existing machines, operate wirelessly. Thanks to the built-in energy harvesting technologies, the sensor units operate energy-autonomously with the help of the existing machine heat. The sensor unit measures acceleration, vibration and temperature data at the monitoring point.

In contrast to the classic methods of vibration analysis, the recorded sensor data is pre-processed directly in the sensor unit and reduced to a few characteristics, which are transmitted to a gateway via low-power radio. In a learning process, the artificial intelligence (AI) there creates the operating states of the monitoring point that can be expected and are classified as non-critical as a so-called indicator value. In the further course of the monitoring, this value is compared with the current operating status and thus ensures the detection of changes in it. The measured values and the indicator value are visualised in a user-friendly way in a web portal.

Cross stripes formed on the surface of a press roll at the mill location in Dettingen over a period of 2-3 months. This phenomenon, also known as barring, makes it necessary to change the roll well before the desired service life in order to avoid quality losses in the product. By monitoring the vibration of the roller bearings of the press roller with the help of the IoT sensor platform, it was possible for the first time to visualise a trend in the development of the vibrations. A previously conducted, classic vibration analysis was able to accurately describe the defect pattern as a point recording, but did not allow any conclusions to be drawn about the formation process of the transverse stripes. The trend displayed by the wireless sensor units showed different rapid changes in the vibrations on the press roll in different time periods. These time periods correlated with the production of different paper grades. Monitoring the press roll using the IoT sensor platform provides important information for investigating the press parameters during the production of a specific paper grade and their influence on the formation of barring at this position.