inSIGHT: remote health industrial services in high-dimensional data

Important results from the project

InSIGHT will develop and test a health check solution for cold rolling mills that combines visualization and artificial intelligence to be able to give early warnings in the event of problems and propose measures. The goal was to develop, test and validate use cases to better understand the rolling process and to be able to more quickly identify and solve problems that have arisen. Data analysis did not show clear connections and the process could not be described using algorithms. Probable cause is that the system around the stressometer is more complex than available data could describe.

Expected long term effects

The desired result was to enable digital interaction between supplier and customer via remote support to decrease the need for travel, and reducing production downtime. This would have strengthened the value chain through new services and increased product quality. The goal of the project was to develop, test and validate use cases and create a rolling mill health check solution. Since the data analysis was unsuccessful, no ML/AI algorithms could be developed. Without algorithms it was no use to continue to work on the health check solution, without changing the project goals.

Approach and implementation

Project management: During the project, monthly project and steering group meetings have been held. User Experience: Kick-off and site visits have been carried out and interviews with staff have been conducted to understand the visualisation requirements. Data collection and preparation: Data has been collected from one of Gränges rolling mills in Finspång and data has been prepared and analyzed. Expected connections and descriptions of errors could not be found, which is why subsequent steps could not be carried out.