ProRail uses UbiOps to prevent asset failures in their network. They saved months of work using UbiOps for easier deployment of their AI algorithm. The connection of this algorithm to the right data sources went also easier under UbiOps’s management of the data science pipelines.
One of the components actively being monitored are switch heater installations. These are critical to guarantee that track switches keep functioning during periods with cold, ice and snow. Together with UbiOps, a team of asset managers, system specialists and data scientists from ProRail have developed an Artificial Intelligence (AI) algorithm which uses IoT measurements and weather data to detect faulty switch heaters.
Usually, the deployment and management of data science pipelines requires at least the same amount of effort as the actual development of the algorithms. However, with UbiOps the whole system was up and running in a matter of days, saving months of work.
The data science team was able to deploy the algorithms with minimal effort and connect them to the required data sources, without having to care about the cloud infrastructure needed to do so. Newer model versions were easily uploaded and operational once available. Where deployment of models previously took months of time and effort, with UbiOps this process was reduced to hours. The logging and monitoring functionality available in UbiOps ensures that the solution runs as required. The rail network is ready for the winter.
With UbiOps we managed to significantly reduce the time and efforts needed for deploying and managing our AI models for detecting failures of our switch heating installations
Cees-Jan Mas
Projectmanager Vernieuwing at ProRail
© 2023 UbiOps.
