By using the Caverion Intelligence AI solution we have been able to detect and mitigate several deviations long before they escalated into real failures.
Sune Jonsson, Manager System and Plant Development, OKG AB
The Swedish nuclear power company OKG wanted to utilise data more efficiently in order to guarantee production reliability at the Oskarshamn nuclear power plant. By utilising the data, we aimed to achieve a timelier understanding of the facility's critical objects and a wider visibility into the status of non-critical devices.
OKG’s goal was to find a modern tool for utilising the data collected from the power plant. Detecting faults early prevents downtimes and makes it possible for staff to plan maintenance work effectively.
All services at the power plant must comply with strict safety and information security regulations to guarantee the safe operation of the plant.
The Caverion Intelligence Anomaly Detection service uses machine learning and the plant's existing process data to monitor the plant's operations. The service understands different operating modes and the connections between different measurements, and finds anomalies in the data to predict disturbances.
The Anomaly Detection service includes regular status meetings with our Caverion experts. To ensure data security, users outside the nuclear power plant don’t have access to the Anomaly Detection system, but we update the analytics and data models together as needed.
The service’s web-based user interface is designed to be very user-friendly so that the staff at the power plant can utilise the data as effortlessly as possible in their daily operations. Data from the system is examined and interpreted together by the plant's various experts, which increases the amount of shared knowledge about the plant's operations.
The nuclear power plant has very comprehensive safety systems in place. Thanks to Caverion Intelligence Anomaly Detection, disturbances can be identified at an early stage, which gives more reaction time to correct disturbances before the set limit values are reached.