Applying AI Techniques to Improve Alarm Display Effectiveness

The Alarm Filtering System (AFS) addresses the problem of information overload in a control room during abnormal operations. Since operators can miss vital information during these periods, systems which emphasize important messages are beneficial. AFS uses the artificial intelligence (AI) technique of object-oriented programming to filter and dynamically prioritize alarm messages. When an alarm’s status changes, AFS determines the relative importance of that change according to the current process state. AFS bases that relative importance on relationships the newly changed alarm has with other activated alarms. Evaluations of alarm importance take place without regard to the activation sequence of alarm signals. The United States Department of Energy has applied for a patent on the approach used in this software. The approach was originally developed by EG&G Idaho for a nuclear reactor control room.