A major division of a Fortune 100 company operating hundreds of turbines around the world began experiencing an average of two unplanned shutdowns per day. They wanted to reduce unexpected interruptions and mitigate production loss by anticipating and avoiding potential shutdowns.
Using the Maana Knowledge Platform, the company navigated volumes of controller alarm data being created by 350 generators. As shown in Figure 5, this data was then combined with turbine specification data and turbine trip data to build a model that predicts imminent trips and alerts staff using alarms. The model draws correlations between historical patterns in support calls and trip events. Now the company can detect unwanted events that could lead to a potential turbine shutdown within 24 hours, gain insights into trip events and ultimately reduce the frequency of unexpected shutdowns.