Mission: Enhance ship mission readiness on the emerging autonomous fleet by predicting equipment faults & remaining life in time to matter. Manage & compress sensor data to support actionable analytics.

Solution: Apply AI-SHIP to summarize sensor data while maintaining mission relevant information, predict & localize equipment faults, and explain fault root cause.

Our team developed an AI-based energy analytics for predictive maintenance capability to localize and predict hard fault activity of a power plant and supporting systems on a autonomous ships for NAVSEA. The ship carried over 30,000 sensors across the equipment without a process of aggregating and analyzing the data effectively.

For the Power Plant Activity Summarization and Hard Fault Classification project, we exceeded the stretch goal of 80% hard fault classification with greater than 90% accuracy and 90% confidence across all ship systems. We showed that prediction of hard faults at least three (3) hours in the future is feasible. We also summarized and compressed ship activity more than 1700 times across all systems with historical visualization.