U.S. Navy Gets Its First Warship Equipped With AI For System Monitoring

The U.S. Navy destroyer USS Fitzgerald (DDG-62) has become the first warship to deploy Enterprise Remote Monitoring Version 4 (ERM v4), an artificial intelligence (AI) powered system designed to predict maintenance issues and improve fleet readiness.

The system, developed by Austin-based company Fathom5, helps crews identify equipment failures before they occur, reducing maintenance disruptions and keeping more ships mission-ready.

ERM v4 is a part of the Pentagon’s Condition Based Maintenance Plus initiative, which aims to use machine learning to enhance maintenance planning for ship crews, shore commands, and logistical units.

The system is expected to increase the Navy’s combat readiness, ensuring more ships are available for deployment in case of a large-scale conflict.

According to Zac Staples, a retired Navy officer and CEO of Fathom5, the Navy currently operates in three cycles: one-third of ships are deployed, one-third undergo major depot-level maintenance, and the remaining third are in varying stages of readiness.

He stated that AI-driven maintenance could help keep more ships prepared for deployment.

ERM v4 replaces the Integrated Condition Assessment System (ICAS), which has been in use since the 1990s. On the USS Fitzgerald, the system analyzes 10,000 sensor readings per second from the ship’s hull, mechanical, and electrical (HME) systems.

The AI then makes maintenance recommendations, which are directly integrated into the ship’s maintenance planning system.

During its deployment, ERM v4 identified a “long lead item” (a critical part that takes a long time to replace) nearing failure. The system alerted the crew in advance, allowing them to order the part and pick it up at the pier- preventing a system failure that could have put the ship out of commission.

Staples did not disclose the specific part, deferring to Naval Sea Systems Command (NAVSEA), which did not respond.

Staples stated that traditional ship maintenance often follows a “run-to-failure” approach, where components are used until they break before being repaired or replaced. He compared the transition to modern car oil changes, where vehicles now track mileage, oil viscosity, and temperature to optimise maintenance timing instead of following a fixed schedule.

The Navy has discussed the need for a cultural shift in its maintenance approach. In 2023, Mathias Haegele, a mechanical engineer at the Naval Surface Warfare Center, Philadelphia Division, acknowledged the challenges of adopting AI-driven maintenance, stating that new technologies require engagement, adoption, and continuous feedback from end users.

ERM v4 also modernises how sailors record system data. Traditionally, sailors manually recorded readings on clipboards but the system introduced digital log-keeping using smartphone-like devices, allowing sailors to enter data more quickly and accurately.

Staples added that sailors adapted well to the change since text-based input is faster and more efficient than handwriting.

ERM v4’s AI recommendations are verified by shipboard maintenance leadership before being implemented. Feedback from sailors is then used to refine the system’s algorithms. The AI is updated four times a year, following a continuous learning model where data is collected, analysed, and redeployed in quick cycles.

By comparison, the Submarine Warfare Federated Tactical System (SWFTS), considered a gold standard for continuous updates, operates on a two-year cycle.

Currently, ERM v4 is focused on engineering maintenance, but Staples says that it could eventually be expanded to combat systems, though this would be a more complex process.

In 2024, the system will be integrated into the Naval Maintenance Repair and Overhaul (NMRO) system, making it the predictive maintenance AI layer for the Navy’s larger logistics and IT portfolio.

In 2025, ERM v4 will be deployed four more times, collecting more data to refine its algorithms. The system is expected to expand to one or two more ships this year, possibly including an amphibious transport dock, with plans to scale up to a dozen or more ships per year starting in 2026.

The Navy has long sought to use data-driven maintenance to reduce downtime and improve readiness. A previous ERM system was tested in 2019, with Rear Adm. Lorin Selby, then NAVSEA’s chief engineer.

However, earlier Condition-Based Maintenance (CBM) efforts faced challenges. One previous approach delayed major refurbishments until necessary, but this led to uncertainty in shipyard schedules, causing longer and costlier maintenance periods.

This time, the Navy sees AI-powered CBM as a way to address smaller maintenance issues before they become major problems.

While some machine learning capabilities already exist on Navy ships, such as in the Aegis Combat System, ERM v4 is the first AI system built specifically for predictive maintenance.

The Navy hopes AI will not only help detect part failures but also improve logistics by pre-positioning replacement components based on predictive data.

The U.S. military has been increasingly integrating AI into maintenance and logistics. The F-35 Joint Strike Fighter program already uses real-time health monitoring and predictive maintenance, and the commercial sector has also embraced AI for optimising vehicle and mechanical system upkeep.

Reference: The War Zone

Source: Maritime Shipping News