The U.S. Navy is awash in data on its ships. This data goes back decades for older vessels in the form of maintenance availability reports, drydocking’s and software upgrades, to say nothing of the constant stream of operational reports that include everything from fuel economy to casualty reports on shipboard systems and equipment.
Much of this information remains stored by the Navy for decades across multiple databases and in some instances paper archives. The information is a vast, untapped resource in building a true digital twin for active naval vessels that serves as the first point of departure in planning and executing maintenance on that ship.
Such an accurate digital twin could guide maintenance planner activities and avoid the costly surprises often found when shipyard workers open sections of a ship usually concealed below deck and behind bulkheads. These actions often reveal additional maintenance problems that can add weeks and even months to a ship’s programmed maintenance availability.
The recent Ticonderoga class cruiser modernization fiasco is a good example of incomplete shipboard knowledge vastly increasing the timeline for overhaul. In eight of those 11 proposed cruiser modernizations the updated cost and timeline made retaining the ships impossible, and they were lost to the service earlier than planned.
The Navy must collect and aggregate its vast database of shipboard information, make it available to artificial intelligence to organize and present in visual and data formats to better inform shipboard availability planning. AI can accelerate this process and help the service to understand what truly is going on aboard a 30-plus year-old ship before timelines widen and costs escalate.
The Ticonderoga cruisers
The CG-47 Ticonderoga class cruisers were the U.S. Navy’s first AEGIS combat system platforms, and at their entry to service at the end of the Cold War they represented the apogee of combat system capability.
Designed to defend aircraft carrier battle groups from massed raids of Soviet cruise missiles fired from submarines, air and surface platforms, these first generation AEGIS cruisers tracked hundreds of targets at a time and could engage dozens with surface-to-air missiles, 20mm Close in Weapon System guns and other passive and active systems.
Their radar and computing systems were a quantum leap over existing U.S. Navy air defense warships of the late 1970’s that might engage six to eight targets. The cruisers have fielded every baseline of the AEGIS combat system, making them some of the most versatile warships every employed by the U.S. Navy.
The Ticonderoga’s, however, aged like any other warship from repeated deployments in support of U.S. interests, as well as suffering from deferred and cancelled maintenance caused by continuing resolutions and sequestration during the 2010’s. The combination of age, lots of deployed sea time and deferred maintenance would make upgrading the remaining CG-47’s a challenge. The Navy warned Congress on several occasions that modernizing the cruisers was not worth the funds likely to be expended, but the failure to produce the CG-X class of replacements and shrinking fleet size compelled the legislature to demand cruiser modernization.
Finding more than expected
Sadly, the cruisers turned out to be much more degraded than the Navy or others expected. While the ships’ combat systems could be upgraded, the degraded hulls, mechanical and electrical problems, many not well documented within existing Navy databases, slowed the process and substantially increased the price of the program by 36%. Ultimately, of the 11 cruisers planned for modernization, only seven entered the program, with only three completing the modernization effort.
A recent Government Accountability Office (GAO) report on cruiser modernization noted significant failures in planning for and executing the cruiser modernization. This is very much a “birds eye view” of the problem, but the GAO report did get close to some of the specific causes.
“The Navy did not sufficiently track and, thus, did not fully understand, the condition of the cruisers prior to modernization. NAVSEA 21, Program Office, and OPNAV officials told us that, in hindsight, the cruisers were in worse condition than they realized. Navy officials noted that this was primarily due to the Navy deferring maintenance by cancelling maintenance periods throughout the lives of these ships,” the GAO found.
This is surprising given the vast amounts of data the Navy routinely collects on warships in active service, especially in terms of equipment casualty reports, regular inspections such as the Board of Inspection and Survey reports and routine messages on fuel and spare parts usage. Getting an accurate picture of a ship at any point in its service life, however, is a challenge due to the relative dispersal of this information across multiple Navy databases.
Even a current crew of a ship may be unaware of specific challenges if they occurred before the longest serving crew member reported aboard. While the Navy has taken steps to better document deferred or lost maintenance on its ships, too much of the overall picture of any ship’s material condition resides in disparate locations and databases.
A more accurate assessment
An artificial intelligence with access to all of the Navy’s vast data on its ships and armed with an accurate assessment algorithm might help to gain a better picture of the material condition of a vessel ahead of major maintenance actions such as the cruiser modernization program.
Had the Navy been better informed as to the real challenges and potential costs involved in the cruiser update, it might have been able to push back more effectively with Congress and resist the legislature’s persistent demands to keep the aging Ticonderoga’s in service beyond 35 years. Historic assessments and those made by the Navy when building the CG-47 class clearly stated and expected lifespan of 30 years.
An AI assessment that contains 30 years of casualty reports, maintenance and repair records, accidents and other data might better persuade Congress in regard to a costly modernization program than weak recommendations or silence on the issue; something Congress has accused the Navy of in the past.
Generating AI assessments for ship maintenance is not an easy task, as it is likely that most reports older than 20 years will be paper copies only. Even if digitized, their data points might not align closely enough for common categorization. Relevant data for even one ship is likely scattered across multiple repositories of Navy information.
But the effort and cost to collect and aggregate this data to develop an AI-driven digital twin of every Navy vessel is more than worth the effort. Given that GAO says the Navy wasted $1.84 billion on four cruisers that entered the cruiser modernization program but did not finish and were retired, the usefulness of an AI-driven digital twin of every U.S. navy warship is clear.
