The Modernization of the Military: Machine Learning and Predictive Maintenance
This is Part 2 of our 3-part series on the shift from preventative maintenance to predictive maintenance in the military. For a general overview of the two concepts and the difference between them, read Part 1 here.
Artificial intelligence (AI) has many different military applications — from cybersecurity to combat training. It provides the ability to intake, process, categorize, and derive insights from copious amounts of data in short amounts of time. This has significant uses for everything from identifying a T-90 main battle tank in a satellite image, identifying high-value targets in a crowd using facial recognition, translating text for open-source intelligence, and generating text for use in information operations.
Recently the Department of Defense has made it clear that advancing reliance on AI and machine learning is imperative to national success. In 2019, on the heels of an Executive Order, DoD released its new AI strategy, prioritizing research in that field. DoD established a new Joint Artificial Intelligence Center, announced a $2 billion program to develop new AI technologies, and launched a collaboration with leading robotics and autonomous technology experts. But even before making these enterprise moves, the military has begun relying on machine learning and AI in its everyday operations, particularly in the way it maintains its complex equipment.
How the Military is Embracing Predictive Maintenance
Avoiding unexpected equipment outages or failures is especially important in the military context, where the consequences could be disastrous. With a move from planned preventative maintenance to AI-driven predictive maintenance, plans are no longer one-size-fits-all. Instead, maintenance is scheduled and undertaken based on each piece of equipment’s unique characteristics, usage, and “symptoms.”
While its adoption is not yet widespread, predictive maintenance technologies have already begun to show promise in realizing the full lifespan of complex machinery and reducing costs associated with excess maintenance. Here are three examples.
Utilizing AI to Repair Army Vehicles
Modernizing military maintenance hinges on access to mass amounts of data. To help, the Army hired Uptake Technologies to implement AI-driven predictive maintenance to increase operational availability and reduce costly breakdowns on troop-carrying vehicles. By leveraging more than 2.1 billion hours of operational machine data, Uptake’s AI Asset IO software identified the Bradley Fighting Vehicles signatures of failure weeks before an actual failure event, and at a significantly higher accuracy than previous methods. As the Army continued to use the technology, the software increased operational availability, improved logistics, reduced maintenance costs, and improved safety, all with a level of precision that increased with time.
AI-Driven Predictive Maintenance for the Army’s Stryker Fleet
In an attempt to upgrade their systems, the Army’s Logistics Support Activity (LOGSA) contracted IBM’s Watson to develop tailored maintenance schedules for its Stryker fleet, predicting failures of major components like transmissions, engines, and steering systems before they actually happened. Sensors were installed on 350 Stryker vehicles, and Watson analyzed maintenance manuals and work orders to create a comprehensive maintenance picture. With that information, the system was able to flag anomalies and predict when components in the vehicles were likely to fail. As a result, the Army could set up maintenance for individual vehicles rather than sending them in groups for scheduled maintenance — almost like a personal treatment plan. This predictive approach to vehicle maintenance kept more military vehicles operational and saved manufacturers and agencies time and money because they would know what parts to keep in inventory for likely repairs.
Machine Learning Models and the Military Sealift Command
When the Military Sealift Command (MSC) needed to move from a preventive, condition monitoring based maintenance approach, Abeyon created a machine learning system to transition them to a more proactive, reliability-based maintenance approach. In this case, full visibility of all machine data was critical. In order to fill information gaps in maintenance and repair data, Abeyon’s team built Clarifi to automatically analyze unstructured data and relationships between equipment data, which allowed them to identify potential weak points. With invaluable information from Clarifi, MSC engineers could continue to pursue new and innovative ways to increase efficiency, reduce costs, and deliver excellent service.
The Long-Term Potential of Predictive Maintenance
It is clear that traditional maintenance in the military is evolving to become more dynamic and flexible. There is no longer a universal approach for all tools and equipment, as assessing specific conditions, variations in usage and operating environment, and even damage are being taken into account. The total impact of this move is only starting to be realized. AI-driven predictive maintenance can prevent disastrous machine malfunctions, saving not only money but also potentially lives.
With an intense focus on preserving and protecting our equipment, it can be easy to neglect the maintenance of the military’s most valued asset — the soldier. AI and machine learning can be leveraged to improve both physical and mental combat readiness.
Find out what predictive maintenance for health and wellness in the military could look like in Part 3 of the Predictive Maintenance series.