Follow Datanami:
September 2, 2020

Air Force Expands Predictive Maintenance

The U.S. Air Force is expanding its embrace of predictive analytics tools to keep pace with maintenance demands for its huge fleet of fighters, bombers, tankers, transports and helicopters.

There is no shortage of U.S. military aircraft, with estimates ranging as high as 5,400 for the Air Force alone. The problem has been keeping that air armada flying. According to Air Force Times, aircraft readiness as measured as a percentage of planes able to fly has steadily decreased over the past decade.

Hence, the service has been enlisting analytics and AI software companies to help get a handle on maintaining increasingly complex aircraft loaded with electronics gear. Those and other modernization efforts have been spearheaded by the Defense Innovation Unit (DIU), the Silicon Valley-based Pentagon unit established in 2015 to accelerate the transfer of commercial technologies to the military services.

The latest example was announced this week by enterprise software vendor C3.ai, which will provide the Air Force’s Rapid Sustainment Office with an analytics suite to be used for predictive maintenance. The latest contract follows a five-year agreement announced in January 2020 to supply the Defense Department with AI-based predictive maintenance software.

(Separately this week, DIU tapped Google Cloud to develop a prototype AI platform for digital pathology. The research project would deliver augmented reality microscopes to DoD facilities along with access to AI models for medical applications such as predictive cancer diagnoses.)

The latest Air Force contract award to C3.ai focuses on the applying predictive maintenance for the HH-60 Pave Hawk helicopter (shown), the company said this week. Earlier Air Force maintenance efforts covered a range of Air Force planes, including the F-35 Lightning II. “This latest award represents the next stage in scaling C3.ai’s predictive maintenance solution across the defense enterprise,” the AI software vendor said.

C3.ai, based in Redwood City, Calif., claims predictive maintenance tools deployed across the U.S. military services could save as much as $5 billion annually. Those savings would be derived from anticipating parts delays that keep aircraft grounded along with unscheduled maintenance and overall improvements in mission capabilities.

The Air Force’s Rapid Sustainment Office “will be able to accelerate scaling AI and [machine learning] capabilities across the Air Force enterprise, and combine data science with Air Force operational maintenance, to digitally transform how we maintain our global fleet,” said Nathan Parker‚ the office’s deputy program executive officer.

DIU has been steadily tapping commercial AI tools for applications ranging from predictive maintenance and military medicine to cybersecurity and surveillance drones.

DoD’s predictive maintenance effort was launched in 2017, and has since been scaled with the deployment of C3.ai platform across Air Force maintenance depots.

“In advance of failures, supervised machine learning can accurately predict the probability of failure of various subsystems over different time horizons,” DIU said in its annual report. “As a result, maintenance technicians are able to identify component-level failures before they occur, pre-position parts in anticipation of failures, and replace components with a high potential for failure.”

Recent items:

Tamr Helps Air Force Wrangle Data

DoD Looks to Scale Predictive Maintenance

Predictive Maintenance Driving Big Gains in Real World

Datanami