Most federal agencies face challenges when it comes to data and technology adoption. These are compounded for the FAA because of the sheer size and scope of its operations: radar centers span the country to track the planes in the air. The FAA wanted to quickly and efficiently identify the root cause of radar and system failures. In addition, there was need to increase the level of understanding of the cost of repairs or the drivers behind the increasing cost of repairs. Other concerns included the intense pressure on budgets, rising inflation, and equipment modernization and service life extension needs.
Repeatable predictive methodology
These challenges, coupled with the COVID-19 pandemic, forced the FAA to reassess the way maintenance is performed and to use data to help weigh different investment options and better define financial priorities. It was clear that the FAA needed to apply data to answer critical business questions, such as which preventive maintenance interventions are most critical for the safety of the National Airspace System (NAS).
As organizational leaders wrestled with these different issues and the challenge of weighing diverse investment options, the FAA turned to the EY organization. FAA had lots of data; in fact, leadership frequently remarked that “they are data rich and insight poor.” But, how do could FAA access data that is spread across vast systems? Our role was to help them source, process and integrate millions of records from multiple types of different data sets.
Unstructured data was translated into structured data that could be used for analysis by employing natural language processing (NLP) to find commonalities in data and identify parts in maintenance logs. A pipeline was created to push data through and allow the FAA to utilize sensors to monitor issues and know where to send technicians to address them. These efforts created a reduction in maintenance hours and an enhanced planned maintenance schedule through remote monitoring activity on a broader scale.
By observing radar data in remote locations, for example, the FAA is able to save maintenance time, as technicians don’t need to drive to these locations. And, the technician can monitor system performance without relying on manual logs. This early detection approach can reduce operational risk and make it possible to proactively respond to potential failure and/or alarms.