Skip to main content

Freya Systems LLC


Christopher MacNeel, Senior Data Scientist


Freya Systems LLC



Business Challenge

A prominent military aviation platform has been historically challenged by low fleet readiness, with almost half the fleet unable to execute missions due to maintenance and supply difficulties. Along with unscheduled maintenance, part delays and personnel shortages, another significant factor contributing to low readiness rates is Long Term Down (LTD) aircraft, which are aircraft that are unable to fly for more than 60 days. It has proved difficult to identify which aircraft will reach an LTD state. Using historical data, an analytics solution was needed to predict when aircraft would become LTD and forecast their estimated maintenance turnaround time.

Analytics Solution

In late 2018, the aircraft’s manufacturers signed a contract with their military customer to help reduce LTD aircraft. Freya Systems developed two multivariate algorithms to help analysts predict maintenance efforts and increase fleet readiness so the military end customer could obtain better readiness levels. The first algorithm utilizes logistic regression to determine when specific aircraft will become LTD with 82.2% accuracy, helping maintenance operators focus on a smaller subset of aircraft rather than a larger pool of potential LTD candidates. The second algorithm utilizes a random forest to estimate the hours required to bring an aircraft back to a flyable state. Resulting data are shared with a quick response maintenance team and used to prioritize aircraft for efficient resource utilization.


This project enabled our customer to increase mission capability rates by over 8% and lower LTD aircraft by 40%. In 2019, the team was credited with returning 26 LTD aircraft back to a flight ready status. Through the use of predictive analytics and strong customer engagement, the fleet experienced increased readiness, allowing more missions to be completed and our client to exceed their customer’s contractual expectations.