A Condition Based Maintenance Approach to Forecasting B-1 Aircraft Parts
Abstract
United States Air Force (AF) aircraft parts forecasting techniques have remained archaic despite new advancements in data analysis. This approach resulted in a 57 percent accuracy rate in fiscal year 2016 for AF managed items. Those errors combine for $5.5 billion worth of inventory that could have been spent on other critical spare parts. This research effort explores advancements in condition based maintenance (CBM), and its application in the realm of forecasting. Then it evaluates the applicability of CBM forecast methods within current AF data structures. This study found large gaps in data availability that would be necessary in a robust CBM system. The Physics-Based Model was used to demonstrate a CBM like forecasting approach on B-1 spare parts, and forecast error results were compared to AF status quo techniques. Results showed the Physics-Based Model underperformed AF methods overall, however outperformed AF methods when forecasting parts with a smooth or lumpy demand pattern. Finally, it was determined that the Physics-Based Model could reduce forecasting error by 2.46 percent or $12.6 million worth of parts in those categories alone for the B-1 aircraft.
Document Details
- Document Type
- Technical Report
- Publication Date
- Mar 23, 2017
- Accession Number
- AD1051628
Entities
People
- Joshua D. Defrank
Organizations
- Air Force Institute of Technology