Content area
Abstract
Approximately 30% of returned radar hardware is classified as “Cannot Duplicate” (CND), indicating a high rate of failure reports that cannot be reproduced. These returns typically result in no repairs being performed, despite extensive testing and investigation. In such cases, a repair depot is unable to replicate the failures reported by field maintenance crews. This can lead to significant waste in terms of time, cost, and resources for a company and a customer. More critically, it disrupts fighter jet operations and reduces overall operational availability. Additionally, CND returns negatively impact the reliability and maintainability (R&M) metrics of radar systems. This praxis presents the development of a machine learning algorithm that utilizes data from the Failure Reporting, Analysis, and Corrective Action System to predict CND cases. The objective is to reduce the CND return rate by 20%, thereby enhancing the R&M of fighter jet radar systems. By minimizing unnecessary hardware replacements, the proposed solution aims to improve operational availability and significantly reduce sustainment costs when implemented at the field or depot level.
