Enhanced Filtering of False Alarm for the 5th Generation Fighter Aircraft using Machine Learning Classification

Prognostics and Health Management (PHM) has become an important capability for the 5th generation fighter aircrafts such as F-35 Lightning II or F-22 Raptor to ensure the safety of flight and reduce the life cycle costs. The on-board PHM System of a 5th generation aircraft uses sensors, software and computing capability to detect component or system faults and generate health reporting codes. Similar to the legacy military aircraft platforms, the PHM system also generated a large number of false alarms, which can reduce the aircraft availability and increases the Operating and Support (O&S;) costs. Studies have been conducted in the past to identify the root causes of false alarms (Ungar 2015) and presented various mitigation strategies including supervised machine learning techniques (de Padua Moreira & Nascimento Jr 2012) to reduce the number of false alarm at the source. However, the goal of the research in this praxis is to train machine-learning classification models using the PHM data and the maintenance records to create an enhanced filtering capability based on the lessons-learned from the operating units. The results of the research have shown that the machine-learning classification model called, RUSBoostedTrees, exceeded the human performance by 4% and 9% in identifying false alarm and faults, respectively, along with the savings of 11 minutes per maintenance event. The results also have shown that the trained algorithm will reduce the diagnostics performance gaps between different operational sites. Lastly, the model can be trained with a fewer number of maintenance events than what it would take human to achieve the equal or better performance in distinguishing false alarms and faults.

Author

• Her, Nathan

Language

• en

Keyword

• Machine Learning
• Aircraft
• PHM
• False Alarm
• Enhanced Filter

Date created

• 2019

Type of Work

• Dissertation

License

• All rights reserved

Rights statement

• In Copyright

GW Unit

• Engineering Management

Degree

• D.Engr.

Advisor

• Grenn, Michael

Committee Member(s)

• Etemadi, Amir
• Malalla, Ebrahim

Persistent URL

•  https://scholarspace.library.gwu.edu/etd/j3860766z