Classification Trees and the Analysis of Helicopter Vibration Data
Loading...
Authors
Larson, Harold J.
Subjects
Classification trees, Clustering, Vibration analysis.
Advisors
Date of Issue
1997-09
Date
1997-09
Publisher
Monterey, California. Naval Postgraduate School
Language
en_US
Abstract
Health and Usage Monitoring Systems are receiving a great deal of interest, in an attempt to increase the safety and operational readiness of helicopters, and to minimize maintenance costs. These systems monitor (and can record) various flight parameters, pilot conversations, engine exhaust debris, metallic chip detector levels in the lubrication system, rotor track and balance, as well as vibration levels at selected locations throughout the airframe and the power drive system. Vibration levels are currently being observed on two operational SH-60B helicopters and on an H-60 power drive system installed in the Helicopter Transmission Test Facility (HTTF) at the Naval Air Warfare Center, Trenton, NJ. This paper employs classification trees to analyse vibration signatures produced in the HTTF, identifying those characteristics which distinguish normal signatures from signatures produced by known faulted parts. These trees are quite successful in separating the two types of signatures and achieve small misclassification rates for HTTF data. They are also applied to vibration data collected from an operational aircraft; assuming the tail gearbox in the operational aircraft has no faults, the trees derived from the HTTF produce a high proportion of false alarms
Type
Technical Report
Description
Series/Report No
Department
Operations Research
Organization
Naval Postgraduate School (U.S.)
Identifiers
NPS Report Number
NPS-OR-97-017
Sponsors
Naval Postgraduate School, Monterey, CA
Funding
Naval Postgraduate School, Monterey, CA.
Format
Citation
Distribution Statement
Approved for public release; distribution is unlimited.