7.3. PERFORMANCE
7.4. CONCLUSION
Studying the raw indicator plots, trends are indeed visible. The detection failure is however due to the exceptionally low magnitude for both indicators in this case. Even though trends are clearly visible to a human analyst, the slopes, i.e. augmentation or diminution per hour, is still small due to the abnormally low magnitude of both indicators. A way to circumvent this problem is to not evaluate the slopes directly, but rather to normalize slopes by scatter level g.w. This will in fact mimic the human approach to .nding trends in noisy data; to identify patterns which clearly rise clearly above the random scatter. Using the test data at hand, this is however not possible. Due to the short duration of some of the fault cases, the scatter energy g.w can not be estimated, thus precluding the use of this method.
7.4 Conclusion
This chapter has been more concerned with fault detection than diagnosis and prognosis. The reason for this being that there is little authentic fault data on record, making it di.cult to train and validate a fault recognition system. The only data existing in rich supply is normal state data. Consequently, it is easier to create a system capable of recognizing deviation from normality, and validating this on the few fault case datasets that do exist. From an operational point of view this is largely su.cient, as any suspected damage will in any case result in manual inspection of the aircraft.
Using the feature extraction methods developed in the previous chapter with training data representing a healthy condition of an asset, it is possible to detect most anomalies without aircraft speci.c con.guration of the fault detection algorithm. Further, this method has a better detection rate than traditional threshold based methods. Although testing only a single failure mode is insu.cient to determine if the method has potential for other fault types, the method still show promise. A strong point of the method is that it does not require any sort of con.guration or training by the user. This is a major advantage compared to traditional methods, which are prone to miss-learning by the user, leaving the system with reduced fault detection capabilities.
Chapter 8
Conclusion
8.1 General
Fault detection in helicopter engines, rotors and transmission systems has been an area of intense research for over one and a half decade. Even with this continuous development into HUMS, as well as other areas of accident prevention, rotorcraft remains overrepresented in the accident statistics com-pared to equal size airplane. Consequently, there is still work to be done within the HUMS domain as well as other safety enhancing technologies.
Another aspect of HUMS is maintenance cost reduction. Military opera-tors have already started to adapted maintenance work to bene.t from the fault detection capabilities of HUMS, retiring components "on condition" [23].
中国航空网 www.aero.cn
航空翻译 www.aviation.cn
本文链接地址：OPTIMIZATION OF FAULT DIAGNOSIS IN HELICOPTER HEALTH AND USA(67)
