7.2 Classi.cation
Most failure modes for most components are identi.able by .uctuations in the expected value or scatter level in one or more indicators associated with the component. Although di.erent in architecture, both the parametric and the non-parametric feature extraction methods developed in the previous chapter extracts this information. For robust fault detection, it is however also necessary to perform a validation of each indicator step as well as direct threshold testing on certain critical indicators.
Edge occurrences should if possible be correlated with the aircraft main-tenance log to verify that they really originate from equipment replacements. This generates the signal bfault which return the step size for every step not occurring at the same moment as a maintenance action. For steps occurring
115
simultaneously with a maintenance action, bfault remains zero.
Certain components, mainly rotors and engine shafts, have global unbal-ance thresholds which they not under any circumstance must supersede. To verify that the vibration levels for these components are within bounds, their unbalance indicators must be tested directly as a supplement to trend anal-ysis. To avoid false alarms due to noisy data, this testing should however be done after the outlier component s.has been removed.
This produces a total four metrics from each indicator at each point in time; a de-noised version of the indicator itself, .uctuations in scatter level, .uctuations in expected value, and unexpected step occurrences (Fig. 7.1).
A
component
is
however
usually
associated
with
several
indicators.
To
identify the condition of a component it is necessary to evaluate the metrics from all the associated indicators. This provides the component state vector, consisting of i . s., ag.w , ac.and bfault for each indicator associated with the component, and describes the condition of the component at given instance in time.
Using traditional HUMS methodology, a component is diagnosed by com-
7.2. CLASSIFICATION
paring the set of associated indicators to a baseline. This baseline must however be adapted to each aircraft, and is subject to change between main-tenance actions. Using the component state vector, it is possible to use a global baseline which is not subject to change. This because the .uctua-tion metrics ac.and ag.w are less sensitive to aircraft speci.c di.erences in the expected indicator value. Consequently, for indicators with large vari-ation in expected value across aircraft, the baseline pays less attention to the absolute value, i . s., and more attention to the .uctuation parameters ac.and ag.w . Inversely, for indicators with little variation between aircraft, such as shaft unbalance indicators, more signi.cance is given to the indicator absolute value.
A simple fault detection method is to assign thresholds to each element in the component state vector. For the slope metrics, ac.and ag.w , both min and max thresholds should be used. For bfault and i.s, it is su.cient to only apply max thresholds. A set of thresholds must be de.ned for each element in the state vector for each component on the aircraft. Once this is done, it will however be possible to apply the same thresholds to all aircraft of a given type.
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(63)
