曝光台 注意防骗
网曝天猫店富美金盛家居专营店坑蒙拐骗欺诈消费者
Although such a method will permit detecting progressions deviating from normality, the method will not permit fault identi.cation, i.e. diagnosis. To enable this, it is not only su.cient to detect that a progression is abnormal, it is also necessary to determine in which way the progression is abnormal. This can by done by a nonlinear mapping tool, such as a radial basis network.
All conditions has an expected value and an uncertainty for each element in the component state vector. This permits using the component state vector as the input to a radial basis network (Fig. 7.2).
The inside of the radial basis network can be seen as a multidimensional space where the number of dimensions is equal to the number of inputs. In this space, each condition has a region, or cluster, at the position corresponding to the vector of expected values. The size of a region along each dimension is given by the uncertainty along the corresponding metric. For each input vector, the network will identify which region, and consequently which condition, the vector falls within. If a vector falls in the void between the clusters, it’s interpreted as a unidenti.able anomaly.
Like with the threshold method, an instance of this network must be adapted to every component on the rotorcraft. Network calibration can be done either through expert knowledge or automated training. It is however di.cult to obtain the necessary training data and expert knowledge to cover all conditions for every component. This due to the large number of fail-ure modes to which a rotorcraft drive-train is susceptible, and due to the relatively low fault frequency on modern helicopters.
In order to perform a go / no-go decisions, it is however su.cient to be able to detect the normal state condition, for which su.cient training data exist for all components. Any component state vector not corresponding to normality must by de.nition be seen as abnormal. To maintain .ight safety, abnormality detection is largely su.cient as any suspected anomaly will result in a manual inspection of the components in question. If a classi.er is designed to only recognize deviation from normality, a threshold-based classi.er is however preferred. This because the threshold-based approach is signi.cantly less complex than a neural network.
7.3 Performance
The threshold-based method is tested using marked training-sets from AS332 L1 and L2 helicopters. The faulty sets are retrieved from clients’ ground sta-tions, isolating the propagation periods for documented defects. The healthy sets are data batches chosen on random outside the periods containing known defects, each case-number represents a di.erent aircraft.
All the fault cases are loss of torque in the left ancillary gearbox interme-diate gear .xing bolts. This allows the gear to assume a "wobbling" rotation patter, causing damage to its own tooth surface as well as the tooth surfaces of adjacent gears. When the gear rotates in an unbalanced manner, it forms a modulation between the shaft rotation frequency and the tooth meshing
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(64)
