Hiberarchy Clustering Fault Diagnosis of Hydraulic Pump

Fault diagnosis is on e of the key technologies of prognostic and health management system (PHM) of aircraft hydraulic system. Aiming at the strong coupling of various fault features of hydraulic pump when multiple faults occur simultaneously, a hiberarchy clustering fault diagnosis strategy was proposed, in which three le vel fault reasoning machine was adopted for five kinds of failures for hydraulic pump. The main idea is to distinguish the obvious failures with individual signal processing first, then figure out the blurry information with data fusion techniques. To the desultorily features, such as increscent clearance of piston/shoe subassembly and off-center of swashplate subassembly, the intensive techniques is uesed to strengthen fault feature under lognitudinal and transverse direction so as to realize the multiple fault diagnosis. Through accumulating the power over +/- 10 Hz of basic axial and severalfold frenquency, the sum of relative power can be exploited to realize the fault feature extraction and fault diagnosis under confused information. Application indicates that hiberarchy clustering method can diagnose the multiple failures correctly of hydraulic pump with high precision even in farraginous condition.