Parameterized local characteristic scale decomposition and its application in composite fault diagnosis

ZHU Wen-min1, YU Yu2, XUE Hai-feng3
(1.The First Construction Division Co., Ltd. of China Railway, Chongqing 401121,China; 2.Changjiu Intercity
Railway Co., Ltd., Nanchang 330000,China; 3.The State Key Laboratory of Mechanical Transimission,
Chongqing University, Chongqing 400030, China)

Abstract: In order to eliminate the discontinuity of the first derivative in local characteristic scale decomposition (LCD) method, the solution of Hermite interpolation combined with adjustable parameters was designed. A parameterized Hermite interpolation method was constructed, which could be further approximated by adjusting the adjustable parameters. A parameterized local characteristic scale decomposition method was proposed, the core of this method was to obtain fitting curve by parameterized Hermite interpolation. The more ideal component signal was got by adjusting adjustable parameters λ, the fitting error of LCD in decomposition process was overcame. The proposed method was applied to simulated and actual signals of composite faults. The results indicate that compared with LCD and empirical mode decomposition (EMD), PLCD can effectively extract the fault features of weak signals.
Key words: parameterized local characteristic scale decomposition (PLCD); parameterized Hermite interpolation; fault diagnosis; composite fault