Study on fault diagnosis method for planetary gearboxes based on optimized VMD and ELM
YANG Rongkun
JIANG Hong
ZHANG Xiangfeng
Abstract：[Objective]Due to the complex structure of planetary gearboxes,fault features are difficult to extract,and traditional methods rely heavily on professional expertise.To solve these problems,a fault diagnosis method integrating beluga whale optimization(BWO)algorithm optimized variational mode decomposition(VMD),multi-scale permutation entropy(MPE),and extreme learning machine(ELM)was proposed.[Methods]Firstly,the BWO algorithm was employed to optimize the decomposition layers K and penalty factor α of VMD using the minimum envelope entropy as the objective function to achieve adaptive signal decomposition.Secondly,the MPE algorithm was used to compute the non-linear features of the intrinsic mode function(IMF)components,and a feature vector consisting of five time-domain indexes was constructed.Finally,the vectors were fed into the ELM for training and diagnosis.Comparative tests were conducted on a planetary gearbox test bench under four working conditions.[Results]The testing results show that the overall accuracy of the proposed method reaches 97.92%,which is significantly higher than that of EMD-ELM and optimized VMD-SVM models.The findings verify that the BWO-VMD effectively improves signal de-noising and adaptive decomposition.This research provides a reliable basis for the health monitoring and precision design of planetary gearboxes.
Keywords：Planetary gearboxVariational mode decompositionBeluga whale optimizationMulti-scale permutation entropyExtreme learning machineFault diagnosis
Publication Date：2026-03-15
Online Publishing Date：2026-09-12(First online date of this platform, not the publication date of the document)
Pages：7( 172-178 )