Constructing ISSA-ELM power electronic circuit soft fault diagnosis method based on VMD
To address the issue of poor accuracy in soft fault diagnosis of power electronic circuits,a fault diagnosis strategy that combines variational mode decomposition(VMD)with an improved sparrow search algorithm(ISSA)to optimize the extreme learning machine(ELM)was designed.This study initially used VDM to decompose collected fault signals to extract intrinsic mode functions(IMF)and then extracted time-domain parameters from the linearly reconstructed IMF,which act as feature vectors for fault diagnosis.Next,the parameters of the ELM were optimized through the ISSA to enhance the diagnostic accuracy of ELM,resulting in the establishment of the ISSA-ELM model.The ISSA first utilized circle chaotic mapping to initialize the population,then introduced a convergence factor for updating the follower's position,and finally applied three methods,including inverse learning and disturbing the optimal solution by Cauchy mutations to improve the sparrow search algorithm(SSA).Through testing with eight benchmark functions,ISSA exhibited a faster convergence speed and higher optimization accuracy compared to four other intelligent algorithms.The results indicated that the VMD combined with the ISSA-ELM model achieved an average accuracy exceeding 99.000 0％in soft fault diagnosis of a 150 W Boost circuit,surpassing the accuracy of other models,which confirms the feasibility of the proposed soft fault diagnosis method for DC-DC circuits.
万方数据
维普
