Residual current recognition based on adaptive VMD and optimized DFNN

In order to realize rapid fault recognition of residual current device (RCD) and improve power safety, a fault residual current recognition method (AVMD-DFNN) based on adaptive variational modal decomposition (AVMD) and optimal dynamic fuzzy neural network (DFNN) is proposed. The decomposition parameters of VMD are determined adaptively by empirical mode decomposition (EMD) to realize the de-noising of the residual current signal. The characteristic parameters of residual current signal are extracted and used as the classification index of DFNN to recognize the type of residual current fault after the dimensionality reduction process. The DFNN is optimized by the minimum output method to remove the redundant fuzzy rule functions, so as to realize the rapid fault recognition of RCD. The simulation results show that AVMD-DFNN has high recognition accuracy and speed, which provides a theoretical reference for the development of new adaptive residual current devices.