A VMD-Enhanced ELM Method for EVT Ratio Error Prediction

To assess future measurement accuracy changes in electronic current transformers, a ratio error prediction model is introduced, integrating Time-Shifted Multiscale Permutation Entropy, Variational (TSMPE) Mode Decomposition (VMD), and Extreme Learning Machine (ELM). First, the signal is decomposed using VMD, guided by TSMPE. Then, the complexity of the resulting Intrinsic Mode Functions is quantified with Permutation Entropy, and the sequence is reconstructed. Finally, ELM predicts the reconstructed sequence, yielding the final results. The case study confirms the method's effectiveness in enhancing current transformer measurement accuracy.