A Fault Diagnosis Method for NPC Three-Level Inverters Based on VMD-SE and SMA-SVM

To tackle the challenges of low accuracy and high computational complexity in diagnosing open-circuit faults for NPC three-level inverters, this paper introduces a novel fault diagnosis approach that integrates Variational Mode Decomposition (VMD), Sample Entropy (SE), and an enhanced Slime Mold Algorithm (SMA) optimized Support Vector Machine (SVM). The method begins by decomposing the fault current signal using VMD and calculating the SE to identify fault characteristics. Next, it employs a correlation coefficient to filter out invalid Intrinsic Mode Functions (IMFs). Finally, an improved SMA is applied to optimize the SVM for accurate fault classification and identification. Simulation results show that this method outperforms traditional techniques in terms of effectiveness and accuracy.