Open-Circuit Fault Detection and Isolation Method for Five-Level PUC Inverter Based on the Wavelet Packet Transform of the Radiated Magnetic Field
Ibtissem Abari, Mahmoud Hamouda, Mohammad Sleiman, Jaleleddine Ben Hadj Slama, Hadi Y. Kanaan, Kamal Al‐Haddad
Abstract
This article proposes a noninvasive open-switch fault detection method based on the analysis of the high-frequency signals. The faulty power switch is detected by metering the emitted near-field above a new topology using the wavelet packet transform and considering a complex control algorithm. The proposed method is tested on a single-phase five-level packed U-cell (PUC5) inverter. The latter is controlled using the finite-control set model predictive control (FCS-MPC) technique that generates the gate signals with a variable switching frequency. Moreover, unlike the conventional pulsewidth modulation strategy, the switching pattern provided by the FCS-MPC technique is not known in advance. Despite these complexities, the proposed fault detection method proves its high performance and effectiveness. Experimental studies are carried out on a single-phase PUC5 inverter using the real-time controller (OP4510) from OPAL-RT and prove the capability of this method to properly identify the faulty switch in this multilevel topology.