Abstract
In addressing the issue of open-switch (OS) fault diagnosis for grid-connected three-phase inverters under unknown parameter conditions, a method combining input-compensated recursive least squares (ICRLS) and a discrete disturbance sliding mode observer (DSMO) is proposed in this paper. First, an ICRLS approach is introduced, which enhances the accuracy of parameter identification by compensating for unknown disturbances, thereby improving fault diagnosis reliability. Subsequently, a novel DSMO is presented for the rapid and precise estimation of three-phase currents. Then, adaptive fault detection variables are designed based on these observations, ensuring the robustness of the detection algorithm. Finally, by constructing a fault phase identification and fault identification mechanism, the method achieves precise identification of 21 different types of OS faults in grid-connected three-phase inverters. The effectiveness and robustness of the proposed method are validated through hardware-in-the-loop (HIL) testing results.
