Model-Based Fault Diagnosis Scheme for Current and Voltage Sensors in Grid Side Converters

The effective operation of the grid side converter is crucial for the reliability of renewable energy systems. Grid side converters are responsible for converting the energy produced by renewable energy sources into a form that can be injected into the grid. A grid side converter controller manages the conversion using feedback from voltage and current sensor measurements from both the alternative current (ac) and direct current (dc) sides of the converter. However, these sensory devices may fail or diverge from their normal operation, which can affect the effective operation of the converter. This article proposes a model-based fault diagnosis scheme for sensors located in the dc and ac sides of the grid-tied converter, taking into account the effects of fault propagation due to the physical interconnection between the ac and the dc sides. First, a discrete-time, nonlinear model of the grid side converter is formulated that considers the presence of ac current and dc voltage sensor faults. Second, a nonlinear estimation scheme is designed by considering the effects of sensor noise and modeling uncertainty. The estimator design utilizes the estimator gain matrix that is obtained by solving the linear matrix inequalities problem. Finally, a sensor fault diagnosis scheme is designed based on the analytical redundancy relations considering the estimations of the nonlinear estimation scheme. The performance of the proposed fault diagnosis scheme is validated through both simulation and experimental results, demonstrating its effectiveness in timely detecting and isolating sensor faults.© 2022 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works. This work was supported in part by the project PVgnosis under Grant P2P/SOLAR/0818/0007 through Horizon 2020 SOLAR-ERA.NET Cofund 2 (Cofund ERA-NET Action, N° 786483), in part by the Cyprus Research and Innovation Foundation, in part by the European Union's Horizon 2020 research and innovation programme under grant 739551 (KIOS CoE - TEAMING), and in part by the Deputy Ministry of Research, Innovation and Digital Policy, Republic of Cyprus