Fault Diagnosis and Tolerant Control of Power Converters in Renewable Energy Systems Using Observer-Based Estimation and Model Predictive Control

Abstract

The power converters are essential parts of the renewable energy systems but their faults like switch failures and sensor faults have great influence to the performance of converters thus they are unstable and lack reliability thus deteriorating the performance of converters. The given paper suggests an all-encompassing system of fault diagnosis and fault-tolerant control of power converters through an observer-based estimation scheme based on the Model Predictive Control (MPC). A strong observer is developed to estimate the states of the system and produce residual signals so that fault detection and fault isolation can be performed in real time. The decision mechanism modeled on threshold allows the correct detection of different types of faults and the MPC strategy restructures the control measures to allow stable operation in the case of faults. The described strategy is tested in simulation in a variety of fault conditions, such as open-circuit switch faults and sensor faults. Key fault diagnostic measurements include performance on detecting faults in terms of detection accuracy, fault detection time, false alarm rate and fault isolation time, and control performance measurements involving settling time, recovery time, total harmonic distortion (THD), and efficiency. Experiments show that the developed method has a high success in detecting faults (>95%), fault and recovery are achieved to be much shorter, and the output is stable with better power quality than the conventional control methods. The proposed integrated observer-MPC framework is both effective and trustworthy when it comes to strengthening the robustness and resilience of renewable energy converter systems, therefore, it can be utilized in real-time and smart grid projects.