Robust Current Sensor Fault Tolerant Control of an Interleaved DC-DC Converter Based on EDSMO with MFPC

Fault diagnosis plays a crucial role in maintaining the integrity of power converters and optimizing maintenance interventions. This paper presents a robust fault diagnosis method for interleaved DC-DC converters that relies solely on the sensor current. A state observer-based model-free predictive control (MFPC) technique is utilized, which integrates an ultralocal model and an estimated disturbance sliding mode observer (EDSMO) for the DC-DC converter. Two state observers are explored: the conventional proportional-integral observer (PIO) and the proposed EDSMO. The approach first rejects model disturbances in the MFPC, followed by the injection of external disturbances to the current sensor to evaluate the robustness of the proposed method. Instead of relying on a mathematical model, the ultra-local model of the converter is employed. Simulation results demonstrate the effectiveness of the MFPC with EDSMO, highlighting its superior resilience and control performance compared to the traditional PIO, particularly in the context of an interleaved DC-DC converter.