The cascaded H-bridge rectifier (CHBR), serving as the front-end equipment in high-speed train traction systems, is highly susceptible to pantograph-catenary separation (PCS), which induces severe voltage/current fluctuations and interferes with IGBT fault detection. This paper establishes a mathematical model of PCS and analyzes its impact mechanism on system residual variations. An adaptive fault detection strategy based on real-time condition recognition is proposed, which dynamically identifies PCS events by monitoring input current, input voltage, and DC-link voltage, thereby effectively suppressing false diagnoses. Simulation studies conducted across diverse operating scenarios demonstrate the robustness and efficacy of the proposed approach.
 

Pantograph-catenary separation, Single-phase cascaded H-bridge rectifier, IGBT failure, Residual variation, Condition identification, Fault detection optimization