Fault detection of converter valve based on the combination of fault tree analysis and multi-frequency impedance testing

Converter valves include numerous components in thyristor-level damping, voltage-sharing, and power-tapping circuits, and their complex structure makes it difficult for conventional impedance tests to localize faults beyond qualitative valve-level judgments. Here, we propose a fault detection method that integrates fault tree analysis (FTA) with multi-frequency comprehensive impedance testing. Multi-frequency small-signal excitation voltages are applied across the terminals of a thyristor level, and the corresponding active/reactive power responses are measured; a multi-objective optimization model with regularization and parameter-bound constraints is formulated to jointly identify key resistance and capacitance parameters, and parameter deviations are subsequently mapped onto the fault tree to relate test observations to fault modes. MATLAB simulations show that the proposed method can accurately identify faulty components in the associated circuits without dismantling internal wiring. By linking multi-frequency impedance features with FTA-based fault-mode reasoning, the method improves fault localization and diagnostic interpretability, offering practical value for converter-valve testing and maintenance.