Multi-Timescale Fault-Propagation Mechanism and Fault Ride-Through Control for Fiber-Optic Communication Failures in VSC-HVDC Converter Valves

The conventional fault-control strategy for fiber-optic communication failures in modular multilevel converters (MMCs) directly bypasses faulty submodules (SMs). However, once communication is restored, the bypassed SMs cannot be reintegrated, and the converter must shut down when redundancy is exhausted. This paper establishes a fault-propagation model and proposes a long-timescale non-bypass fault-tolerant control strategy that avoids unnecessary SM bypassing and improves system reliability. First, a mathematical fault-propagation model is developed for downlink command loss and uplink status-feedback interruption. The model reveals a positive correlation between modulation-index deviation and current fluctuation during downlink faults and derives the coupling mechanism between switching states and DC-voltage fluctuation during uplink faults. On this basis, a non-bypass fault-tolerant control method is developed in which a faulty SM switches to a local constant-voltage closed-loop mode. This mode preserves the SM’s electrical connection and voltage stability while enabling rapid resynchronization and recommissioning after fiber-optic communication is restored. Hardware-in-the-loop (HIL) results show that the proposed strategy limits current fluctuations to within 0.1% under transient faults and DC-voltage fluctuations to within ±1% under permanent faults while substantially improving system availability compared with the conventional bypass scheme. The proposed method provides a cost-effective and practically implementable approach to maintaining uninterrupted operation of VSC-HVDC MMCs under communication-link failures.