Quantified Fault Severity Based Adaptive Commutation Failure Prevention Strategy for UHVDC System with Hierarchical Connection

• An adaptive commutation failure prevention strategy is proposed for the UHVDC system with hierarchical connections, based on quantified fault severity, to reduce unnecessary DC power reduction. • A differentiated Taylor-series-based DC current prediction is developed and embedded into non-faulted-layer commutation failure prevention control. • A hierarchical configuration table is developed to implement a graded control strategy, effectively suppressing commutation failure while reducing unnecessary DC power reduction. In ultra-high-voltage direct-current (UHVDC) systems with hierarchical connection (HC), an AC fault in one layer may trigger commutation failure (CF) not only in the faulted-layer converter but also in the non-faulted-layer converter because of inter-layer coupling and the shared DC current path. Existing prevention strategies are mainly triggered by local voltage signals and usually do not explicitly consider the rapid DC-current transient in the non-faulted layer or the different control requirements under different fault severities. To address this issue, this paper proposes an adaptive CF prevention strategy based on quantified fault severity for UHVDC systems with HC. First, the formation mechanism of CF in the faulted and non-faulted layers is analyzed, and an approximate analytical model of the post-fault DC current is derived. Then, a DC-current prediction method based on current derivatives is developed and embedded into the predictive commutation-current-area control and the voltage-dependent current-order limiter (VDCOL). Furthermore, the mapping between fault-layer voltage sag and CF risk is used to establish a graded control-engagement scheme. Simulations conducted in PSCAD/EMTDC demonstrate that the proposed method improves the CF immunity of UHVDC systems with HC and provides a practical basis for severity-aware adaptive protection and control design.