Directional Fault-Response Matrices for Multi-Infeed LCC-HVDC Siting

Multi-infeed LCC-HVDC siting requires a static strength test, a finite-fault propagation assessment, and an auditable rule for selecting the subset of candidates that receives detailed electromagnetic-transient (EMT) evaluation. This paper develops a directional fault-response matrix whose row denotes the responding converter bus and whose column denotes the fault-source terminal. A two-infeed derivation, stated for scalar or common-angle impedances and checked numerically for unequal X/R ratios, shows that a stronger inter-bus tie can raise the weaker terminal’s simplified MSCR while increasing the mutual impedance and the cross-terminal fault response. Structural properties of the nonnegative matrix are derived, and a closed-form first-order surrogate calibrated by one pooled least-squares gain screens all 15 four-infeed combinations of six synthetic sites and routes seven candidates (28 single-source switching-EMT runs) to detailed verification. The complete 60-run library is additionally simulated offline and used only as ground truth: the reduced and EMT composite scores have Spearman coefficients of 0.750 over the routed set and 0.911 over all 15 candidates (0.911 under leave-one-candidate-out calibration), the reduced hard screen produces zero false retentions and zero false rejections, and the matrix score is compared with MSCR-only and impedance-ratio MIIF baselines. ABCE is the best verified candidate, with a minimum simplified MSCR of 2.008 and a maximum EMT cross-response of 0.0395 p.u. A 0.05 ohm stress profile activates the transient limit for 9 of the 15 candidates. Element-level error, weight and threshold sensitivity, measured simulation times, and model limitations are reported. The method remains a screening and diagnostic framework rather than a stability law or a proof of global EMT optimality.