Non-interconnected power systems operate with limited flexibility and tight transmission margins, making them particularly sensitive to single-element outages and post-contingency congestion, especially as renewable energy sources (RESs) increase variability and operational stress. This paper presents a reproducible, planning-stage screening workflow for N−1 security assessment in an islanded transmission network, combining lossless DC contingency analysis with an interpretable overload-severity formulation and directional PTDF/TLR sensitivities. The framework is demonstrated on the autonomous 150 kV power system of Rhodes (Greece), an island grid with conventional generation and distributed wind/PV resources and strongly seasonal demand.

The workflow proceeds in four stages. First, an N−1 contingency set is executed and all post-contingency thermal violations are extracted under the selected branch ratings. Second, each violation is quantified using an MVA exceedance metric, Severity _c,e=max(S_e - S_e^max , 0) (MVA) , which captures overload criticality in a comparable form across contingencies and elements. Third, directional PTDF/TLR sensitivities are computed for incremental transfers from candidate buses to a designated sink (balancing) bus, and the sign of the response is used to retain actions aligned with relieving the violated flow direction. Fourth, severity-weighted, signed relief contributions are aggregated to produce a bus-level relief score, enabling ranking of candidate locations and identification of the dominant contingencies and binding network elements that govern thermal insecurity.

To address a frequently overlooked modeling dependency, the study also evaluates how the assumed balancing (sink) location influences PTDF directions and the resulting rankings, demonstrating that siting conclusions can change materially with different balancing assumptions. Overall, the proposed method transparently bridges contingency screening and controllability assessment, providing decision support for prioritizing flexible resources (e.g., storage discharge or redispatch capability) and guiding where more detailed AC, OPF, or dynamic studies should focus.