What the study found
The most severe outage was T04 7-8, which islands bus 8. Among solved contingencies, the ranking combines voltage-limit breaches, thermal exceedance, loss increase, and failure-state penalties. The base AC solution was cross-checked with a second solver before contingencies ran.
Network model
The test grid contains 14 buses, 15 lines, five transformers, five voltage-controlled sources including the slack, and 259 MW of load. Slack, PV, and PQ buses are treated according to which voltage and power variables are specified.
Fast screening versus full AC
PTDFs map balanced bus injections to line-flow changes. LODFs use those sensitivities to estimate how pre-outage flow redistributes after a branch trip. Across the selected comparison set, mean signed-flow absolute error was 1.17 MW; the bridge outage is deliberately not screened because its LODF denominator identifies islanding.
Engineering honesty
The source case does not provide continuous branch ratings. I therefore use clearly labeled synthetic study ratings for a reproducible stress test; no result is presented as an actual utility limit. The model is a teaching and portfolio study, not an operating decision tool.