Preferred action order

Verify the contingency definitions, monitored elements, and limits.
Group critical contingencies by common corridor, interface, or weak area.
Apply preventive actions first: redispatch, shunts, tap changes, or topology changes that help multiple contingencies.
Evaluate corrective actions next: post-contingency switching, redispatch, or RAS schemes with explicit triggers.
Use load shedding only as a last-resort action and state that clearly.
For recurring problems, recommend reinforcement or a revised operating guide.

Working rules

Do not trust contingency results from a bad base case.
Prefer actions that improve a family of contingencies rather than a single outage.
Re-run the full monitored set after any candidate fix.

Start with a solved base case and a ranked list of critical contingencies.

Verify the contingency definitions, monitored elements, and limits.
Group critical contingencies by common corridor, interface, or weak area.
Apply preventive actions first: redispatch, shunts, tap changes, or topology changes that help multiple contingencies.
Evaluate corrective actions next: post-contingency switching, redispatch, or RAS schemes with explicit triggers.
Use load shedding only as a last-resort action and state that clearly.
For recurring problems, recommend reinforcement or a revised operating guide.

Do not trust contingency results from a bad base case.
Prefer actions that improve a family of contingencies rather than a single outage.
Re-run the full monitored set after any candidate fix.