 |
Center for Nonlinear Studies |
We consider the problem to find AC power flow feasibility regions where the AC power flow equation are guaranteed to have a feasible solution. Construction of these regions is based on statistical machine learning accelerated by importance sampling driven by feasibility proxies. Resulting regions have simple analytic description in the space of power flow injections and can be efficiently used for various emergency control problems. Efficiency and tightness of the approach is validated on a number of test cases.
Host: Chris Neale