Using Network Metrics to Achieve Computationally Efficient Optimal Transmission Switching

Recent studies have shown that dynamic removal of transmission lines from operation ("Transmission Switching") can reduce costs associated with power system operation. Smart Grid systems introduce flexibility into the transmission network topology and enable co-optimization of generation and network topology. The optimal transmission switching (OTS) problem has been posed in on small test systems, but problem complexity and large system sizes make OTS intractable. Our previous work suggests that most economic benefits of OTS arise through switching a small number of lines, so pre-screening has the potential to produce good solutions in less time. We explore the use of topological and electrical graph metrics to increase solution speed via solution space reduction. We find that screening based on line outage distribution factors outperforms other methods. When compared to un-screened OTS on the RTS-96 and IEEE 118-Bus networks, the sensitivity-based screen generates near optimal solutions in a fraction of the time.