Comparison of stochastic and deterministic security constrained optimal power flow under varying outage probabilities

Abstract

Increasing penetration of renewables and focus on cost efficiency means that power systems are operated closer to security limits. It has been suggested that the traditional N-l approach to security is unsatisfactory as it does not consider the likelihood and severity of contingencies. In this paper we investigate a stochastic version of security-constrained optimal power flow (S-SCOPF), which considers outage probabilities when deciding the optimal dispatch. Both a priori dispatch costs and realized real time costs are considered to evaluate the usefulness of the S-SCOPF a s opposed to the deterministic SCOPF (D-SCOPF). For low outage probabilities the S-SCOPF gives lower total costs, achieved by reducing the security for contingency situations. For high outage probabilities the two methods give the same dispatch. The robustness of the S-SCOPF to errors in the outage probability is also investigated. Underestimating the true outage probability results in the S-SCOPF giving higher total costs.