Stochastic optimal reactive power dispatch: Formulation and solution method

Classical optimal reactive power dispatch (ORPD) is usually formulated as a deterministic optimization problem, such that the network structure and load power injections are known and fixed. Hence, the influences of load uncertainties and branch outages are not typically considered. This paper proposes a chance-constrained programming formulation for ORPD that considers uncertain nodal power injections and random branch outages. A solution method combining both probabilistic load flow and a genetic algorithm is proposed and demonstrated in order to solve the problem. Simulations on several test systems show that the proposed method can prevent under-compensation or over-compensation of reactive power and increase voltage security margins. These advantages are achieved with the acceptable expense of a small increase in active power loss when compared with the results of classical deterministic ORPD. Several techniques are presented in order to improve the efficiency of the proposed method and case studies demonstrate the effectiveness of the techniques.