A cooperative PSO algorithm for Volt-VAR optimization in smart distribution grids

Abstract

The penetration level of photovoltaic (PV) systems is set to increase in the following years and already distribution networks (DNs) are straining to overcome the adverse voltage effects caused by reverse power flows and intermittent generation phenomena. Moreover, regular voltage optimization approaches that employ PV inverters as control devices, cannot deal effectively with the escalating dimensionality of the problem. This paper introduces a reactive power optimization method for PV-heavy DNs based on cooperative particle swarm optimization (CPSO). The proposed approach makes use of multiple swarms, each swarm containing a group of design variables that are interrelated with respect to the optimization objective; a community detection algorithm is employed to assign the design variables to the different swarms, by identifying voltage-decoupled zones of the grid. The different swarms cooperate by exchanging information in order to better explore the search space, while still solving the optimization problem as a whole. The feasibility and effectiveness of the proposed scheme are demonstrated through comparisons with other approaches for various load and generation scenarios on the IEEE 123-bus distribution system.