Agent-based power routing in Active Distribution Networks.

The expected large-scale implementation of distributed generation (DG) requires a change in the current structure and operation of distribution networks. The future distribution network must be able to manage power flow in a bidirectional way, cope with uncertainties of renewable power generation and adjust to demands of more sophisticated customers. This paper introduces the concept of a power routing function to avoid congestion, minimize the operating cost, and adequately serve the requirements of customers. This function considers the optimal power flow as a problem of minimum cost flow in the graph theory. The Scaling Push-Relabel (SPR) algorithm is used to solve that problem. It will be implemented in a distributed agent environment which is suitable with a design concept of Active Distribution Networks. The performance of the power routing function is tested on a simulation of the medium voltage 32-bus network. Simulation results show the effectiveness and flexibility of the proposed function in dealing with issues of load demand increases and network configuration changes. [ABSTRACT FROM PUBLISHER]