The research on the autonomous power balance framework for distribution network based on multi-agent modeling

The penetration of DG has brought many problems to the operation and management of distribution network. One of which is how to reasonably coordinate and allocate DCs outputs, how to achieve maximum use of energy and rational power efficiency under security and stability as well as how to meet the requirements of both global optimization and local autonomous control. Based on Muti-Agent Technology (MAS) and Replicator Dynamics (RD algorithm) of Evolutionary Game Theory, this paper analyzes the optimal scheduling of DG's output based on autonomous control of local power, and compares the influence of "cost", "rated power capacity" and "electrical distance" on the results of the game. By analysis, the rules are found that DGs with lower generation cost and greater rated power capacity has more opportunities to undertake load demand and the electrical distance could be ignored which will make the solution much easier and more efficient. The hierarchical structure of the whole system is described as the power grid, distribution network and Autonomous power units. The principle of RD algorithm is clear, and the fitness function is simple and efficient. Case study for IEEE test system proves the proposed algorithm and the game models are correct and reasonable.