Stochastic Electricity Social Welfare Enhancement Based on Consensus Neighbor Virtualization

This paper tries to develop a multiagent consensus-based algorithm to handle the energy management problem (EMP) in the smart grids in the presence of dispatchable loads and distributed generators. To this end, the neighbor virtualization concept is employed to improve the system performance by introducing a novel communication structure for distributed algorithms. Accuracy, quickness, and scalability are the most important benefits of this algorithm that make it applicable in real power systems. The increasing penetration of wind turbines and the stochastic nature of load demand can inject much uncertainty in our problem. In order to deal with this issue, unscented transform as a powerful tool is used to model the problem uncertainties. The proposed problem is formulated in the form of a single-objective stochastic optimization framework maximizing the generators and consumers welfare. The feasibility and high performance of the proposed framework are examined on two test systems, including the IEEE 39-bus test system and a large grid with 1305 agents in MATLAB simulation. The simulation results advocate the high capability and effectiveness of the proposed framework for EMP of smart grids.