Distributed robust economic dispatch in power systems: A consensus + innovations approach

The paper presents a fully distributed approach for economic dispatch in power systems. The approach is based on the consensus + innovations framework, in which each network agent participates in a collaborative process of neighborhood message exchange and local computation. The distributed approach is shown to converge to the optimal dispatch under rather weak assumptions on the agent communication network connectivity. Intuitively, the proposed approach includes a consensus term which achieves convergence to a common incremental cost value and an innovation term which ensures that the total generation is equal to the load to be supplied. Further, in the proposed approach each network bus only needs to be aware of its local cost parameters and the local predicted load and the neighborhood communication involves exchanging only the locally determined marginal energy price with a few neighbors. Robustness of the proposed approach and techniques for convergence rate improvement are discussed. Finally, simulation studies on the benchmark IEEE 14 bus system demonstrate the effectiveness of the approach.