Hybrid Centralized-Decentralized Economic Dispatch Based on A Distributed Finite-Step Consensus Algorithm with Divided Regional Incremental Costs

Hanwen Zhang, Hanqing Yang, Tieshan Li, Yue Long

TL;DR: This paper investigates a hybrid centralized-decentralized economic dispatch method that incorporates geographical region division, aiming to achieve economical power allocation without the need for full graph connectivity.

Abstract:

The economic dispatch problem of power systems, which is typically based on multi-agent networks, often necessitates conditions like full graph connectivity. However, as the scale of power systems expands, maintaining full graph connectivity in communication systems could result in increased costs and communication burdens. Consequently, this paper investigates a distributed finite-step consensus method that incorporates geographical region division, aiming to achieve economical power allocation without the need for full graph connectivity.

In this method, we initially derive a Laplace matrix from the distributed topology of generators in multi-agent power systems. Subsequently, we construct another Laplace matrix by geographically dividing regions, thereby illustrating the communication relationships between these regions. Subsequently, we calculate the incremental cost of each region, taking into account the power constraints of the generators. Moreover, to expedite convergence, we employ a distributed finite-step consensus algorithm to address the economic dispatch problem, leveraging the divided regional incremental costs. Finally, we validate the effectiveness and accuracy of the proposed method through results on various multi-agent topologies and comparisons with other iterative algorithms.

Submission Number: 63