A distributed power routing method between regional markets based on Bellman-Ford algorithm

Xue‐Cheng Tai, Hongbin Sun, Qinglai Guo, Zihao Li

Abstract

With the deregulation of the power industry, large power users can buy electricity directly from power producers. For trades between different regional markets, it is necessary to research methods to find the cheapest route for power exchanging. Moreover, as there exists no coordinator above regional markets, the power routing problem must be solved in a distributed way. This paper proposes a method based on the Energy Management Systems (EMS) of multiple markets, which saves cost by eliminating expensive electronic transformers and has high efficiency in systems with numerous nodes by considering the regional markets as generalized nodes. We take transmission costs and congestion costs into consideration and use the Bellman-Ford algorithm to determine the cheapest route for power transmission between different regional markets. This method proves effective and convergent when there is no loop among the markets. A four-market loop-free case was used to illustrate the efficiency of the method.

Introduction: The Challenge of Inter-Market Power Trading

As the power industry deregulates, large consumers can now buy electricity directly from producers across different regions. However, the power grid is not a single entity; it is a patchwork of regional markets. Finding the most cost-effective path for a transaction is difficult because there is often no overarching coordinator to manage trades between regions. Furthermore, previous "energy router" concepts relied on expensive power electronic transformers at every node, and electricity must follow physical constraints (Kirchhoff’s laws), meaning transactions can cause congestion that adds hidden costs. This research proposes a distributed software approach that works with existing infrastructure to solve the "cheapest route" problem.

Methodology: Markets as Virtual Routers

The proposed method models interconnected regional markets as an undirected graph where vertices are regional markets and edges are the transmission lines between them. Instead of new hardware, the existing Energy Management System (EMS) of each regional market acts as a "virtual router," adjusting local generation to facilitate through-flow power.

The "weight" of a path is calculated as the sum of:

  • Transmission Fees: Linear costs for using lines and market infrastructure.
  • Congestion Fees: The extra cost incurred when a transaction forces a market to run more expensive local generators to maintain grid stability.

To find the optimal path, the system employs a modified Bellman-Ford algorithm. To prevent the "count-to-infinity" problem common in distance-vector protocols, the algorithm is modified so each node records and updates its full path information during iterations.

Case Study: A Four-Market Radial System

The method was evaluated using a synthetic four-market case study. Each regional market is modeled as a standard 9-node power system, connected in a loop-free (radial) topology to test convergence and efficiency.

Topology of the four-market test case consisting of interconnected 9-node systems.

By treating an entire regional market as a single "generalized node," the complexity of the global routing problem is significantly reduced. This allows the EMS to calculate dynamic congestion fees on the fly using Optimal Power Flow (OPF) equations.

Key Results: Convergence and Congestion Management

The simulation demonstrated several key findings:

  • Linear Convergence: In loop-free networks, the algorithm successfully identified the optimal path in at most $|V|-1$ iterations.
  • Efficiency: In the 4-market test, the system identified the "A-B-C-D" path as the cheapest route for a 100MW transaction.
  • Congestion Impact: When line limits were tightened, the algorithm correctly captured economic bottlenecks. For instance, the transaction cost in Market A rose from $7,965.05/h to $8,103.13/h, reflecting the increased cost of managing grid stability.
Transaction costs across markets A, B, C, and D under specific grid congestion conditions.

Limitations and Future Outlook

While effective, the current proof of convergence only applies to loop-free (radial) networks. Real-world power grids are often highly meshed, which can lead to circulating power flows and convergence issues for simple distance-vector algorithms. Additionally, the current model processes transactions sequentially (first-in-first-served). Future research will focus on extending this distributed logic to systems with loops and exploring simultaneous auction-style routing for high-volume markets.

Summary updated

Abstract

With the deregulation of power industry, large power users can buy electricity directly from power producers. For trades between different regional markets, it is necessary to research on the method to find the cheapest route for power exchanging. Moreover, as there exists no coordinator above regional markets, the power routing problem must be solved in a distributed way. Most previous research are based on power electronic transformers. Some research proposed EMS of subnetworks as energy routers, but failed to consider the power flow constraints or did not converge in some specific cases. Our method is based on the EMS of multiple markets, which saves cost by eliminating expensive electronic transformers and has high efficiency in systems with numerous nodes by considering the regional markets as generalized nodes. We take transmission costs and congestion costs into consideration, and uses Bellman-Ford algorithm to determine the cheapest route for power transmission between different regional markets and make convergence analysis about our algorithm. This method proves effective and convergent when there is no loop among the markets. A four-market loop-free case was used to illustrate the efficiency of the method.
BibTeX
@misc{tai2017distributed,
  author = {Xue‐Cheng Tai and Hongbin Sun and Qinglai Guo and Zihao Li},
  title = {{A distributed power routing method between regional markets based on Bellman-Ford algorithm}},
  year = {2017},
  eprint = {1711.02470},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/1711.02470},
}
Figure