Fault-tolerant Network Design for Bounded Delay Data Transfer from PMUs to Control Center

TL;DR

This paper addresses the design of fault-tolerant communication networks for power grid monitoring, ensuring data from PMUs reaches the control center within delay constraints even if some links fail.

Contribution

It introduces the RDCMCCC problem for fault-tolerant network design with delay constraints, providing complexity analysis, an ILP formulation, and a heuristic solution.

Findings

Heuristic performs well on real substation data.

Optimal solutions are computable via ILP.

Fault-tolerant design ensures data delivery despite link failures.

Abstract

Communication network design for monitoring the state of an electric power grid has received significant attention in recent years. In order to measure stability of a power grid, it is imperative that measurement data collected by the Phasor Measurement Units (PMUs) located at the Sub-Stations (SS) must arrive at the Control Center (CC) within a specified delay threshold delta. In earlier papers we formalized the design problem as the Rooted Delay Constrained Minimum Spanning Tree (RDCMST) problem. However, RDCMST does not provide any fault-tolerance capability, as failure of just one communication link would prevent PMU data from reaching the CC. In this paper, we study the optimal cost network design problem with fault tolerance capability. In our model the PMU data from the SSs will have a path to reach the CC in spite of the failure of at most R links within the delay threshold delta. If R = 1, each SS will have two link disjoint paths of length at most delta to the CC. In other words, each SS will be on a cycle with the CC. We refer to this problem as the Rooted Delay Constrained Minimum Cost Cycle Cover (RDCMCCC) problem. We provide computational complexity analysis, an Integer Linear Programming formulation to find the optimal solution and a heuristic based on the sweeping technique. We evaluate the performance of our heuristic with real substation location data of Arizona.