PMU Placement Optimization for Efficient State Estimation in Smart Grid

Ye Shi, Hoang Duong Tuan, Trung Q. Duong, H. Vincent Poor, Andrey V. Savkin

Abstract

This paper investigates phasor measurement unit (PMU) placement for informative state estimation in smart grid by incorporating various constraints for observability. Observability constitutes an important property for PMU placement to characterize the depth of the buses' reachability by the placed PMUs, but addressing it solely by binary linear programming as in many works still does not guarantee a good estimate for the grid state. Some existing works have considered optimization of some estimation indices by ignoring the observability requirements for computational ease and thus potentially lead to trivial results such as acceptance of the estimate for an unobserved state component as its unconditional mean. In this work, the PMU placement optimization problem is considered by minimizing the mean squared error or maximizing the mutual information between the measurement output and grid state subject to observability constraints, which incorporate operating conditions such as presence of zero injection buses, contingency of measurement loss, and limitation of communication channels per PMU. The proposed design is thus free from the fundamental shortcomings in the existing PMU placement designs. The problems are posed as large scale binary nonlinear optimization problems involving thousands of binary variables, for which this paper develops efficient algorithms for computational solutions. Their performance is analyzed in detail through numerical examples on large scale IEEE power networks. The solution method is also shown to be extendable to AC power flow models, which are formulated by nonlinear equations.

Introduction

To manage a smart grid effectively, operators require a precise, real-time map of the system's state (voltages and currents), a process known as State Estimation. Phasor Measurement Units (PMUs) are the high-tech sensors that provide the necessary data for this task. However, PMUs are expensive to install and maintain, necessitating an optimized placement strategy.

Historically, researchers have approached PMU placement in two ways: connectivity-focused methods that ensure every node (bus) is observable, and quality-focused methods that optimize for Mean Squared Error (MSE) or Mutual Information (MI). The former often fails to provide accurate estimates, while the latter frequently ignores practical observability constraints due to the computational complexity of large-scale binary nonlinear optimization. This paper bridges this gap by maximizing estimation quality while strictly adhering to operational and observability constraints.

Methodology

The authors develop analytical models for MSE and MI that are explicitly functions of a binary selection vector. To solve the resulting large-scale optimization problem without resorting to exhaustive searches, they propose:

  • Exactly Penalized Optimization: Discrete binary constraints are transformed into continuous ones using a mathematical penalty trick, allowing the use of continuous optimization techniques on discrete problems.
  • Scalable Path-Following Algorithms: The paper introduces two algorithms: Algorithm 1 for constrained MSE/MI optimization and Algorithm 2, a faster, greedy-style path-following algorithm for baseline performance.
  • Real-world Constraints: The model incorporates Zero Injection Buses (ZIBs), Power Flow Measurements (PFMs), single-PMU outage contingencies, and limits on the number of communication channels per PMU.

Experimental Evaluation

The researchers evaluated their algorithms on standard IEEE power system benchmarks and large-scale European/Polish grids, ranging from the IEEE 30-bus system to the 2383-bus Polish network.

Diagram of an IEEE 30-bus system illustrating the placement of PMUs and generators within the network topology.

Key Results and Insights

  • Scalability and Efficiency: The proposed algorithms successfully handled the 2383-bus Polish network, a feat that traditional semi-definite programming (SDP) relaxations struggle with. For the IEEE 300-bus system, the path-following approach was significantly faster and more accurate than traditional convex relaxation methods.
  • The "Trivial Estimate" Trap: The study highlights that ignoring observability constraints can lead to useless "unconditional mean" estimates for unobserved components. Observability is the necessary floor for meaningful estimation.
  • Communication Impact: A key finding is the impact of communication channel limits. If a PMU can only report on a few incident lines, the number of required units increases significantly.
  • Contingency Robustness: The model successfully placed PMUs such that the grid remained observable even during a single PMU failure, providing a robust framework for grid stability.

Conclusion

This research provides a scalable framework for PMU placement that balances mathematical rigor with practical engineering constraints. By moving beyond simple connectivity to optimize for information quality, the proposed algorithms ensure that smart grid operators have the most accurate data possible for system management. Future work will explore extending these models to dynamic state estimation and addressing physical layer security for PMU networks.

Summary updated

Abstract

This paper investigates phasor measurement unit (PMU) placement for informative state estimation in smart grid by incorporating various constraints for observability. Observability constitutes an important property for PMU placement to characterize the depth of the buses' reachability by the placed PMUs, but addressing it solely by binary linear programming as in many works still does not guarantee a good estimate for the grid state. Some existing works have considered optimization of some estimation indices by ignoring the observability requirements for computational ease and thus potentially lead to trivial results such as acceptance of the estimate for an unobserved state component as its unconditional mean. In this work, the PMU placement optimization problem is considered by minimizing the mean squared error or maximizing the mutual information between the measurement output and grid state subject to observability constraints, which incorporate operating conditions such as presence of zero injection buses, contingency of measurement loss, and limitation of communication channels per PMU. The proposed design is thus free from the fundamental shortcomings in the existing PMU placement designs. The problems are posed as large scale binary nonlinear optimization problems involving thousands of binary variables, for which this paper develops efficient algorithms for computational solutions. Their performance is analyzed in detail through numerical examples on large scale IEEE power networks. The solution method is also shown to be extendable to AC power flow models, which are formulated by nonlinear equations.
BibTeX
@article{shi2019pmu,
  author = {Ye Shi and Hoang Duong Tuan and Trung Q. Duong and H. Vincent Poor and Andrey V. Savkin},
  title = {{PMU Placement Optimization for Efficient State Estimation in Smart Grid}},
  journal = {IEEE Journal on Selected Areas in Communications},
  year = {2019},
  doi = {10.1109/jsac.2019.2951969},
}
Figure