Two-Stage Dynamic Partitioning Strategy Based on Grid Structure Feature and Node Voltage Characteristics for Power Systems

Lixia Sun, Xianxue Sha, Shuo Zhang, Jiahao Wang, Yiping YuEnergies 2025 · DOI

TL;DR

This paper introduces a novel two-stage dynamic partitioning strategy for power grids, enhancing adaptability to transient events. By combining electrical coupling strength and transient voltage correlation, the method creates more cohesive and effectively decoupled grid regions, leading to improved voltage control and system stability.

What Problem Does This Paper Address?

Modern power grids are increasingly complex, with the integration of renewable energy sources introducing significant volatility and uncertainty. This makes traditional methods of grid partitioning, often based on static or geographical features, insufficient for effective voltage control and stability management, especially during transient events like faults. The challenge lies in creating dynamic, functionally coherent partitions that reflect the actual electrical behavior of the grid under stress, ensuring strong internal connections within a partition and weak connections between partitions. This is crucial for implementing localized voltage control strategies and improving overall system resilience.

Methods

The proposed solution employs a two-stage dynamic partitioning strategy:

  1. Pre-Partitioning based on Electrical Coupling Strength: This stage focuses on identifying initial groupings.
  2. Electrical Coupling Strength: This is quantified by combining two metrics:
  3. Short-circuit impedance: Measures the direct electrical connection between reactive power source nodes and load nodes.
  4. Reactive power-voltage sensitivity: Assesses how effectively reactive power sources can influence voltage at load nodes.
  5. These two metrics are normalized and weighted to create an "electrical coupling strength" index.
  6. Load nodes are then assigned to partitions based on their highest electrical coupling strength with reactive power source nodes, forming initial, functionally related groups.
  7. Second-Stage Partitioning based on Modularity Optimization: This stage refines the initial partitions.
  8. Node Voltage Correlation: Transient voltage behavior across various fault scenarios is captured using cosine similarity on voltage feature vectors. This creates a weighted graph where edge weights represent the similarity of transient voltage responses between nodes.
  9. Modularity Maximization: A community detection algorithm is used to optimize the modularity of the weighted graph. Modularity measures the strength of community structures within a network, aiming to maximize intra-partition connections while minimizing inter-partition connections.
  10. The process iteratively merges adjacent partitions, selecting the merger that yields the highest modularity, until an optimal partitioning scheme is achieved. This approach automatically determines the optimal number of partitions.

Datasets

The study utilizes a modified IEEE 39-bus system for its case study. This system is enhanced with renewable energy sources (wind and PV) and a DC feeder. Various operating scenarios were simulated using the PSD-BPA software (version 2.8), including different load levels, fault locations (on 35 buses), fault distances from line origins, and fault clearance times. A total of 1260 scenarios were generated to collect transient voltage data for all nodes.

Key Insights and Ideas

  • Structure-Function Coupling is Key: The paper emphasizes that effective partitioning requires considering both the physical network structure (topology, impedance) and its functional behavior (voltage response, reactive power control).
  • Transient Behavior Matters: Static grid characteristics are insufficient for dynamic voltage control. Capturing how nodes respond to faults is crucial for creating robust partitions.
  • Modularity as an Optimization Goal: Leveraging community detection algorithms and modularity maximization provides a principled way to find optimal partitions that balance internal cohesion and external decoupling.
  • Automatic Partition Number Determination: The modularity optimization approach eliminates the need for pre-defining the number of partitions, making the method more adaptable.
  • Electrical Coupling Strength as a Pre-Partitioning Guide: Using electrical coupling strength helps to create meaningful initial partitions, preventing unrealistic groupings of nodes that might appear similar topologically but behave differently electrically.

Key Results

The proposed two-stage dynamic partitioning strategy demonstrated significant improvements compared to existing transient voltage-based and conventional complex network-based methods on the modified IEEE 39-bus system:

  • Modularity Improvement: The proposed method achieved the highest modularity, with an improvement of 69% compared to other methods.
  • Reduced Voltage Deviation: The maximum post-fault voltage deviation was reduced by 38.6%, indicating more consistent voltage behavior within partitions.
  • Highest Regional Decoupling Rate: The method achieved the highest regional decoupling rate (0.2654), signifying strong internal coupling within regions and weak coupling between regions.
  • Effective Partitioning: The results show strong intra-regional cohesion and weak inter-regional connectivity, validating the strategy's effectiveness in enhancing adaptability and decoupling under transient conditions.

Limitations and Future Work

The authors acknowledge several limitations and suggest future research directions:

  • Limited Fault Scenarios: The current study focuses on single-contingency faults. Future work should extend the method to handle multiple or cascading failures to improve robustness.
  • Computational Burden: Analyzing diverse fault types, locations, and operating conditions for large-scale networks can be computationally intensive. Future research could explore methods for selecting representative fault sets to reduce data dimensionality and computational load.
  • Ideal Data Assumption: The method assumes perfect data availability. Real-world applications may face data uncertainty. Future work could incorporate data uncertainty modeling and robust partitioning strategies.
  • Large-Scale Systems: The paper suggests applying the method to larger power systems, which will require efficient filtering and selection of severe fault scenarios to manage computational complexity and data dimensionality.

Summary updated

Abstract

To enhance the adaptability of grid partitioning under transient scenarios, this paper proposes a two-stage dynamic partitioning strategy based on structure–function coupling. Electrical coupling strength is first characterized using short-circuit impedance and the sensitivity between reactive power and voltage, while transient voltage correlation is incorporated through cosine similarity as edge weights in a graph model. Grid partitioning is then conducted by maximizing modularity through a staged approach that ensures network connectivity and automatically determines partition numbers. Case studies on the modified IEEE 39-bus system demonstrate that compared with transient voltage-based partitioning and conventional complex network methods, the proposed approach improves modularity by 69%, reduces the maximum post-fault voltage deviation by 38.6%, and achieves the highest regional decoupling rate. The result shows strong intra-regional cohesion and weak inter-regional connectivity, verifying the strategy’s effectiveness in enhancing adaptability and decoupling under transient conditions.
BibTeX
@inproceedings{sun2025two,
  author = {Lixia Sun and Xianxue Sha and Shuo Zhang and Jiahao Wang and Yiping Yu},
  title = {{Two-Stage Dynamic Partitioning Strategy Based on Grid Structure Feature and Node Voltage Characteristics for Power Systems}},
  booktitle = {Energies},
  year = {2025},
  doi = {10.3390/en18102544},
}
Figure