Modeling Method for the Coupling Relations of Microgrid Cyber-Physical Systems Driven by Hybrid Spatiotemporal Events

Xiaoyong Bo, Xiaoyu Chen, Huashun Li, Yunchang Dong, Zhaoyang Qu, Lei Wang, Yang Li

Abstract

Traditional microgrid models often treat cyber and physical layers as static, separate entities, failing to capture the dynamic interactions during events like cyber-attacks. This paper proposes a novel modeling framework driven by hybrid spatiotemporal events. Utilizing Multi-Agent Systems (MAS) and hybrid automata, the framework simulates the continuous "perceive-control" loop. Validation through False Data Injection Attack (FDIA) scenarios demonstrates the model's ability to track physical consequences of digital interference, offering a robust tool for microgrid security analysis.

Introduction

Microgrids are quintessential Cyber-Physical Systems (CPS), integrating a "cyber" layer of sensors, controllers, and communication networks with a "physical" layer of power generation and storage hardware. Most existing research relies on static network models that struggle to describe the dynamic, cyclical conversion between information flow and energy flow.

The layered architecture of a Microgrid CPS, showing the interaction between the cyber control layer and the physical power layer.

In real-world operations, delays or malicious data injections in the cyber layer ripple through the physical system, potentially causing cascading failures. This paper addresses this gap by creating a model that captures the real-time "perceive-control" loop, accounting for both spatial and temporal constraints.

Methodology

The proposed framework employs a multi-layered architecture centered on two core technologies:

  1. Multi-Agent Systems (MAS): Every physical component—such as wind turbines, photovoltaic arrays, or battery packs—is paired with a digital "agent." These agents are self-organizing and coordinate to make decentralized decisions.
  2. Hybrid Automata: This mathematical tool bridges the gap between discrete cyber events (packet transmissions) and continuous physical processes (voltage/current fluctuations).
Detailed schematic of the microgrid components and their corresponding agent modules, illustrating the decentralized control structure.

The coupling between layers is defined through a Four-Stage Event Chain:

  • State Perception: Sensors convert physical energy states into digital data.
  • Network Communication: Data is routed through the cyber network.
  • Intelligent Decision-Making: Agents process data to determine control strategies.
  • Action Control: Digital decisions are converted back into physical actions, such as adjusting inverter outputs or tripping breakers.

Experimental Setup

To validate the model, a high-fidelity co-simulation environment was developed:

  • RT-LAB: Simulates the continuous physical power system (3 generators, 3 loads).
  • OPNET: Simulates the discrete cyber network (3 routers, 6 agent nodes).
  • MATLAB/Simulink: Manages the underlying control logic.

This setup enables the injection of "False Data Injection Attacks" (FDIA) to observe system reactions in a controlled environment.

Results and Discussion

The model was tested under two primary attack scenarios:

  • Measurement Attack: Tampering with load current phase angles led agents to incorrectly identify faults, causing unnecessary system adjustments.
  • Control Command Attack: Injecting a "remote trip" command resulted in instantaneous load-shedding and grid instability.

Key insights include the necessity of modeling the "perceive-control" loop to understand system vulnerabilities. The research highlights that cyber-attacks are defined by spatiotemporal constraints—where and when the injection occurs is as critical as the data itself.

Conclusion

While effective for microgrids, scaling this model to larger smart grids remains a computational challenge. Future work will focus on integrating AI for autonomous "attack antigen" identification and expanding the framework into Cyber-Physical-Social Systems (CPSS) to account for human behavior in energy management.

Summary updated

BibTeX
@article{bo2021modeling,
  author = {Xiaoyong Bo and Xiaoyu Chen and Huashun Li and Yunchang Dong and Zhaoyang Qu and Lei Wang and Yang Li},
  title = {{Modeling Method for the Coupling Relations of Microgrid Cyber-Physical Systems Driven by Hybrid Spatiotemporal Events}},
  journal = {IEEE Access},
  year = {2021},
  doi = {10.1109/access.2021.3053402},
}
Figure