Dynamic Integral-Event-Triggered Control of Photovoltaic Microgrids with Multimodal Deception Attacks

Zehao Dou, Liming Ding, Shen YanSymmetry 2025 · DOI

TL;DR

This paper introduces a novel control strategy called Dynamic Integral-Event-Triggered Scheme (DIETS) for photovoltaic (PV) microgrids. DIETS significantly reduces communication load by intelligently triggering control updates only when necessary, while also providing enhanced security against sophisticated deception attacks.

What Problem Does This Paper Address?

Modern microgrids, especially those integrating renewable energy sources like PV, rely heavily on communication networks for coordinated control and optimization. However, this reliance creates two major challenges:

  1. Communication Bandwidth Limitations: Traditional periodic communication is inefficient, wasting network resources. While event-triggered control (ETC) offers a solution by communicating only when needed, existing ETC methods can still be prone to redundant triggers due to noise or disturbances, and their fixed thresholds can lead to delays.
  2. Cybersecurity Threats: Communication networks are vulnerable to attacks. Deception attacks, in particular, can manipulate sensor data or control commands, leading to incorrect system decisions, instability, and even collapse of the microgrid. Existing security measures often struggle against multimodal deception attacks that mimic legitimate data.

This paper aims to address both these issues by developing a control strategy that is both communication-efficient and robust against advanced cyber threats.

Methods

The authors propose a two-pronged approach:

  1. Dynamic Integral-Event-Triggered Scheme (DIETS):
    • Average Historical State Processing: Instead of relying solely on instantaneous system states, DIETS uses the average of historical state data. This smooths out variations caused by noise or transient disturbances, reducing unnecessary triggering events.
  2. Adaptive Triggering Threshold: A time-varying triggering threshold is designed, integrating real-time system states and historical trends. This allows the threshold to adapt dynamically, further optimizing triggering and preventing Zeno behavior.
  3. Mathematical Foundation: The DIETS is formulated using mathematical expressions that define the triggering condition based on the discrepancy between integrated historical states and triggered values, incorporating a weighting matrix and a memory interval.
  4. Secure Control Strategy against Multimodal Deception Attacks:
    • Bernoulli Stochastic Process: Multimodal deception attacks are modeled using a Bernoulli stochastic process to capture their time-varying nature and different attack modes.
  5. Lyapunov-Krasovskii Functional and Bessel-Legendre Inequality: These advanced mathematical tools are employed to analyze system stability under attack. The Bessel-Legendre inequality is used to handle integral terms arising from historical states, aiming to reduce conservatism in stability analysis.
  6. Co-design of Controller and Triggering Matrix: The controller gain and triggering matrix are designed together as symmetric linear matrix inequalities (LMIs) to ensure system stability and security.

Datasets

Not discussed in the paper. The paper uses simulated data based on a PV microgrid model.

Key Insights and Ideas

  • Beyond Instantaneous States: The core innovation lies in moving beyond instantaneous system states for triggering. Averaging historical data is a simple yet effective way to filter out noise and reduce communication load without sacrificing control performance.
  • Adaptive Thresholds are Crucial: A fixed triggering threshold is a bottleneck. The dynamic, adaptive threshold in DIETS allows for more intelligent triggering, balancing communication efficiency with control responsiveness.
  • Multimodal Attacks Require Sophisticated Defense: The paper highlights that simple deception attack models are insufficient. The use of a Bernoulli process to model multiple attack modes and their probabilistic nature is a step towards more realistic cyber-attack scenarios.
  • Integrated Design for Security and Efficiency: The co-design of the controller and triggering mechanism is key. This ensures that the system's communication efficiency efforts don't compromise its ability to withstand attacks, and vice-versa.

Key Results

The paper demonstrates significant improvements in communication efficiency and system robustness:

  • Communication Efficiency: DIETS achieved a substantial reduction in triggering events. It recorded only 99 triggering events, which is a 52.6% reduction compared to dynamic ETC and a 55.2% reduction compared to normal ETC. The average trigger interval was also significantly longer (0.05s for DIETS vs. 0.024s for dynamic ETS and 0.023s for normal ETS).
  • Security Performance: The proposed secure controller effectively stabilized the PV microgrid under multimodal deception attacks. State responses under attack with the proposed controller were significantly better than those with a standard controller, which showed misjudgments and ineffective control strategies.
  • Controller vs. Standard: When comparing the proposed secure controller against a standard controller (without specific deception attack handling) under deception attacks, the secure controller reduced triggering events by 16.8% (99 vs. 119 events) while maintaining better state responses.

Limitations and Future Work

The authors acknowledge several limitations and suggest future research directions:

  • Time-Delay Issues: The current work does not explicitly model or analyze time delays in communication networks, which are critical in practical microgrid systems. Future work will incorporate delay-dependent analysis.
  • Static Attack Models: The multimodal deception attack model is based on static patterns and prior assumptions. Future research will explore adaptive learning-based attack models to design more sophisticated and dynamic attack strategies.
  • Load Frequency Control (LFC): The authors plan to extend their research to LFC in large-scale power grids, considering random delays and cyberattacks.
  • Hardware Constraints: The current study is primarily theoretical. Future work will involve validating the proposed strategies through hardware-in-the-loop experiments and exploring multi-node scalability challenges.

Summary updated

BibTeX
@article{dou2025dynamic,
  author = {Zehao Dou and Liming Ding and Shen Yan},
  title = {{Dynamic Integral-Event-Triggered Control of Photovoltaic Microgrids with Multimodal Deception Attacks}},
  journal = {Symmetry},
  year = {2025},
  doi = {10.3390/sym17060838},
}
Figure