Malicious Mode Attack on EV Coordinated Charging Load and MIADRC Defense Strategy

Yichen Zhou, Weidong Liu, Jing Ma, Xinghao Zhen, Yonggang LiarXiv 2021 · arXiv

Abstract

As electric vehicles (EVs) become integral to the smart grid, their coordinated charging systems present a new cyber-physical attack surface. This paper introduces the Malicious Mode Attack (MMA), a coordinated IoT-based threat that induces grid resonance through periodic power fluctuations. To mitigate this, a Multi-index Information Active Disturbance Rejection Control (MIADRC) strategy is proposed. Simulations and hardware-in-the-loop tests demonstrate that MIADRC can suppress malicious oscillations with high accuracy, ensuring grid stability even under heavy load conditions.

Introduction

The transition to a sustainable energy future relies heavily on the integration of Electric Vehicles (EVs) as flexible grid assets. Managed via Coordinated Charging Load Control Systems (CCLCS) and the Internet of Things (IoT), these vehicles can optimize energy consumption. However, this connectivity introduces significant vulnerabilities. This research uncovers the Malicious Mode Attack (MMA), where attackers manipulate thousands of EV chargers to pulse power at the grid's natural resonant frequency, potentially leading to catastrophic instability.

The Malicious Mode Attack (MMA)

The MMA is a sophisticated, distributed cyber-threat. Unlike traditional attacks that target single points of failure, MMA leverages a botnet-like infection of the CCLCS to coordinate thousands of devices.

The three stages of a Malicious Mode Attack: Penetration, Infection, and Attack.

As illustrated in Figure 3, the attack unfolds in three distinct phases:

  1. Penetration: The attacker identifies weak-password energy routers and gains initial access to the network.
  2. Infection: A malicious loader program spreads to other vulnerable devices, eventually downloading a specialized Trojan from a remote file server.
  3. Attack: A central control service sends instructions to the infected devices, forcing them to fluctuate their power draw in a sinusoidal pattern.

If the frequency of these fluctuations matches the grid's "weakly damped electromechanical modes," it triggers high-amplitude forced oscillations. This resonance effect can cause system-wide instability, similar to how rhythmic movements can compromise structural integrity in physical bridges.

Characterizing the Threat

The authors derived five "laws" governing MMA, concluding that the most lethal scenarios occur when the malicious frequency aligns perfectly with the grid's natural oscillation frequency. Because the attack is distributed across thousands of chargers, localizing and disconnecting the source in real-time is practically impossible for grid operators. This necessitates a defense that makes the grid robust enough to ignore the attack rather than just blocking it.

Defense Strategy: MIADRC

To counter MMA, the paper proposes a Multi-index Information Active Disturbance Rejection Control (MIADRC) strategy. Traditional controllers often fail because they cannot accurately model the complex, dynamic nature of a large-scale power grid in real-time.

The MIADRC structure and its operational flow chart for suppressing MMA-induced oscillations.

As shown in Figure 4, MIADRC is a decentralized approach that treats the malicious attack as an external disturbance to be rejected. Key features include:

  • Multi-Index Observation: By integrating multiple signals—voltage, frequency, and power—the controller gains a comprehensive view of the grid's state, allowing it to "see" the hidden malicious commands.
  • Active Compensation: The system uses an Extended State Observer (ESO) to estimate the disturbance in real-time and generates a counter-signal to neutralize the oscillations before they can resonate.
  • Self-Healing Logic: The control flow involves real-time identification; once an MMA is detected, the MIADRC is put into operation to compensate for the disturbance and suppress forced oscillations.

Experimental Results and Validation

The effectiveness of the MMA model and the MIADRC defense was verified using standard power system benchmarks (IEEE 4-machine and 10-machine systems) and a custom Hardware-in-the-Loop (HIL) testbed.

  • Model Accuracy: Hardware tests closely mirrored theoretical simulations with error rates below 1%, confirming the physical feasibility of the MMA threat.
  • Superior Suppression: In the New England 10-machine system, the proposed MIADRC strategy achieved a 91.8% suppression rate of forced oscillations.
  • Adaptability: MIADRC outperformed standard Wide-Area Power System Stabilizers (WAPSS), particularly under "tense" grid conditions (heavy load) where the grid is naturally more prone to shaking.

Conclusion and Future Outlook

The Malicious Mode Attack represents a critical shift in cyber-threats against the smart grid, moving from data theft to physical resonance-based destruction. The proposed MIADRC strategy provides a robust, "passive-active" defense, enabling the grid to maintain stability by actively rejecting malicious disturbances. While highly effective, future work is needed to reduce the identification lag (currently ~4 seconds) and integrate these physical defenses with cyber-intrusion detection systems for a holistic security framework.

Summary updated

BibTeX
@misc{zhou2021malicious,
  author = {Yichen Zhou and Weidong Liu and Jing Ma and Xinghao Zhen and Yonggang Li},
  title = {{Malicious Mode Attack on EV Coordinated Charging Load and MIADRC Defense Strategy}},
  year = {2021},
  eprint = {2110.13681},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2110.13681},
}
Figure