An artificial insurance framework for a hydrogen-based microgrid to detect the advanced cyberattack model.

Area of Science:

• Electrical Engineering
• Cybersecurity
• Renewable Energy Systems

Background:

• Microgrids offer flexibility but are vulnerable to cyberattacks, especially in grid-connected mode.
• A novel False Transferred Data Injection (FTDI) attack model manipulates power flow and voltage stability.
• Detecting FTDI attacks requires monitoring both the value and direction of power flow.

Purpose of the Study:

• To propose and evaluate a novel advanced attack model (FTDI) for microgrids.
• To develop a robust detection mechanism for FTDI attacks using a learning generative network.
• To enhance the cybersecurity of microgrid systems operating in grid-connected mode.

Main Methods:

• Development of a False Transferred Data Injection (FTDI) attack model.
• Implementation of a Generative Adversarial Network (GAN) based detection model.
• Testing on a 24-bus IEEE microgrid system with diverse renewable energy sources.

Main Results:

• The proposed GAN-based model effectively detects malicious data injection.
• Achieved high detection rates with scores of 0.95% (Hit rate), 0.92% (C.R. rate), 0.7% (F.A. rate), and 10% (Miss rate).
• Demonstrated the model's capability to identify changes in both power value and direction.

Conclusions:

• The GAN-based approach provides an effective solution for detecting FTDI attacks in microgrids.
• The study highlights the importance of considering power flow direction in cybersecurity detection.
• The proposed model significantly improves the resilience of microgrids against sophisticated cyber threats.