Proposed algorithm for smart grid DDoS detection based on deep learning

Area of Science:

• Cybersecurity
• Electrical Engineering
• Artificial Intelligence

Background:

• The Smart Grid relies on digital technology for enhanced dependability, security, and efficiency.
• Real-time analysis and state estimation are crucial for effective Smart Grid control.
• Smart Grid communication infrastructure is vulnerable to cyber-attacks, threatening grid availability.

Purpose of the Study:

• To develop an effective intrusion detection algorithm to mitigate cyber-attacks on Smart Grid communication infrastructure.
• To address the specific threat of Distributed Denial of Service (DDoS) attacks.
• To improve the overall security and reliability of the Smart Grid.

Main Methods:

• Proposed a hybrid deep learning algorithm integrating Convolutional Neural Network (CNN) and Gated Recurrent Unit (GRU).
• Focused on detecting Distributed Denial of Service (DDoS) attacks targeting Smart Grid communication systems.
• Validated the algorithm using a benchmark cybersecurity dataset from the Canadian Institute of Cybersecurity Intrusion Detection System.

Main Results:

• The proposed hybrid deep learning algorithm demonstrated superior performance compared to existing intrusion detection methods.
• Achieved a high overall accuracy rate of 99.7% in detecting cyber intrusions.
• Effectively identified Distributed Denial of Service (DDoS) attacks within the Smart Grid communication network.

Conclusions:

• The hybrid CNN-GRU deep learning model offers a robust solution for detecting cyber-attacks in Smart Grid environments.
• This advanced intrusion detection system significantly enhances the security and resilience of critical energy infrastructure.
• The findings highlight the potential of deep learning for safeguarding the Smart Grid against evolving cyber threats.