Power Intelligent Terminal Intrusion Detection Based on Deep Learning and Cloud Computing

Area of Science:

• Computer Science
• Electrical Engineering
• Cybersecurity

Background:

• Power systems face increasing internal and external cyber threats, jeopardizing national economic infrastructure.
• Ensuring the information security of power networks requires high-standard security measures beyond general computer information security.

Purpose of the Study:

• To analyze intrusion threats in power information networks.
• To develop and investigate an effective intrusion detection technology for power systems.
• To enhance the security and reliability of critical power infrastructure.

Main Methods:

• Deep learning methods were employed to analyze the structure of power knowledge networks and cloud computing environments.
• A novel network intrusion detection model was developed, integrating abuse detection and anomaly detection techniques.
• Big data network data retrieval and analysis were accelerated using deep learning on data components.
• A fuzzy integral method was utilized to optimize intrusion prediction accuracy.

Main Results:

• The proposed model effectively detects new attack variants, overcoming limitations of traditional abuse analysis models.
• Rapid and accurate retrieval and analysis of big data network traffic were achieved.
• The intrusion prediction accuracy for the power information network reached 98.11%, an improvement of 0.6%.

Conclusions:

• The developed deep learning-based intrusion detection model significantly enhances the security of power information networks.
• The integrated approach of abuse and anomaly detection provides robust protection against evolving cyber threats.
• Optimized prediction accuracy ensures a more reliable and secure power system operation.