Advanced mathematical modeling of mitigating security threats in smart grids through deep ensemble model

Area of Science:

• Cybersecurity
• Electrical Engineering
• Artificial Intelligence

Background:

• Smart grids (SGs) are vulnerable to cyberattacks, necessitating robust security measures.
• Intrusion detection systems (IDSs) are crucial for protecting SG data and ensuring a reliable power supply.
• Deep learning (DL) offers advanced capabilities for identifying complex attack patterns in SGs.

Purpose of the Study:

• To develop a novel intrusion detection technique for smart grids.
• To enhance the resilience and security of smart grid environments against cyber threats.
• To leverage deep learning and metaheuristic optimization for improved intrusion detection.

Main Methods:

• A new Mountain Gazelle Optimization with Deep Ensemble Learning based intrusion detection (MGODEL-ID) technique was developed.
• Data preprocessing involved Z-score normalization, and feature selection utilized the Mountain Gazelle Optimization (MGO) model.
• Intrusion detection was performed using an ensemble of Long Short-Term Memory (LSTM), Deep Autoencoder (DAE), and Extreme Learning Machine (ELM) classifiers, with hyperparameter tuning by the Dung Beetle Optimizer (DBO).

Main Results:

• The MGODEL-ID model demonstrated superior performance in detecting intrusions within the smart grid environment.
• Simulation results confirmed the enhanced security outcomes achieved by the proposed MGODEL-ID technique.
• Experimental validation indicated that MGODEL-ID outperforms existing intrusion detection models.

Conclusions:

• The MGODEL-ID technique provides a robust and effective solution for smart grid cybersecurity.
• The integration of deep learning and metaheuristic optimization significantly enhances intrusion detection capabilities.
• The developed model contributes to the integrity, resilience, and reliability of the electricity supply chain.