Hybrid deep learning-driven explainable AI framework for fault detection and classification in smart power grids

Area of Science:

• Electrical Engineering
• Artificial Intelligence
• Smart Grid Technology

Background:

• Stable operation and reduced downtime are critical for smart power grids.
• Effective fault detection is essential for grid reliability and efficiency.

Purpose of the Study:

• To propose a hybrid deep learning model for accurate and interpretable fault detection and classification in smart power grids.
• To capture spatial and temporal attributes of electrical signatures for enhanced fault analysis.

Main Methods:

• A hybrid model combining Convolutional Neural Networks (CNN) and Long Short-Term Memory (LSTM) with Explainable Artificial Intelligence (XAI).
• Utilized real-time sensor data and simulated fault conditions (LG, LL, LLG, three-phase faults).
• Incorporated SHAP and attention mechanisms for improved interpretability.

Main Results:

• Achieved 97.84% classification accuracy and 97.08% F1-score, surpassing conventional models by over 3%.
• Demonstrated stable performance in noisy environments with less than 2% performance decrease.
• Achieved a low inference latency of 18 ms, suitable for real-time deployment.

Conclusions:

• The hybrid CNN-LSTM-XAI model offers a powerful, scalable, and interpretable solution for intelligent fault management in smart grids.
• The model's ability to handle noisy data and provide real-time insights enhances grid stability and reduces operational risks.