INTELLIGENT ELECTRICAL FAULT DETECTION AND CLASSIFICATION USING ENSEMBLE DEEP LEARNING ARCHITECTURE

Abstract

The electrical power transmission lines defect detection system uses deep learning algorithm techniques, taking with regard the rising need for electricity in contrast to the slow development of transmission capacity. This research introduces a unique solution to detect electrical faults utilizing an ensemble deep learning algorithm. This research proposed a deep learning-based ensemble ACG model that integrates Artificial Neural Networks (ANN), Convolutional Neural Networks (CNN), and Gated Recurrent Units (GRU) to combine feature compression, spatial feature extraction, and temporal pattern learning within a unified framework. In addition, several deep learning architectures, including Artificial Neural Networks (ANN), Recurrent Neural Networks (RNN), Long Short-Term Memory (LSTM), and Convolutional Recurrent Neural Networks (CRNN), are implemented and evaluated to analyze their effectiveness for electrical fault detection and classification. Sensor data collected from electrical systems are preprocessed, normalized, and encoded to assure the model has high-quality inputs. In particular, ACG model performance is examined using different datasets, including datasets comprising various fault types. The proposed ensemble ACG model obtained 99.66% accuracy, 99.54% precision, 99.72% recall, and 99.63% F1 Score. In addition, in order to validate ACG model performance, a K-fold cross validation was employed, and results indicated that ACG model performed better by obtaining 99.79% accuracy, 100% precision, 99.52% recall, and 99.76% F1 score. Experiment results indicate that proposed framework is more accurate, reliable, and precise than previous models utilized in the process of classifying and identifying faults.