Anomaly Detection and Classification Method for Loop Resistance Testing Signals Based on the Attention Mechanism

Loop resistance testing is an essential tool for the maintenance and fault detection of electrical equipment. However, traditional detection methods often struggle with low accuracy and are susceptible to human errors when handling complex signals. To address this issue, this paper proposes an integrated model based on Convolutional Neural Networks (CNN), Long Short-Term Memory Networks (LSTM), and the Attention mechanism for anomaly detection and classification of loop resistance testing signals. The model first extracts local features of the signal using CNN, then captures the temporal dependencies of the signal using LSTM, and finally emphasizes key information in the signal through the Attention mechanism. Experimental results demonstrate that the proposed model significantly enhances the accuracy of anomaly detection in loop resistance testing signals, particularly in identifying anomalies such as short circuits and insulation faults. Compared to traditional machine learning methods, the proposed model offers clear advantages in both detection accuracy and robustness, providing effective support for the safe detection of electrical equipment. The experimental results validate the potential and prospects of this method in practical applications, especially for monitoring complex and dynamically changing electrical signals.