Research on current transformer signal recognition method based on convolutional network

With the rapid development of smart grid and industrial automation, the accuracy and reliability of current transformers as key components in the power system have become increasingly important. This article proposes a current transformer signal recognition method based on convolutional neural networks, aiming to improve the efficiency and accuracy of distorted signal recognition. By systematically studying the key components of neural networks such as convolutional layers, pooling layers, Flatten layers, fully connected layers, BN layers, and Dropout layers, an efficient signal feature extraction method has been designed. Additionally, a distortion signal model for current transformers has been constructed, and a denoising solution has been proposed to reduce interference factors during the signal processing. The experimental section validates the effectiveness and stability of the proposed method in identifying distorted signals of current transformers through a series of tests in a simulated environment. The results demonstrate that the new method achieves high recognition accuracy in various complex scenarios, providing technical support for the intelligent operation and maintenance of power systems.