Research Article
Research on Feature Extraction and Recognition of Inverter Fault Data Based on Neural Networks
@INPROCEEDINGS{10.1007/978-3-031-65126-7_30, author={Jingpeng Hu and Zhiguo Xiong}, title={Research on Feature Extraction and Recognition of Inverter Fault Data Based on Neural Networks}, proceedings={Quality, Reliability, Security and Robustness in Heterogeneous Systems. 19th EAI International Conference, QShine 2023, Shenzhen, China, October 8 -- 9, 2023, Proceedings, Part I}, proceedings_a={QSHINE}, year={2024}, month={8}, keywords={Neural network Inverter failure Data feature recognition}, doi={10.1007/978-3-031-65126-7_30} }- Jingpeng Hu
Zhiguo Xiong
Year: 2024
Research on Feature Extraction and Recognition of Inverter Fault Data Based on Neural Networks
QSHINE
Springer
DOI: 10.1007/978-3-031-65126-7_30
Abstract
This article proposes a fault detection method for cascaded inverters that combines digital signal processing technology and neural networks. This method determines the location and type of faults by detecting and analyzing the output voltage of the inverter. Fast Fourier transform (FFT) is used to analyze the frequency spectrum of output voltage signal to extract fault characteristics. By simplifying input data through Principal Component Analysis (PCA), the structure of neural networks can be improved. The feature recognition technology of inverter fault data based on neural networks can timely and effectively improve training speed and generalization accuracy.
Copyright © 2023–2025 ICST
