摘要
以光伏系统中直流串联电弧的检测为研究对象,基于自编码器模型提出了一种通用且复杂度相对较低的故障电弧检测方法。首先,通过MATLAB/Simulink仿真进行理论预演,并构建了模拟实际光伏系统应用场景的实验平台。该实验平台为深度学习算法生成训练数据,并从实验正常运行状态数据中提取编码特征。然后,使用训练完成的深度学习模型,对实验平台生成的故障电弧数据进行分析。通过各项指标评价,验证了该算法在光伏系统中检测直流串联电弧的性能和有效性。实验结果表明,该方法能够准确地检测光伏系统中的故障电弧,为提高系统安全性和防止火灾危险提供了可行的解决方案。
This paper focuses on the detection of DC series arc faults in photovoltaic systems.A general and relatively low-complexity fault arc detection method is proposed based on the autoencoder deep learning model.Firstly,theoretical simulations are conducted using Matlab/Simulink,and an experimental platform is constructed to simulate real-world scenarios of photovoltaic system applications.This experimental platform generates training data for the deep learning algorithm and extracts encoding features from normal operation data.Then,the trained deep learning model is used to analyze the generated fault arc data from the experimental platform.Through the evaluation of various indicators,the performance and effectiveness of the algorithm in detecting DC series arc faults in photovoltaic systems are validated.
出处
《工业控制计算机》
2024年第5期58-59,62,共3页
Industrial Control Computer
