基于小波变换和人工智能的电力系统故障诊断

Fault Diagnosis of Power System Based on Wavelet Transform and Artificial Intelligence

针对电力系统故障诊断的问题,研究了一种基于小波变换和人工智能的方法。先介绍了电力系统故障诊断方法的基本框架,随后重点研究了基于小波变换的时频域特征提取和基于卷积神经网络的集成学习方法。在实验中,利用MATLAB平台构建了一个模拟数据集,并对传统卷积神经网络(Convolutional Neural Network,CNN)和该方法在准确率、召回率以及F1值上的性能进行了比较。结果表明,新方法相对于传统CNN在所有指标上均取得了显著的提升,表明其在电力系统故障诊断中具有更高的正确性和有效性。

Focuses on the problem of fault diagnosis in power systems and investigates a method based on wavelet transform and artificial intelligence.Firstly,the basic framework of fault diagnosis methods for power systems was introduced,followed by a focus on time-frequency domain feature extraction based on wavelet transform and ensemble learning methods based on convolutional neural networks.In the experiment,a simulated dataset was constructed using the MATLAB platform,and the performance of traditional Convolutional Neural Network(CNN)and this method in terms of accuracy,recall,and F1 value was compared.The results show that the new method has achieved significant improvements in all indicators compared to traditional CNN,indicating that it has higher accuracy and effectiveness in power system fault diagnosis.