电力物联网背景下基于HHT-CNN的智能变电站故障诊断

Fault diagnosis of smart substation based on HHT-CNN under the background of power internet of things

在电力物联网的背景下,为提高传统智能变电站故障诊断能力,提出基于希尔伯特-黄和卷积神经网络相融合的智能变电站故障诊断方法.将智能变电站中的故障录波数据作为故障诊断数据,利用希尔伯特-黄变换提取综合电流的故障特征,通过训练好的卷积神经网络实行故障定位.以典型的110 kV智能变电站为例进行仿真测试,测试结果表明:增加数据增强模块能有效提高卷积神经网络模型的泛化能力;选择合适的卷积神经网络模型参数能有效提高故障诊断正确率和降低训练时间;相对于其他2种方法,该方法有较高的故障诊断正确率.

Under the background of power internet of things,in order to improve the traditional fault diagnosis technical capability of smart substation,a fault diagnosis method of smart substation based on convolutional neural network was proposed.The fault recording data in smart substation was used as fault diagnosis data.Then fault characteristics of comprehensive current were extracted by Hilbert-Huang transform.Finally,the fault was located by using the trained convolutional neural network.A typical 110 kV smart substation was taken as an example,the test results showed that the whole fault diagnosis model had strong generalization ability by adding data enhancement module in the training process of convolutional neural network.And the fault diagnosis accuracy was improved and the training time was reduced by selecting appropriate convolutional neural network model parameters.In addition,compared with the other two methods,the method proposed in this paper had higher fault diagnosis accuracy.