利索-全球专利检索

  1. 登录
  2. 注册
发明公开
CN108020759A 一种基于PSOGSA神经网络的XLPE电缆振荡波故障识别方法 无效 - 驳回

一种基于PSOGSA神经网络的XLPE电缆振荡波故障识别方法
摘要：
本发明提出了一种基于PSOGSA神经网络的XLPE电缆振荡波故障识别方法，1)设计4种XLPE电缆绝缘缺陷，并采集局放信号，然后进行小波包变换，得到各子带系数图；2)提取并计算各子带系数的归一化能量值，将其作为特征向量；3)将上述特征参数输入到PSOGSA神经网络模型，并进行类型识别，即输出相应的XLPE电缆振荡波故障类型。
摘要（英）：

The invention provides an XLPE cable oscillation wave fault recognition method based on a PSOGSA neural network, which comprises the steps of 1) designing four XLPE cable insulation defects, collecting partial discharge signals, and then performing wavelet packet transform to obtain sub-band coefficient maps; 2) extracting and calculating a normalized energy value of each sub-band coefficient, andenabling each normalized energy value to serve as a feature vector; and 3) inputting the feature parameters into a PSOGSA neural network model, performing type recognition, and outputting the corresponding XLPE cable oscillation wave fault type.

0/0