基于最值归一化的MMC-MVDC单极接地故障识别方法
A method for identifying single-pole ground faults in MMC-MVDC based on min-max normalization
摘要
针对小电流接地方式的辐射状模块化多电平中压直流(MMC-MVDC)系统发生单极接地故障时健全极电压升高的问题,本文提出一种基于数据归一化的故障检测方法.通过分析不同接地方式下的暂态电流回路与极电流突变量特征,指出直接利用极电流突变量进行故障检测的局限性.运用最值归一化技术,使其实现无量纲化,在保留波形形态特征的同时有效放大故障特征差异,据此构建选线选极准则.PSCAD 仿真验证,该方法仅利用 1 ms 数据窗便能准确识别线路与母线故障,不仅耐受高过渡电阻能力强,且能良好适应负荷波动、极短线路及低采样率等复杂工况,具备显著的工程应用价值.
This paper addresses the issue of increased healthy pole voltage in a radial modular multilevel medium voltage direct current(MMC-MVDC)distribution network with small current grounding when a single-pole ground-ing fault occurs.A fault detection method based on data normalization is proposed.By analyzing the transient cur-rent loop and the characteristics of pole current fault component under different grounding methods,the limitations of directly using pole current fault component for fault detection are pointed out.This method utilizes min-max nor-malization technology to render the quantities dimensionless.By doing so,it effectively amplifies the differences in fault characteristics while preserving the waveform morphological features,thereby establishing the criteria for fault line and pole selection.PSCAD simulation verification demonstrates that this method can accurately identify line and busbar faults using only a 1 ms data window.It not only exhibits a strong ability to withstand high transition re-sistance but also adapts well to complex conditions such as load fluctuations,extremely short lines,and low sam-pling rates,indicating significant value for engineering applications.
年,卷(期):
2026, 45(9)
分类号:
TM721.1
机标分类号:
TM930
TM773
TP391.41
在线出版日期:
2026年10月14日
基金项目:
2025JC-YBQN-642:陕西省自然科学基础研究计划项目2030)(2024ZD0802304:智能电网重大专项,