LNG船电力系统故障检测的仿真研究  

On Simulation of Fault Signal Detection for Electric Power Systems in LNG Carriers

作　　者：包艳[1] 施伟锋[1] 

机构地区：[1]上海海事大学物流工程学院,上海201306

出　　处：《计算机仿真》2014年第12期446-450,共5页Computer Simulation

基　　金：高等学校博士学科点专项科研基金(20123121110003);上海市教委科研创新重点项目(12ZZ155)

摘　　要：研究电力推进LNG船电力系统故障有效识别的问题。LNG船电力系统中含有大功率推进电机,其随机变化易造成电力系统故障,产生的故障暂态信号蕴含大量噪声,具有随机、非平稳的特点。传统方法不能有效提取这类故障信号特征,故障检测准确度低。为解决上述问题,提出了一种基于聚类经验模型分解(Ensemble Empirical Mode Decomposition,EEMD)的希尔伯特-黄变换(Hilbert-Huang Transform,HHT)故障检测方法。首先,将故障时刻的电网电压信号进行EEMD分解,得到固有模态函数分量;然后,将上述分量的希尔伯特边际谱进行时频分析,提取较为准确的故障特征信息。仿真结果表明,HHT方法能弥补传统信号分析方法的不足,最大限度的抑制噪声和保留故障信号特征,提髙故障检测准确率。Random changes of high power loads in LNG carriers propulsion motors will cause electric power sys- tem faults and even lead to the system crashes, and the fault transient signals are often random and non -stationary. Traditional fault detection methods cannot accurately extract these kinds of fault features. Based on the Ensemble Em- pirical Mode Decomposition （ EEMD）, a new fault detection method of Hilbert - Huang Transform （HHT） is presen- ted for electric power systems. The voltage signal is decomposed by EEMD at the time of failure, and the Intrinsic Mode Function （IMF） is obtained. By analyzing the Hilbert spectrum of timefrequency, the fault feature informa- tion can be extracted accurately. Simulation results show that the proposed method can make up for deficiencies in the traditional detection methods, suppress noise, reserve the fault signal feature, and improve the accuracy of fault detection.

关 键 词：电力推进船 特征抽取 聚类经验模型分解 希尔伯特-黄变换 故障诊断 

分 类 号：TM743[电气工程—电力系统及自动化]