摘要
为提高故障自动判别结果的可靠度与准确度,优化电气系统运行,本文引进改进LSTM,以某电气系统为例,开展绝缘老化故障自动判定方法的设计与研究。首先,将改进LSTM感知单元作为一个呈离散化特点的非线性时变动态网络,在网络前端输入电气系统运行中的时序信号,根据输出的信号构建电气系统的状态空间模型,实现基于改进LSTM的电气系统运行状态感知;其次,结合电气系统运行状态,提取电气系统绝缘老化故障特征分量;最后,引进支持向量机,自动判定系统绝缘老化故障,划分故障类别。实验结果表明,该方法实际应用效果良好,按照规范自动判定电气系统绝缘老化故障,可提高判定结果的置信度、准确度。
In order to improve the reliability and accuracy of automatic fault identification results and optimize the operation of electrical system,this paper introduces an improved LSTM and takes an electrical system as example to carry out the design and research of automatic insulation aging fault determination method.Firstly,the improved LSTM sensing unit is regarded as a nonlinear time-varying dynamic network with discretization characteristics,the time sequence signal of the electrical system is input in the front end of the network,and the state space model of the electrical system is constructed according to the output signal,realize the operating state perception of the electrical system based on the improved LSTM.Secondly,combined with the running state of electrical system,the characteristic components of insulation aging fault of electrical system are extracted.Finally,support vector machine is introduced to determine the insulation aging fault automatically and divide the fault categories.The experimental results show that the design method has good effect in practice,and according to the specification automatic determination of insulation aging fault in electrical system can improve the confidence and accuracy of the judgment result.
作者
张瑶桐
ZHANG Yaotong(Inner Mongolia Electric Power Survey and Design Institute Co.,Ltd.,Hohhot,Inner Mongolia 010010,China)
出处
《自动化应用》
2023年第22期47-49,共3页
Automation Application
关键词
改进LSTM
判定方法
自动
绝缘老化故障
电气系统
improved LSTM
decision method
automatic
insulation aging failure
electrical system