摘要
为提高电力系统中故障诊断的效率,文中基于人工智能技术,开发了一套电力系统故障诊断系统。该系统利用人工智能技术中的深度置信网络,采用先预训练和微调参数的方式构建了电力系统故障诊断模型。搭配网络系数约束和网络平滑约束,以便突出连接矩阵中部分重要的连接,以辅助限制波尔兹曼机抓住暂态故障的局部特征,提高故障识别能力。测试表明,本系统能够准确识别电力系统中设备故障的种类,评估准确率较高,具有较强的时间优势,能有效推进电网信息化的发展。
In order to improve the efficiency of fault diagnosis in power system,this paper develops a power system fault diagnosis system based on artificial intelligence technology.The system uses the deep confidence network in artificial intelligence technology to construct a power system fault diagnosis model by pre-training and fine-tuning parameters.Network coefficient constraints and network smoothing constraints are provided to highlight some important connections in the connection matrix to assist the RBM in capturing local features of transient faults and improving fault identification.Tests show that the system can accurately identify the types of faults of various equipment in the power system,higher evaluation accuracy,and has obvious time advantages,effectively promoting the development of grid information.
作者
王哲
刘梓健
邱宇
WANG Zhe;LIU Zi⁃jian;QIU Yu(Information Center,Guangdong Power Grid Co.,Ltd.,Guangzhou 510000,China)
出处
《电子设计工程》
2020年第2期148-151,156,共5页
Electronic Design Engineering
关键词
故障诊断
人工智能
深度置信网络
电力系统
fault classification
artificial intelligence
deep trust network
power systems