摘要
故障检测和诊断 (FDD)技术目前已发展为四大类 ,适用于电网故障检测和诊断的有基于状态估算的FDD法和基于人工智能的故障诊断法。基于状态估算的FDD法 ,直接反映被控系统的运行状态 ,通过重构被控过程状态 ,构成残差序列 ,再通过构造模型 ,采用统计检验法 ,并利用状态观测器、滤波器等把故障检测出来 ;基于人工智能的故障诊断法又细分为专家系统和故障树等诊断法 ,基于专家系统的故障诊断法 ,主要利用计算机的推理能力、电力专家的丰富经验及电网中的因果关系 ,快速找到故障 ;基于故障树的诊断法是一种常用、有效的故障诊断法 ,通过逐级提问 ,构成一个递阶故障树 ,再通过启发性搜索 ,查找到故障。总之 ,运用自动控制中的故障检测和诊断技术来提高电网中计算机在线应用中实时数据库的有效性和可靠性 ,从而成为电力系统遥测校正和状态估算的重要辅助手段。
The fault detection diagnostics technique (FDD) has been developed into four kinds, while FDD method based on state estimation and fault diagnostics method based on artificial intelligence are suitable for power network fault detection and diagnostics. FDD based on the state estimation reflects directly the operation state of the controlled system, in which the residual error sequence is composed through re-constitution of the controlled process state, then through the structural model, by means of statistic inspection, state observer, wave filer and so on the fault will be detected. The fault diagnostics method based on artificial intelligence is divided into expert system and fault tree, etc.. The fault diagnostics based on the expert system mainly uses the derivation capability of computer, rich experiences of the power experts and causality in the power network to quickly discover the faults. The fault diagnostics based on the fault tree is a kind of normal and effective method. It discovers the fault through inquisition step by step to constitute a gradient step fault tree and then through the inspired searching. In general speaking FDD technique in automatic control is used to improve the effectiveness and reliability of real-time database for application of computer on-line in power network, which will become an important additional measures for remote measurement calibration and state estimation in power system.
出处
《电力建设》
北大核心
2003年第3期39-41,共3页
Electric Power Construction
关键词
故障检测和诊断
遥测校正
状态估算
电网
fault detection diagnostics
remote measurement calibration
state estimation
power network
