摘要
电力系统状态估计是能量管理系统的核心基础模块,然而当量测系统中存在不良杠杆量测或多个强相关的不良数据时,传统的含不良数据辨识程序的最小二乘估计不能很好地排除不良数据对状态估计的影响。提出一种指数型目标函数电力系统抗差状态估计模型,该估计模型能够在状态估计过程中自动排除不良数据影响,无需不良数据辨识环节。可证明该模型等价于最小化Renyi二次熵定义下的信息损失。介绍该估计模型的理论基础,分析其数学特性。并结合一个简单的电力系统状态估计问题,研究该估计模型的局部最优点问题及其解决方法。
Power system state estimation is a crucial basic function in energy management system. The traditional weighted least square based state estimation method with bad data identification function cannot suppress bad data efficiently when bad leverage data or multiple interacting bad data exist. A robust state estimation model with an exponential objective function was proposed, which can reject bad data automatically with no need of extra bad data identification procedure. It had been verified that the proposed method is equivalent to minimizing information losses defined by Renyi quadratic entropy. Theoretical foundation and mathematical characteristics of the proposed method were well analyzed. Local optimum problem and its solution in application of the method were also studied.
出处
《中国电机工程学报》
EI
CSCD
北大核心
2011年第4期67-71,共5页
Proceedings of the CSEE
基金
国家自然科学基金项目(50707013)~~
关键词
电力系统
状态估计
RENYI熵
不良数据辨识
抗差估计
核密度估计
power system
state estimation
Renyi entropy
bad data identification
robust estimation
kernel density estimation
