基于改进灰自助法的可靠性原始参数预测  

Forecast of Original Reliability Parameter Based on Improved Grey Bootstrap

作　　者：陈柔伊 许亮[2] 刘希喆[2] 

机构地区：[1]南方电网科学研究院,广东广州510080 [2]华南理工大学电力学院,广东广州510640

出　　处：《广东电力》2012年第3期5-9,39,共6页Guangdong Electric Power

基　　金：广东省部产学研结合项目资助(2010B090400137)

摘　　要：针对电力系统可靠性评估中原始参数缺乏的问题,提出原始参数预测的改进灰自助法,它将GM(1,1)模型与自助(Bootstrap)法相结合,利用Bootstrap法产生大量再生样本,用改进灰色模型来减小误差,避免小样本特别是极小样本情况下自助法可信度不高的问题。采用改进灰自助法对算例进行仿真结果表明:在点估计、区间估计以及预测的置信度和可信度上都优于Bootstrap法以及传统的极大似然法,可信度和实用性高,解决电力系统可靠性评估解决原始数据不足的问题。Aiming at inadequacy of original parameter in reliability assessment of power system,the paper introduces grey bootstrap for original parameter forecast,which combines GM（1,1） model with Bootstrap,generates abundant regeneration samples by means of Bootstrap and reduces errors via improved grey Bootstrap,averting low reliability of Bootstrap in case of small samples,especially minimum ones.The example simulation through improved grey Bootstrap shows that it is superior to Bootstrap and maximum likelihood method in respects of point estimation,interval estimation and confidence and reliability of forecast.It is of high reliability and practicability and can cope with inadequacy of raw data in power system.

关 键 词：可靠性原始参数 小样本 BOOTSTRAP法 改进灰自助法 极大似然法 可信度 

分 类 号：TM711[电气工程—电力系统及自动化]