摘要
提出了一种基于稳健回归的绝缘子污闪试验数据处理方法:根据试验电压与盐密和灰密数据,通过复加权最小二乘迭代算法求解回归系数;迭代计算中采用权重函数,权重系数为上次迭代的残差的函数,以此减少奇异值对回归系数的影响。通过比较试验数据与稳健回归及最小二乘回归的预测结果,验证了基于稳健回归方法在绝缘子污闪试验数据处理中的有效性。
A method to predict insulator pollution flashover voltage based on robust regression was proposed. According to testing voltage and the data of salt deposit density （SDD） and non-soluble deposit density （NSDD）, the regression coefficient was solved by complex weighting least square iteration algorithm. In iterative calculation the weight function was adopted and the weight coefficient was the function of the residual error of last iteration to weaken the influence of singular values on regression coefficient. Via comparing the prediction results of the data from tests, robust regression and least square regression respectively, this paper verifies effectiveness of the proposed robust regression-based predicting method.
出处
《电工电气》
2011年第4期51-54,共4页
Electrotechnics Electric
基金
浙江省电力公司重点科技项目(ZDK069-2010)
关键词
绝缘子
污闪
稳健回归
数据处理
insulator
pollution flashover
robust regression
data processing
