| 摘要: |
| 将秩次回归应用于系统谐波阻抗分析与谐波评估。基于系统和用户的等效电路，可获得系统谐波阻抗、背景谐波和公共连接点谐波的解析表达式。利用在公共连接点同步测量得到的谐波电压、电流数据，将系统谐波阻抗和背景谐波作为回归方程的回归系数求出其最小二乘估计。使用残差秩次得分函数得到基于秩次回归的估计函数。采用最速下降法求取基于秩次回归估计函数的极小值即可获得基于秩次回归估计，并计算出用户谐波发射水平。该方法能够克服样本中异常数据对回归系数的影响，同时对误差项无正态性分布要求。仿真结果证明了该方法的有效性和准确性。 |
| 关键词: 谐波阻抗 谐波发射水平 秩次回归 得分函数 最速下降法 |
| DOI： |
| 分类号: |
| 基金项目:国家科技支撑计划（2007BAA12B05）；西南交通大学科研基金（2007A12） |
|
| Harmonic impedance analysis and harmonic assessment based on rank regression |
|
XIE Shaofeng
|
|
School of Electrical Engineering，Southwest Jiaotong University，Chengdu 610031，China
|
| Abstract: |
| The rank regression is proposed for harmonic impedance analysis and harmonic assessment. Based on the equivalent circuit of system and customer，the analytical expressions of system harmonic impedance， background harmonic and harmonic at PCC（Point of Common Coupling） are obtained. With the harmonic voltage and current measured at PCC，the least-square estimation of system harmonic impedance and background harmonic are calculated as the regression coefficients of regression equation. The rank regression based estimation function is deduced from the scoring function of residual rank. The steepest descent method is applied to calculate its minimum，and the rank-regression-based estimation and the harmonic emission level are thus calculated. With the proposed method，the impact of abnormal sample data on the regression coefficients is avoided and the normal distribution of errors is not required. Its effectiveness and correctness are verified by the simulative results. |
| Key words: harmonic impedance harmonic emission level rank regression scoring function steepest descent method |
