基于相关性分布式卡尔曼滤波的谐波检测方法  

Harmonic detection method based on distributed related Kalman filter

作　　者：赵婵娟[1] 王建平[1] 孙伟[1] 朱程辉[1] 

机构地区：[1]合肥工业大学电气与自动化工程学院,合肥230009

出　　处：《电子测量与仪器学报》2016年第9期1333-1341,共9页Journal of Electronic Measurement and Instrumentation

基　　金：国家自然科学基金(51177034;51307041;51304058)资助项目

摘　　要：微电网谐波问题严重。针对微电网谐波动态特性复杂的特点,探索了一种基于相关性分布式卡尔曼滤波的微电网时变谐波检测方法。首先,针对微电网系统相邻母线的谐波在一定程度上是相关的特点,提出了代表邻居节点间相关程度的邻节点相关系数这一概念,并给出一种微电网谐波测量邻节点间的量测值和协方差矩阵逆的两个融合变量计算方法;其次,对邻节点间估计值作了分布式融合处理,从而保证了邻节点间估计结果具有一致相关性;再者,通过重构迭代过程,简化了该算法的计算过程。通过在IEEE-14节点系统的仿真实验结果表明,该方法可实现多个检测点分布协同地完成微电网多母线谐波电压的状态估计,且相比传统卡尔曼滤波算法具有更好的估计精度和抗扰动性能,实时性和动态性也表现出较好的性能。Harmonic interference of micro-grid system is serious. According to the dynamic characteristics of the micro-grid system harmonic, a distributed related Kalman filter method is presented for time-varying harmonic estimation. First, considering the fact that micro-grid system harmonic injection at neighbor bus is correlated to some extent, the concept of neighbor correlation coefficient is introduced to represent the degree of correlation between neighbor nodes. And then, a method to calculate the neighbor node fusion variable （the fused inverse- covariance matrices and the fused sensor data） which is suitable for harmonic measurements is proposed. Second, distributed fusion processing among the neighbor nodes of estimated values is utilized to ensure the unanimous correlation of the estimation results between neighbors. Finally, to simplify the calculation, the iterative process of the algorithm is reconstructed. The algorithm is simulated on IEEE-14 bus system. The results show that the proposed algorithm can use multiple sensor nodes state to estimate multi-bus harmonic voltage state cooperatively. It has better anti-disturbance performance, more accurate estimation in comparison to the conventional Kalman filtering, and it displays high real-time and good dynamic characteristics.

关 键 词：时变谐波检测 相关性分布式卡尔曼 相关系数 IEEE-14节点系统 抗扰动 

分 类 号：TM711[电气工程—电力系统及自动化] TN711[电子电信—电路与系统]