基于线路阻抗辨识的微电网无功均分改进下垂控制策略

Improved Droop Control Strategy Based on Line Impedance Identification for Reactive Power Sharing in Microgrid

在孤岛运行的微电网中,分布式电源(DG)以并联方式运行,在其等效线路阻抗不匹配的情况下,传统的下垂控制难以按无功功率-电压的下垂方式对无功功率进行合理分配。为此,提出了一种基于线路辨识的改进下垂控制来消除无功功率分配的偏差。首先,详细分析了线路参数对传统下垂控制效果的影响;其次,针对单个分布式电源,构建了基于本地信号的线路观测环节,以对分布式电源的连接等效线路阻抗进行精确辨识,并在传统下垂控制中基于线路辨识结果加入了电压补偿项,使它在等效线路阻抗不匹配时,仍能保持无功功率的合理分配;最后,将提出的改进下垂控制策略应用于一个典型的微电网系统,并利用MATLAB/Simulink仿真验证了所提方法在多工况下的正确性和有效性。仿真结果表明:所提出的主动线路观测器可基于本地信号较精确地辨识出等效线路阻抗的实际值,基于该线路辨识结果,可有效对DG等效连接线路的不匹配电压降进行补偿,从而实现无功功率的合理分配;该改进方案无需依靠实时通信系统交换各DG源间的信息,也不需要利用中央控制器发送同步信号来保持改进控制启动的同步性。研究结果证明了所提改进下垂控制策略的有效性。

In islanded mode microgrid, the distributed generation（DG） units are operated in the parallel way. In the unequal line impedance condition, the traditional droop control is unable to proportionally share the reactive load. To solve this issue, we proposed an improved droop control strategy based on the line impedance estimator to eliminate the sharing error. Firstly, the relation between the droop control and line impedance was analyzed; secondly, the line impedance estimator of single unit based on the local measurement was built to identify the line impedance with high accuracy. In addition, a voltage compensation term related to the identification result was added to ensure propor- tional sharing in non-trival impedance condition. Finally, the proposed scheme was applied to a microgrid in MATLAB/Simulink software environment.The simulation results demonstrate the proposed active observer for line impedance is able to identify the line impedance accurately. Based on the estimation result of line impedance, it is able to compensate unmatched voltage drop accurately and make reactive power to share by a reasonable way. The proposed method does not need transmission system to transfer message of DGs, and does not meet the central controller to send synchronizing signal to make sure the synchronism of improved control. The research results prove that the proposed improved droop control strategy is effective.