Study On Harmonic Source Localization In Distribution Network Considering Three-phase Model

With the rising number of renewable distributed generations such as wind power generations and photovoltaic power generations are connected to the power grid especially the distribution network, the distributed generation may become a harmonic source because of a failure, affecting the power quality and threatening the safety and economic operation of the power grid. Given the three-phase asymmetry of the harmonic waves, the localization of the harmonic source in distribution network considering three-phase model is important to distinguish the harmonic responsibility and to handle harmonic pollution more specifically.In the distribution network, not all the nodes are equipped with the harmonic measurement instruments. It is important to satisfy the measure requirement of the harmonic source localization with the least measure instruments. This paper will study the measurement configuration of the harmonic source localization. A harmonic measurement system based on WAMS and the basic measurement equation will be introduced. Then a PMU optimal placement method will be studied and tested in the IEEE 33-bus distribution network.Harmonic state estimation is an effective way to locate the harmonic source. Though single-phase model are used in most papers, the 3-phase model should be considered in the localization of the harmonic source as the parameters are not symmetrical. This paper will study a method to locate the harmonic source in 3-phase distribution network using total least square method, with the parameter errors and the measurement errors considered. Several criterias will be presented to identify the main harmonic source and evaluate the affection to the power network. The proposed method will be tested in the IEEE 33-bus distribution network.The distribution network usually is not observable. It is hard to identify the harmonic source by the traditional estimation methods. This paper will use the sparsity of the harmonic sources to reduce the number of the state variables. A method is proposed to turn the under-determined state estimation equation into a determined even over-determined equation to improve the estimation accuracy. The proposed method will be tested in the IEEE 33-bus distribution network compared with the least square method.