Dynamic State Estimation Of Distribution Network With High Proportion Of Photovoltaics

With the high proportion of distributed photovoltaics connected to the distribution network,the dynamic process of the distribution network is becoming increasingly complex.The static state estimation based on conventional data acquisition and monitoring systems is unable to meet the requirements of advanced applications in the power distribution management system.At present,there are a few studies on the dynamic state estimation of the distribution network,and even fewer studies on the dynamic state estimation considering new energy resources such as photovoltaics connecting to distribution network.Therefore,this paper studies the dynamic state estimation method of the distribution network with a high proportion of photovoltaics providing data support for the safe and economic operation of the distribution network.This article focuses on the following aspects:Aiming at the basic theory of distribution network state estimation,a three-phase distribution network modeling method is studied,a three-phase power flow model for the distribution network is established,and an adaptive Levenberg-Marquard method is used for robust power flow calculation.The measurement transformation method of distribution network is studied,and the quadratic constraint linear measurement model of distribution network is established.The state transition model of distribution network is studied,and the state estimation method based on unscented Kalman filtering is established.Aiming at the change of distribution network characteristics after a high proportion of photovoltaics are connected,the conventional photovoltaic model,virtual synchronous machine PV model,and dynamic load model are considered,and the static state estimation and dynamic state estimation models are improved.A static robust state estimation method based on generalized maximum likelihood estimation and a strong tracking dynamic state estimation method based on adaptive noise estimation are proposed.Aiming at the problem of lacking information interaction between static estimation and dynamic estimation in a novel two-stage estimation architecture,a two-stage dynamic state estimation data fusion method is proposed.An actual test system model was established,and the tracking performance and robust performance of the two-stage dynamic state estimation were verified,and the validity of the proposed dynamic state estimation algorithm for the distribution network was verified.