Distribution Network State Estimation Based On Historical Data Modeling

With the rapid development of smart distribution network technology,the requirements for real-time monitoring of the automation level and system status of the distribution network are getting higher and higher.Distribution network status estimation is the basic core part of the distribution management system,providing basic data support for analyzing the distribution network operation status.However,at present,there are fewer configuration devices in the distribution network,and the measurement redundancy is low.It is difficult to ensure the observability of the system,and the estimation accuracy cannot meet the requirements.On the other hand,the load of the distribution network and the three-phase imbalance of the line are not balanced.The problem also brings more difficulties and challenges to the estimation of traditional state.Therefore,this paper studies a state estimation method based on historical data modeling.Firstly,the relationship and difference between power flow calculation and state estimation of distribution network are discussed,and the basic methods of power flow calculation for distribution network are introduced.Then considering the problem of three-phase unbalance of distribution network,the three-phase model of several main components in the distribution network and the mathematical model of the distribution network measurement system are described.Finally,the state estimation of distribution network under weighted least squares and differential evolution algorithm is introduced.Secondly,the operational state data of the distribution network historical time is analyzed and a multivariate regression model is established.The results obtained by the model not only supplement the information of the unmeasured device nodes,but also expand the distribution network state estimator measurement and improve th e system's observability;Compared with the traditional pseudo-measurement,it provides more accurate pseudo-measurement data for subsequent state estimation,which is beneficial to improve the estimation accuracy.Finally,taking the amplitude and phase angle of the node voltage as the state quantity of the state estimation,the pseudo-measurement data provided by the historical data model is used to simulate the three-phase balanced IEEE33 node system and the three-phase unbalanced IEEE13 node system.By analyzing the influence of historical data model on the observability and estimation accuracy of distribution network state estimation,the optimization effects of traditional weighted least squares algorithm and differential evolution algorithm on state estimation of distribution network are compared.The results show that the addition of historical data modeling values as a new pseudo-measurement improves the state estimation accuracy,and the estimation accuracy under the differential evolution algorithm is better,which verifies the effectiveness of the proposed method.