Research On Identification And Estimation Method Of Network Parameter Errors Considering Network Division

State estimation of power system is a very important part of energy management system. Its result directly affects the intelligent analysis and decision-making of power dispatching. In practical opration powe system, it is difficult to obtain correct real-time measurement data and network parameters, which musr cause the result of state estimation incorrect. Therefore, it is important to study erroneous parameters along with bad telemetry measurements which is referred to as bad data for short. Existing methods can not identify and estimate multiply bad data effectively. To solve this problem, an identification and estimation method of network parameter errors considering network division is presented. The main contents are as follows.Firstly, a network topology dvision method based on minimum degree search is proposed. The network is divided into radial network, simple meshed network and complex meshed network to form many independent subareas by network topology searching. And the information of branchs, nods and the public boundary nod in every subarea are established automaticly, thus an important foundation for the divisional realization of parameter identification and estimation is laid.Secondly, a divisional network parameter identification approach is presented. Combining network division with Lagrange identification approach, Lagrange approach is carried out to identify bad data in each subarea. The interaction of bad data between different subareas can be avoided by the proposed method. When multiply bad data exist in different subareas, all bad data can be identified simultaneously. Thus, the identification of multiply bad data of many subareas in the same network is realized efficiently.Lastly, a function weighted least squares augmented method is proposed. If a parameter is identified as erroneous in each subarea, weighted least squares augmented method is used to estimate its value. The weight of least squares state estimation is corrected dynamically via the use of weight function, so the effect of bad data during parameter estimation is inhibited. If there are more than one parameter errors in a subarea, the parameter-residual sensitivity parameter estimation method is carried out again to estimate all of them. The interaction of multiply parameter errors in the same subarea is avoided, thus, the accuracy of estimating erroneous parameter is greatly improved. The proposed method combine network division and Lagrange approach and weight function estimation method, the effectiveness of identifying multiply bad data and the accuracy of estimating erroneous parameter are improved, thus it has great practical value. The validity of the proposed method is verified by the simulations on the IEEE 30-bus system and IEEE 118-bus system.