Using Branch Information For Network Topology Identification

Power system state estimation is the base of power system related aspects, such as dispatch, control, and security evaluation, and also the core component of electric energy management system. After the introduction of state estimation into the power system area, the skills of power system estimation was significantly improved, and many research achievements were obtained in the following respects: estimation criterion, detection and identification of bad data, stability of state estimation algorithms, estimation and identification of system parameters, identification of network topology error, and so forth. However, lots of unsolved problems are still existed in certain aspects of power system state estimation. Moreover, with the improvement of techniques and the change of requirements on power system state estimation, amelioration becomes necessary in some fields of power system state estimation.This thesis focuses on state estimation algorithm and network topology error identification. First of all, summarized the state estimation and reviewed the classic algorithm of state estimation. After summarizing the existing research results in the view of network topology error detection and identification problem, put forward specific solution. The network structural topology error is defined as the unmatching between the switch remote signaling and the switch actual position in the long-range systems and need to test, estimate, identify and correct.Using branch information for network topology identification is developed in this thesis. First of all, suppose the branches running state can be said as branch parameters which are continuous and between [0, 1], when the branch parameter equals 0, means branch off; when it equals 1, branch on; Branch operation state is not sure when it betweens(0,1). Specifically speaking, via introducing a real parameter, the branch admittance can be expressed. So, the state variable can be extended into real part and imaginary part of the voltage, and the parameter; and the Jacobi matrix can be extended into a derivative matrix via this parameter of Jacobi matrix and measurement equation of the traditional state estimation. In the process of state estimation, due to lacking restriction to voltage in the target function, it is probably that the correction is too big, and correspondingly the problem about astringency is produced, which could be solved through transformation into linear programming problems. Therefore construct the branch parameter function and rule it and measuring residual in least squares criterion respectively. Then take the adding form of the two as the new objective function to estimate branch parameters.For example, via the method described in this article, after select the suitable parameters to detect the topology error of test system, the lines possessed topology errors can be identified, and the effects of error measurements on topology structure can be eliminated. In conclusion, the test of the least absolute value state estimation algorithm on this actual testing system demonstrates that the numerical value stability is increased, the iterative times is reduced, and the calculating speed is high, which excels the conventional method.