Voltage Sag State Estimation Based On The Optimal Allocation Of Monitoring Nodes

With the rapid development of industrial technology, a large number of sensitive electrical equipments have been applied to power system, which has brought forward higher requirements on power quality. Among all the power quality problems, voltage sag caused by short-circuit fault has been acknowledged as one of the uppermost problems. To solve the problem, it is of vital importance to have an accurate grasp of voltage sags in the whole power system. Therefore, the voltage sag monitoring and state estimation have important practical significance.From the monitoring point of view, if the expensive power quality monitoring equipment is installed at each bus node in the huge power network, it not only needs a huge amount of construction and maintenance costs, but also brings tremendous pressure to communication and data handling. So practically it is feasible to realize the monitor for the voltage sags of the entire power network by using a limited number of monitoring nodes. Therefore, the key to have a systematic grasp of voltage sags in the whole power system is to make effective estimation of voltage sags information of those unmonitored nodes by using the limited monitoring data.In order to achieve the optimal allocation of the monitoring nodes, this thesis firstly analyses the vulnerable areas of the system bus nodes under various types of short-circuit faults by using voltage sag analytical expression. Then, a voltage sag monitoring node optimal allocation model is established, whose object function is to minimize the number of monitoring nodes and whose key constraint is to guarantee the observability of the voltage sag of each node in the whole system.Secondly, based on the improved optimal allocation of the voltage sag monitoring system, this thesis proposes a new method for the voltage sag frequency state estimation by adopting hybrid particle swarm optimization (HPSO) algorithm. In the first place, on the basis of the state estimation models for power systems, the mathematical model of voltage sag state estimation is constructed by using the vulnerability area matrix. Then, by combining Constriction Factor-PSO with genetic algorithm (GA) and simulated annealing (SA) algorithm, a HPSO algorithm is developed and applied to solve voltage sag state estimation model. Finally, the efficiency and the accuracy of the proposed method are validated through the performance illustrated in the IEEE 39-bus test system.In order to further improve the practicability of voltage sag state estimation, this thesis provides a voltage sag state estimation method based on the fault position method. In this method, the voltage sag measurement equation is analyzed by applying the fault position method instead of the analytic expression method, and then a mathematical model of voltage sag state estimation is built according to the analysis, and it is finally solved by adopting the singular value decomposition technique. By simulation and comparison, the efficiency and the accuracy of the proposed method are verified, on which basis the various factors affecting the result of the state estimation are analysed through simulation experiments.Finally, on the basis of the fault position method, this thesis proposes a technique for multi-fault voltage sag state estimation method. Combining the nodal voltage equation and the boundary condition under multiple faults, the computational method of multi-fault voltage sag is proposed. According to this method, the basic idea of fault position method is introduced into the analysis of multi-fault voltage sag state estimation, and thereby the model of multi-fault voltage sag state estimation is constructed. Eventually, the singular value decomposition method is used to obtain the voltage sag frequency of those unmonitored nodes. Both the efficiency and the accuracy of the multi-fault voltage sag state estimation method are testified through simulation analysis.