Research On Bad Data Detection And Identification As Well As Methods To Improve The Qualified Rate Of State Estimation

With the continuous improvement of the automation level in power system, there is higher and higher requirement of real-time data accuracy. However, due to the effects of measurement device fault, unfavorable communication and so on, the measurements contain much bad data. The existence of bad data does harm to the secure operation of power system. As one of the important function of state estimation, bad data detection and identification can improve the reliability of real-time data. To solve this problem, this paper studies on the bad data detection and identification algorithm as well methods to improve the qualified rate of state estimation.Firstly, the paper expounds the basic conception and research status of bad data detection and identification in power system. Then, the common methods of bad data detection and identification have been put forward. The advantages and disadvantages of these methods are compared.Secondly, the theory and basic methods of cluster analysis have been provided. Based on COPS(clusters optimization on preprocessing stage), a new method to identify the bad data in power system is put forward. The proposed method regards standard residuals and the difference value between adjacent sampling times as the clustering feature. The method agglomerative generates the hierarchical partitions of dataset first. Subsequently, a curve of the clustering quality with respective to the varying partitions is constructed. The partitions corresponding to the minimum of the curve is used to estimate the number of clusters. Experimental results show that the new method can not only identify the bad data rapidly and accurately, but also avoid residual pollution and residual submersion effectively. The simulation results indicate that the proposed method is suitable for the practical power system..Finally, the methods to improve the qualified rate of state estimation were studied. Based on the D5000 smart grid dispatching supporting system of Hebei provincial power dispatching center, the paper analyze the influential factors of the qualified rate of state estimation. According to the maintenance work of state estimation software, some advices and debugging methods are presented. The implementation of the presented proposals and methods effectively guide the maintainers. And the qualified rate of state estimation was significantly improved.