No abstract

No abstract

Machine learning techniques have been widely used in transient stability prediction of power systems. When using the post-fault dynamic responses, it is difficult to draw a definite conclusion about how long the duration of response data used should be in order to balance the accuracy and speed. Besides, previous studies have the problem of lacking consideration for the confidence level. To solve these problems, a hierarchical method for transient stability prediction based on the confidence of ensemble classifier using multiple support vector machines (SVMs) is proposed. Firstly, multiple datasets are generated by bootstrap sampling, then features are randomly picked up to compress the datasets. Secondly, the confidence indices are defined and multiple SVMs are built based on these generated datasets. By synthesizing the probabilistic outputs of multiple SVMs, the prediction results and confidence of the ensemble classifier will be obtained. Finally, different ensemble classifiers with different response times are built to construct different layers of the proposed hierarchical scheme. The simulation results show that the proposed hierarchical method can balance the accuracy and rapidity of the transient stability prediction. Moreover, the hierarchical method can reduce the misjudgments of unstable instances and cooperate with the time domain simulation to insure the security and stability of power systems.Energies 2016, 9, 778 2 of 20 from among massive sets of data, have been used to predict the transient stability [2][3][4][5][6][7][8][9][10][11][12][13][14][15][16][17][18][19]. The transient stability prediction can be treated as a two-class classification (stable and unstable) problem and solved by machine learning methods. A set of appropriate features/attributes is selected to generate the offline training sets, then an appropriate classification method is utilized to predict the transient stability status.There are two kinds of machine learning-based methods with different types of inputs [2]. One uses pre-fault steady-state variables as the original data, the machine learning method will be used to build the mapping between the steady-stable variables and the transient stability status with respect to an anticipated but not yet occurred contingency [2][3][4][5]. Once the current status is identified as insecure, preventive control can be carried out to modify the system to a secure state. However, when a serious fault happens, or failure of primary relay protection, the power system may still be unstable even if the preventive control has been conducted. Therefore we should emphasize importance of the study on transient stability prediction using post-fault responses. Because the post-fault responses carry information about the influence of the faults on the power system, the prediction is independent of the faults. This arises the second way of transient stability prediction based on machine learning techniques.With the development of wide-area measurement systems, the dynamic response of power systems...

No abstract

No abstract