Search

Fault diagnosis is important to stable operation of power systems, and the machine-learning-based fault-diagnosis models were widely studied because of their strong generalization ability. However, the model structures are generally designed according to the topology of power systems. Once there are changes in topology, the system fault characteristics might change, and the models structure and parameters are often required to be adjusted for applying to the new systems. To avoid frequent adjustment work of fault-diagnosis models when system topology changes, we propose a novel fault-diagnosis model for power systems. First, a new data preprocessing using gradient calculation and similarity assessment is presented, and the gradient similarities among the multichannel electrical signals are converted to the visualized similarity images, which are fed to the neural network for further processing. Second, the spatial pyramid pooling (SPP) and hashing classifier (HC) are used in the convolution neural network. With the aid of the SPP and HC techniques, the structure of the fault-diagnosis model can maintain unchanged even though there are topological changes in the power systems. To validate the effectiveness of the proposed model, several state-of-the-art fault-diagnosis models are used for comparison. The results show that the proposed model with unchanged structure is well performed in accuracy and noise immunity, and friendly to the parameters setting.

In this paper, the issue of open-switch faults of voltage source inverter (VSI) is studied, an adaptive fast voltage-based real-time open-switch fault location method is proposed. Firstly, for each switching period, the voltage deviation is calculated by the expected voltage and estimated actual voltage, they are obtained via the mixed logical dynamic model (MLDM) and measured currents, respectively. Then, the voltage deviation of each phase is employed as a fault variable. By this way, avoiding the use of additional hardware, and embedding easily in system. Furthermore, considering the impact of sampling error, parameter error, dead time, delay time, and transition time, the adaptive threshold is designed. Finally, simulations and experiments are executed to verify the effectiveness of the proposed method.

Large-scale wind farms (WFs) generally consist of hundreds of wind turbines (WTs), and the WFs simulation model construction would be complex and even impossible if we develop each individual WT in detail. Therefore, the WFs equivalent simulation model with required accuracy is essential to be developed to explore the WFs operation characteristics. This article proposes an equivalent method for large-scale WFs using incremental clustering and key parameters optimization. Firstly, to acquire more comprehensive and distinguishable representations of WTs operation characteristics, the time series of WT active power, reactive power, voltage and current are selected as the multi-view clustering indicator (CI). Then, considering the computer memory pressures encountered by traditional clustering algorithms in dealing with large-scale WFs, a novel clustering algorithm namely multi-view incremental transfer fuzzy C-means (MVIT-FCM) is proposed, and this algorithm can process the WTs clustering problems without requiring to consider the scale of the WFs. Finally, to further increase the equivalent accuracy of the WFs equivalent simulation model, key parameters in the equivalent model are found using Sobol' criterion and then optimized using the designed Q-learning based non-dominated sorting genetic algorithm II (NSGA-II). To verify the effectiveness of the proposed method, the modified WFs system in China is utilized for case study, and the performance of the proposed model is compared with several state-of-the-art models. Simulation results show that the equivalent accuracy of the proposed model is higher when comparing with other models. Also, the proposed model has the advantage of processing the WFs equivalent problems with any scales.