No abstract

Machine learning methods are promising tools for transient stability assessment (TSA) of power system. Support vector machine (SVM) is used to assess the transient stability of power system after faults occur on transmission lines. Single machine attributes is studied as inputs of the SVM classifier. Experimental results in IEEE 50-generator test system shows that, attributes of single machine with small inertia coefficient are effective in TSA, and the SVM classifier with RBF kernel using these single machine attributes can achieve satisfying classification accuracy.

Cardiovascular disease (CVD) is considered one of the leading causes of death worldwide. In recent years, this research area has attracted researchers’ attention to investigate heart sounds to diagnose the disease. To effectively distinguish heart valve defects from normal heart sounds, adaptive empirical mode decomposition (EMD) and feature fusion techniques were used to analyze the classification of heart sounds. Based on the correlation coefficient and Root Mean Square Error (RMSE) method, adaptive EMD was proposed under the condition of screening the intrinsic mode function (IMF) components. Adaptive thresholds based on Hausdorff Distance were used to choose the IMF components used for reconstruction. The multidimensional features extracted from the reconstructed signal were ranked and selected. The features of waveform transformation, energy and heart sound signal can indicate the state of heart activity corresponding to various heart sounds. Here, a set of ordinary features were extracted from the time, frequency and nonlinear domains. To extract more compelling features and achieve better classification results, another four cardiac reserve time features were fused. The fusion features were sorted using six different feature selection algorithms. Three classifiers, random forest, decision tree, and K-nearest neighbor, were trained on open source and our databases. Compared to the previous work, our extensive experimental evaluations show that the proposed method can achieve the best results and have the highest accuracy of 99.3% (1.9% improvement in classification accuracy). The excellent results verified the robustness and effectiveness of the fusion features and proposed method.

With the continuous improvement of substation automation, the number of communication equipment has risen sharply, the network topology is more complicated, and the problem of unable to locate the fault of process-level communication networks quickly and accurately has become more and more prominent. Under this background, to improve the maintenance efficiency of process-level communication networks in the smart substation, this article proposes a method for fault location of process-level communication networks based on deep neural networks. Based on the redundant monitoring of the fault state, the fault feature information of different monitoring nodes in the message transmission process is analyzed, and the characterization method of the fault feature information is proposed accordingly. Based on the emergence principle, it realizes the automatic generation of fault samples according to the physical connection, logical connection, and message subscription relationship of process-level communication networks. Combined with the training rules in deep learning theory, a fault location model of process-level communication networks based on deep neural networks is established, and the real-time location steps are given. Taking the process-level communication network of a typical 110kV smart substation as an example, simulations verify the effectiveness and accuracy of the proposed fault location method in single fault environments and multiple fault environments. Besides, accurate results can be obtained even when the fault feature information is wrong or missing, and the anti-interference ability is excellent.

No abstract