Identification of Power Quality Transient Disturbance Based on S Transform and Wavelet Transform

Abstract

With the rapid advancement of the country's industrialization, intelligence, and information technology, the nonlinearity, asymmetry, and impact of power system load are becoming increasingly prominent. On the other hand, users' demands for power quality are increasing. The national grid inevitably faces severe challenges. The power quality problem in circuit systems is particularly important. The type identification of power quality transient disturbances is of great significance for analyzing power quality problems and solving power quality problems. In this paper, a method based on S transform and wavelet transform for transient power quality disturbance identification is proposed. Firstly, the wavelet signal is used to denoise the power signal. Then, after obtaining the feature vector of the power signal by the S transform and the wavelet transform, The Relieff algorithm is used to obtain the feature weights of the feature vectors generated by S transform and wavelet transform and to select features. Finally, training and classification are performed using the Support Vector Machine (SVM) classifier. The experimental results show that the power quality transient disturbance classification method proposed in this paper can accurately identify the type of disturbance and have strong robustness, and fully meet the requirements of the relevant industry for the accuracy of power quality transient disturbance recognition.

Power quality transient disturbances detection system based on db5 wavelet

“…The short-time Fourier transform can divide the disturbance waveform into several smooth waveforms, overcoming the spectral leakage defect of the Fourier transform, but the shape and width of the window are fixed, which is not suitable for the analysis of transient disturbance signals [7] . The S-transform uses a Gaussian window function whose window width is inversely proportional to the frequency, eliminating the choice of the window function and improving the defect of fixed window width, with good time-frequency characteristics, but for signals with sudden changes in time, the singularities of the signal cannot be detected [8] . The Hilbert yellow transform is more adaptive than the S-transform and does not require prior selection of the basis function, which is suitable for the analysis of non-stationary nonperiodic signals, and can quickly detect voltage sags, short-term interruptions, and other types of disturbances, however, the method also suffers from constraints such as endpoint effects and modal mixing phenomena, making it difficult to achieve real-time detection in hardware [9] .…”