Research On Sampling And Identification Of System Transient Disturbance Signals Using Cable HFC

With the advance in the smart grid and new energy generation,a large number of power electronic devices,intelligent control devices,and high-power nonlinear loads are continuously put into use,making the transient disturbances in the power system normalized and complicated.These transient disturbances have a large impact on a vast number of sensitive devices in the new power system and even cause chain accidents,affecting the safe and stable operation of the power system.However,the signals are not easy to capture due to the wide dispersion of transient disturbances.Meanwhile,the transient characteristics are random and variable because of a wide range of sources,It seems he related research on transient disturbances is still immature.This paper focuses on the sampling method,feature extraction,and classification identification of transient disturbance signals to provide a basis for further quantitative analysis and control.To address the sampling problem of transient disturbance signals,this paper proposes a novel sampling method to solve the defect that transient disturbances distributed in various parts of the power system are not easily captured by substation monitoring equipment.The method makes full use of the high-frequency current transformer(HFCT)installed at the cable attachment as a high-speed signal acquisition device,which has the advantages of broadband,high sampling,and high sensitivity,and can more comprehensively detect all kinds of transient disturbance signals in the system.At the same time,the HFCT is easy to install and close to the power equipment,providing a convenient source of signal acquisition for transient disturbance signal analysis.The method solves the shortcomings of low sampling rate and limited detection range of power quality monitors and fault recorders in substations,and fully extends the utilization value of HFCT.This paper uses real HFCT samples for transient disturbance signal sampling experiments,demonstrates the feasibility and effectiveness of HFCT to acquire transient disturbance signals,and verifies the superiority of the method,which has strong practicality and promotion value.For the feature extraction method of transient disturbance,this paper proposes a feature extraction method based on multiple time-frequency feature matrices by deeply analyzing the specific features of high-frequency prominence and signal weakness of transient disturbance signal after HFCT filtering.The method firstly decomposes the high-frequency weak waveform signal into multiple components with different center frequencies by using parameter-optimized variational modal decomposition,from which the waveform feature matrix is extracted;then further Wigner-Ville distribution time-frequency analysis is performed on the different frequency components to obtain the time-frequency map and extract the time-frequency map feature matrix of each component;finally,the waveform feature matrix is fused with the time-frequency map feature matrix Finally,the multiple time-frequency feature matrix is constructed by fusing the waveform feature matrix and the time-frequency feature matrix.The analysis results based on the experimental data show that the proposed feature extraction method effectively solves the problem of difficult parameter selection in the variational modal decomposition,and improves the decomposition accuracy and self-adaptive capability.Meanwhile,the influence of cross-interference terms in the Wigner-Ville distribution is eliminated,and the time-frequency resolution of the Wigner-Ville distribution is improved.The multiple time-frequency feature matrix proposed in this paper can fully explore the local details of high-frequency weak transient disturbance signals so that they can be distinguished from each other with a high degree of differentiation,which is conducive to their further classification and identification.Based on the proposed multiple time-frequency feature matrix,this paper uses the long short-term memory network(LSTM)to achieve effective recognition of transient disturbances.The method uses LSTM as the classifier,and the extracted multiple time-frequency feature matrix is input into the LSTM network for training and learning,to realize the classification and recognition of transient disturbance signals.The experimental data show that the method has a high recognition rate under the high noise environment with a low SNR level,and the anti-noise performance is strong,which is suitable for the practical application environment of HFCT.At the same time,the recognition rate of the method is not mainly affected by the disturbance distance,as long as the disturbance can be sensed to start the recognition process and correctly recognized.This paper provides a new monitoring scheme for system transient disturbance identification,which has been verified by experimental tests and has certain practical application potential.