Study On Identification Of 500kV Substation Overvoltage Based On EMTP Simulation Data

The impact of overvoltage on the electrical equipment in substations is the main factor responsible for the insulation damage of the equipment,and the magnitude of overvoltage is also an important reference factor for the insulation coordination design of electrical equipment in substations.There are various kinds of overvoltages occurring in substations,which amplitude,duration and frequency are significantly different.The fast and accurate identification of overvoltage in substations is not only conducive to dealing with faults in time,but also helps to improve insulation coordination of electrical equipment in substations.Based on the simulation overvoltage data of a 500 kV substation,the problem of feature extraction and classification of overvoltage is studied.Firstly,it based on EMTP software to builds an overvoltage simulation model that according to the specific parameters of an actual 500 kV substation.Seven typical substation overvoltages namely single-phase short circuit overvoltage,switching off unload transformer overvoltage,switching off unload line overvoltage,switching on unload line overvoltage,lightning induced overvoltage,lightning back-strike overvoltage and lightning shielding failure overvoltage are calculated.The simulation waveforms of these seven kinds of overvoltages are consistent with theoretical analysis,which verifies the correctness of the simulation model and the calculation results.Secondly,the characteristics of the above seven kinds of over voltage signals are extracted by the time-frequency domain analysis method.In the frequency domain,the harmonic distortion rate of each overvoltage signal is calculated,and to some extent its value can reflect the characteristics of the overvoltage,so that it is proposed as the characteristics of overvoltage identification.In the time-frequency domain,multi-scale analysis method of wavelet transform is used to analyze the overvoltage signal,its results show that the relative energy distribution of overvoltage is significant difference in each frequency bands.Based on this,the relative distribution characteristics of overvoltage energy are considered as the characteristics of overvoltage identification.Meanwhile,the S transform is used to analyze the zero sequence component of the overvoltage signal,and singular value decomposition(SVD)of the S matrix is carried out to construct the characteristics of the overvoltage identification.Finally,the support vector machine theory and its classification principle are briefly introduced,and the BT-SVM classification method is used to construct the substation overvoltage classifier.Cross validation and grid search method are used to optimize the parameters of SVM.The simulation data verification show that the overvoltage classifier can identify the overvoltage effectively.