The Application Of Wavelet Energy Spectrum In EHV Transmission Line Transients Recognition

Extra High Voltage (EHV) transmission line is a very important part of long distance transmission. For high quality of power transmission, the stability and reliability of transmission line are so crucial. For example, a disturbance signal, which could not be recognized and cleared soon, may lead to large fault. It will affect the normal operation and great loss. Correct recognition of different type and location of transients is very important in improvement of the stability and reliability of power net and can avoid greater fault. With the development of power net and rise of voltage, recognition of transmission line transients becomes more and more important.On the base of analysis of EHV transmission line, the paper do some work on the resent research of its transient signal recognition and wavelet energy spectrum application in this filed.Based on multi-resolution, the wavelet energy spectrum is proposed. The matrix similarity and relative entropy, which are deduced by wavelet energy, are analyzed. This analysis provides a good foundation of further simulation and algorithm. In order to weaken the effect of different current magnitude, a normalized wavelet energy matrix is given. This normalized matrix could make all different data range mapped in a same range.A 500kV EHV transmission line is modeled in PSCAD/EMTDC. Four common kinds of transients, single phase to ground, breaker operation, faulted lightning and lightning disturbance, are modeled. Their wavelet energy spectrum is analyzed based on the method of math space distance, contrastively. From the contrast, it is easy to find that there are differences between different transient signals in both high and low frequency band. Thus, wavelet energy spectrum is a potential tool for transient signal recognition.With the different features of signals of different types and locations, the normalized wavelet energy matrix which can sharpen those differences is used in this paper to classify signals. Matrix similarities between signals and standard matrixes are calculated. When the similarity gets its max value, the signal is considered to be that standard type. In the simulation, transient signals of different types and different breaker locations are recognized. The simulation results shows the method based on wavelet energy matrix similarity is effective in classifying different signals and meets the needs of practical use.Finally, the application of wavelet relative entropy in EHV is studied. The relative entropy of three kinds--the entropy between signal and its standard matrix, the entropy between signals of different types, the entropy between high and low frequency band of certain signal, are analyzed, contractively. Based on this analysis, a recognition method based on wavelet relative entropy is proposed. Simulation is done in the same model above. Two kinds of entropy are adopted in this simulation. Shown by the simulation results, the wavelet relative entropy could distinguish different kinds of transient signals. This method is easy to done and needs little calculation. It is a potential tool for transmission line signal recognition.