The Study Of Lightning Induced Fault Identification And Optimization Of Traveling Wave Fault Location

Fault location technology contributes to finding permanent fault and removing hidden danger,so some instant faults can be avoided.The state monitoring systenm,fault analysis recorder and lightning location system can record data of the same fault or disturbance,especially the lightning location system(LLS)and traveling wave fault location system can record the lightning induced fault from different recording system.Massive recording data recorded by LLS and traveling wave traveling wave fault location system which is widely used.It is easily to use data from different source to identify fault cause and precise fault location.The time interval of traditional relay is "second" level,which is much difficult to relate to the "millisecond" time interval of LLS.The time interval of traveling wave fault location system is "microsecond" level,which can relate to the record of LLS.The traveling wave fault location system calculates the distance of the line,but the LLS record the lightning by lattitude and longuititute.Thus the Geographic Information System is used to extract the altitude of the transmission line to transform fault distance to fault tower which achieves the relating of fault tower GPS and lightning GPS.The shielding failure,lightning striking tower and lightning striking grounding line are analyzed to extract the lightning wave characteristic.The impulse corona and tower structure is considered to study the impact of trasient current on transmission line and insulator.The traveling wave data for a fault is unique,but lightning location system always records more than one cloud-ground flash shortly before and after the fault accuring,thus the lightning records need to be The distance and time difference between a fault location result and lightning location system record are used as a kind of characteristic vector as the input of k-Means clustering to sifting the possible lightning records which induced a fault.The number of lightning detection station,polarity of lightning and polarity of traveling wave are involved to form the correlation function to find the most possible lightning record induced a fault.The traveling wave involves quantization noise,electromagnetic noise and other noise,so the wave contains so many sigularity that the traveling wave reflected from fault point can not be easily automatic detected by fault location software,especially that there is only one transmission line on the remote busbar leads to the same polarity of initial traveling wave and reflecting wave from remote busbar.When the lightning induced fault occurs,the suspected fault distances are calculated by the suspected reflection from fault point,and the correlation function is used to find the real reflection from fault point.The calculated fault point,lightning flash point and direction of line corridor are used to form space trend approach to optimizing the traveling wave fault location result.The traveling wave contains much follw-up oscillation after the traveling wave arriving,thus the appropriate filter is hardly chosen.Speeded-up Robust features(SURF)is used to test the similarity of two images and to distribute the initial traveling wave and reflection of fault point to two non-overlapping time windows.So the reflection of fault point is guaranteed to be the first obvious sigularity in its time window.The wavelet is used to extract the time difference between the initial traveling wave and reflection of fault point.The possible error source of lightning location system is analysed.The strict matching of lightning records and traveling wave data is used to extract the lightning induced fault examples.The distance and altitude curve between the initial flash location and detection station by Geographic Information System(GIS).The Fractal dimension is used to calculate the complexity of altitude curve,which is as a kind of weight function for optimization for lightning flash location results.