Algorithm Research Of Power Cable Fault Location Based On Oscillatory Voltage Partial Discharge Detection

With the upgrading of power grids, XLPE power cables gradually become an important part of urban power network due to their own advantages. Cable’s reliability is an important factor for the safe operation of power grid. However, as the using time prolonged, cable insulating material is affected by the external environment, occurring cable insulation degradation phenomenon. Faulty cable will occur frequent partial discharges(PD) with applied high voltage. Oscillation Wave detection System(OWTS) is the most widely used method in recent years, owning to its advantage of high precision.Firstly, the paper studies the mechanism of PD and the technique of OWTS deeply, dissecting the problems and insufficiencies of present methods. That is, the existing methods mostly adopt correlation analysis to match the PD pulses, but seldom combine PD pattern recognition with correlation analysis, so that the accuracy of PD detection and location is not ideal.Secondly, the paper collects OWTS detection data on scene, and extracts waveform characteristics and equivalent time-frequency characteristics T-F form PD pulses as well as noise pulses, and adopts support vector machine(SVM) to build an effective PD pulse recognizer. This recognizer could extract PD pulse and get rid of noise pulse form test data, instead of pulse recognition dependent only on experiential threshold value.For pulse matching and fault location, the paper combines the algorithm of feature recognition with dynamic time wrapping(DTW)- correlation coefficient, making the PD incident pulse match accurately with its reflected pulse. Farther more, a concentrated PD detection algorithm is put forward, in order to achieve the fault location identification automatically. In addition, through verification of laboratory as well as actual test data from knowledge base, the paper proves the algorithm is effective and accurate.At last, the paper introduces the intelligent mobile terminal detection and analysis system for power distribution equipment. The OWTS mobile intelligent system is developed by MATLAB, Microsoft.Net and Java EE, which are severally responsible for analysis algorithm, interface business logic and database control.On the whole, the fault location algorithm and mobile intelligent terminal detection system could not only improve fault location accuracy, prevent unnecessary maintenances, but also be beneficial for big data processing, as well as the further popularization and application of OWTS, in response to the intelligent and integration trend of PD detection.