Analysis And Research On Distribution Line Fault Based On μPMU Measurement

The safe and reliable operation of the distribution line is the basis for ensuring the reliability of the power supply.With the gradual deepening of the development of the distribution network intelligence,there is a higher requirement for the recovery of the distribution line after the fault,so the fast after the fault line accurate diagnosis is of great significance to improve the reliability of power supply.Limited by the defects of low sampling accuracy and low sampling frequency of traditional measuring devices,many fault diagnosis techniques cannot be applied;microsynchronous phasor measurement unit(μPMU)can provide sufficient electrical quantity information for fault diagnosis of distribution network lines.On this basis,combined with the data characteristics of the μPMU and the electrical characteristics of the distribution line after the fault,the three aspects of fault line selection,fault classification,and fault location are respectively studied.First of all,in terms of fault line selection,in order to solve the current problem of difficult line selection for single-phase ground faults in distribution lines under resonant grounding mode,for the shortcomings of the current line selection method,the empirical mode decomposition is used to extract the transient high-frequency component of the zero-sequence current is taken as the main feature of fault line selection,and the average current amplitude difference of the auxiliary feature is used to define the reliability of the fault as the basis for judging the line selection in this paper.In PSCAD / EMTDC,a distribution line simulation model for fault line selection was established,and the effectiveness of the line selection method proposed in this paper under various fault factors(fault distance,transition resistance,fault initial angle)was verified through simulation.Secondly,in terms of fault classification,in order to distinguish the types of short-circuit faults,three types of fault characteristics(three-phase correlation coefficient,three-phase energy coefficient,and zero-sequence current)have been established;In terms of classifier,in order to provide high accuracy classification,combined with the random forest classifier in integrated learning for short circuit fault classification.Build a simulation model in the simulation software,extract a large number of different failure samples by changing various failure conditions,and use them for training and testing of the classifier respectively.The test results show that the classifier established in this paper has high classification accuracy.Finally,in terms of fault location,in order to achieve the accurate location of the fault location,a set of ranging system with parameter identification and fault location is established.Combined with the characteristics of distribution lines,mathematical models for parameter identification and fault location were established.In terms of real number optimization algorithm,adaptive genetic algorithm is applied to achieve low error fault distance measurement.The applicability of the fault location method proposed in this paper under different fault factors(line type,fault distance,fault initial angle)is verified by simulation,and compared with the traditional fault location method to verify that the fault location method proposed in this paper can achieve high-accuracy ranging.