Fault Feature Mining And Fault Cause Identification For HVAC Transmission Lines

High voltage transmission lines work as an important part of the grid,covering widely.Exposed to wicked weather and geographical conditions,high voltage transmission line faults occur frequently due to natural disasters and human activities.As the backbone of the power network,faults of HVAC transmission lines above 220kV will inevitably bring the impact to the power grid and threaten the safe and stable operation of the grid.At the same time,there is no obvious fault traces because of the rapid action of the relay protection device,increasing the difficulty of fault searching.After such an outage has occurred,quick and correct identification of the fault cause would contribute significantly to narrow down the searching areas and expedite the restoration procedure,which helps increase the system reliability and availability,improve customer satisfaction,and reduce economic loss.At present,little research has been studied on the identification of transmission line faults at home and abroad,most of the work is based on the statistics of some faults and the prevention and control measures.No systematic identification method has been worked out.In this paper,we study six typed of single-phase faults of transmission lines caused by lightning,wind,bird damage,insulators contamination,tree contact and wildfire.Based on the deep nderstanding of the fault mechanism,fault features are analyzed in depth.Studies based on external characteristics and internal characteristics of the faults are summarized combing the historical experience:fault weather,season,time,the reclosing situation,the dc component and the third harmonic component of the fault phase current,zero-crossing distortion of the fault phase current and the transition resistance.Analysis based on the actual fault recorder data verify their valid.Furthermore,fisher fractional method is used to realize the importance degree of the features of the models and the main factors are sorted out,so as to establish different fault identification model.In view of the incomplete fault of the transmission line,this paper adopts the SVM algorithm with strong ability to realize the classification and identification of the cause of the fault,which can avoid the over-learning problem to a certain extent.The relevant features are introduced for each fault category.Six types of the fault cause identification model are established and trained using these features,and PSO is adopted to optimize the parameters or the moaeis.The test sample is brougnt into eacn tault moaei,ana the probability of the fault type is obtained,and the reason of the maximum value is the result of the judgment.The results show that the performance of the SVM algorithm based on PSO is improved,and the important features of each type can be obtained.In conclusion,according to the comprehensive feature mining of six types of transmission line faults,the paper proposes a self-learning fault cause identification method based on PSO-SVM combing external characteristics and internal characteristics,which can effectively identify various fault cause types.On the basis of calculation and analysis of actual fault data,the proposed algorithm has strong theoretical basis,and it is proved to achieve high identification accuracy.Moreover,the fault information needed in the method can be obtained easily and timely,and it can be extended combing the fault features of the studied area to ensure the identification accuracy,which can meet the practical engineering requirements.