Research On Line Fault Identification And Location Method Of Distribution Network Containing Distributed Generatio

The distribution network structure has become increasingly complex,with many branch lines and various operating states,and is prone to various types of faults.Due to the harsh and variable fault conditions,it is difficult to extract the fault signal characteristics,especially with the widespread application of distributed generation(DG),which makes the fault identification and localization problems of distribution networks more complicated,resulting in the results of the existing fault diagnosis methods are difficult to achieve the expected results.Therefore,it is necessary to carry out research on fault identification and localization in DG-containing distribution networks to meet the needs of distribution network diagnosis and protection.Firstly,this paper introduces the current research status of fault identification and localization at home and abroad;then analyzes the steady-state and transient fault characteristics of single-phase ground fault in small-current grounding system,and summarizes the change law of characteristic quantity after the occurrence of fault;at the same time,it summarizes the principle of grid connection of DG,analyzes the impact of system connection to DG on the fault identification and localization of traditional distribution network,and builds a simple simulation model of distribution network with inverter-type DG.Secondly,the identification of line fault types in DG-containing distribution networks is studied,and a combination of Modified Ensemble Empirical Mode Decomposition(MEEMD)and Convolutional Neural Network(CNN)is used for fault type identification.The signal analysis method MEEMD is used to decompose the three-phase voltage,three-phase current and zero sequence voltage before and after the fault on the low-voltage side of the main transformer of the system one by one,build them into a time-frequency matrix,and transform them into the form of a gray map,and then obtain a time-frequency gray map to realize the transformation from one-dimensional time-domain information into a two-dimensional image;then the dimensionally adjusted time-frequency gray map is input into the CNN,and the fault features are extracted by the CNN autonomously for classification to realize the identification of line fault types in distribution networks containing DG.Thirdly,based on the results of fault type identification in DG-containing distribution networks,a fault line selection study was conducted for single-phase grounding faults with high occurrence rate.Since the fault signals are often weak and contain noise components,there are difficulties in fault feature selection and extraction,which causes the problem of poor routing results.To this end,an adaptive Variational Mode Decomposition(VMD)based single-phase grounding fault sizing method was proposed.The method uses parameter-optimized VMD to adaptively decompose the zero-sequence current of each line,extracts the transient characteristics of its low-frequency components for two-by-two correlation calculation,and then finds the integrated correlation coefficient to realize fault line selection.The experimental results show that the line selection method is not affected by DG access and can adapt to various fault conditions,and its reliability is verified by the measured data.Finally,on the basis of successful line selection,a single-phase ground fault section location method based on multiple fault characteristics was proposed for multi-segment and multi-branch DG-containing hybrid lines.Considering the limitations of single fault characteristics for location,the fault section location is achieved by using the distribution characteristics of transient zero-sequence current in each section,extracting transient direction characteristics,transient energy characteristics and similarity characteristics for comprehensive comparison.The effectiveness of the method is verified by simulation,and the accuracy is still guaranteed with good adaptability in the case of high-resistance ground fault,noise interference and missing data.