Research On Distribution Network Large Data Fault Diagnosis And Distributed Wind Power Configuration For Emergency Repair

In response to the development of intelligent distribution network service from passive to active,the fusion and processing of multi-source fault data is an important prerequisite to clarify the fault situation and master the fault law.In order to ensure the stability and efficiency of the following active repair and dispatch services,it is necessary to fuse large data from multiple sources in the intelligent distribution network and to make sure that the optimal configuration of the distribution network with distributed power supply is in line with the stability constraints of the system.For this purpose,the following work is done in this paper:Firstly,the data of emergency repair are preprocessed,the data from different sources are checked hierarchically,and the overall data situation and time-effect are judged.The overall quality of the data is predicted.An anomaly detection method of emergency repair data based on OPTICS(Ordering points to identify the clustering structure)algorithm is established.Then,clustering analysis of emergency repair big data,extracting data anomaly items,through the algorithm settings and case analysis,it is clear that this method can still effectively identify data anomaly for emergency repair big data with fault data.Secondly,it clarifies the logic of distinguishing various kinds of faults,finds out the main influencing factors of various kinds of faults,and chooses characteristic quantities for fault diagnosis.Combining with the data fusion model,this paper puts forward a pertinent processing method for various fault data characteristic quantities.The function space of each eigenvalue for fault type evaluation is defined,and a comprehensive fault diagnosis model of distribution network based on DS(Dempster Shafer)evidence theory is proposed.According to the time domain characteristics of the fault,the fault is fused in time domain first,and then in space domain between platforms,so as to achieve the purpose of comprehensive research and identification of the fault.Finally,stochastic fuzzy variables are used to describe wind speed,which covers more abundant uncertain information,so as to enhance the distribution system's ability to deal with N-1 faults and to realize load-loss switching after faults.By judging the static safety index of the system and considering the AM(Active Management)measures,combining the fan capacity allocation planning with the operation of the planning scheme,a stochastic fuzzy chance-constrained programming bi-level programming model considering time series characteristics is proposed.The simulation of IEEE 14-node distribution network system demonstrates that this method fiully takes AM measures into account in the planning stage to improve the operation of distribution network,and has better adaptability under the premise of improving DG(Distributed Generation)acceptance.In this paper,the fault situation of distribution network is deeply studied,and the distributed uncertain configuration of distribution network under fault data processing and diagnosis is considered.From the fault diagnosis of credible data to the reasonable allocation scheme after fault,it can provide reference for active distribution network technology and optimal allocation and analysis of wind power system.