Research On Application Of Precision Planning Of Distribution Network

The distribution network is an important part of the power system,which is located at the middle and end of the power network.The main function is to realize the distribution of electric energy.Distribution network planning is to build a safe,reliable,economical and environmentally friendly distribution network system on the basis of meeting the electricity demand of power users in the region.On the one hand,with the proposal and construction of active distribution network,ubiquitous Internet of Things,digital power grid and energy Internet,higher requirements and standards are put forward for distribution network planning.On the other hand,problems such as scattered data,unqualified data,single planning mode and lack of planning means have always plagued distribution network planners.All this has led to the construction of the distribution network not keeping up with the demand,new problems will arise after the construction of the planning scheme,and the planning efficiency has never been improved.How to realize the effectiveness and accuracy of distribution network planning and effectively guide the construction of distribution network using planning technology,has important research significance.Therefore,this thesis takes precision as the core idea and application as the ultimate goal,and explores the precise planning method of power distribution.Firstly,the business process of precise planning of distribution network is introduced.The functional requirements of precise planning of distribution network are explained: data integration,status problem diagnosis,load forecasting,planning proposal,project optimization and electrical advanced calculation.The distribution network Data requirements for accurate planning are analysised,a list of typical problems is sorted out.Three main problems faced by accurate planning of distribution network are described:insufficient power load data quality,low efficiency of diagnosis of current problems,and unreasonable planning schemes.It is pointed out that this thesis mainly aims to solve these problems.Three problems are targeted to carry out the application research of precise planning of distribution network.Secondly,in view of data missing and mutation drift in power load data,three general methods for improving data quality,linear interpolation,convolutional neural network and fuzzy C-means clustering,are introduced,and their advantages and disadvantages are explained.Based on the low-rank characteristics of load data,it is found that the use of low-rank characteristics of power load data can solve its data missing and mutation drift problems.The algorithm solves the recovery model,and the soft threshold operator and the singular value reduction operator are introduced in the solution process.The data of the model and algorithm in this thesis are verified in data missing scenarios,mutation drift scenarios and mixed scenarios through the analysis of the Irish user data set.Compared with the general method,the method in this thesis has achieved a more ideal recovery effect.At the same time,the method in this thesis is applied to the actual data set.The application results show that the method in this thesis can meet the needs of practical applications.Then,the characteristics of the multi-source heterogeneous data of the distribution network are analyzed,and the integration of the multi-source heterogeneous data is realized.The specific process of automatic diagnosis of the current situation of the distribution network is proposed.The multi-level information and topology static model of "station-line-transformation",using the Java programming language to realize the problem identification program,taking the feeder overload problem as an example,the specific realization process and realization effect of the distribution network automatic diagnosis method are explained in detail.Compared with the problem database sorted out by the planners of a certain city bureau,the validity of the automatic diagnosis model in this thesis is verified,and it is proved that the model in this thesis can significantly improve the accuracy and efficiency of problem diagnosis,and provide planning reference for planners.It has practical application value.Finally,some general solutions to the current situation of the distribution network are proposed.Taking the problem that the feeder does not meet the minimum configuration requirements for self-healing as an example,a bi-level optimization model of sectionalizing switches and tie lines considering operation mode adjustment is established.The minimum sum of purchase and installation costs,operation and maintenance costs,customer power outage loss costs and line loss costs are taken as the objective function,and binary particle swarm optimization and Cplex solver are used to solve the upper and lower layers of the model.The method in this thesis is applied to the RBTS test system.In the BUS2 simulation network and an actual distribution network,the experimental results show that the method in this thesis can not only effectively solve the current situation of the distribution network,but also minimize the total investment cost and realize the precise planning of the distribution network.