Research On Distribution Network State Identification Based On Set Theoretic Estimation

As a fundamental and important part of the Distribution Management System(DMS),distribution network state estimation can increase the accuracy of state variables and measurements.Thus,it can provide complete and reliable real time data for the following network analyses and control.Existing distribution network state estimation methods have some problems with determining the relationship between the estimated results and their actual values,which leads to the uncertainty of the credibility of the estimated results.Consequently,the estimated results cannot be utilized in the closed-loop automated control of the power systems.In order to solve these problems,the main works of this dissertation are listed as follows:Firstly,according to the structure characteristics of the distribution networks compared with transmission networks,dealing with the three-phase unbalanced situation,the detailed mathematical models of three-phase components,for instance,lines,transformers,loads and shunt capacitors,are built,which lays a theoretical foundation for the subsequent chapters.Secondly,in the case of bounded measurement errors,drawing lessons from the basic idea of the set theoretic estimation in transmissions systems,models are built for all kinds of information of distribution networks in the form of sets and the state variables are chosen to form the measurement equation and constraint equation.The problem of calculating intervals of state variables is transformed into the problem of solving the optimization model,which can be solved by modern interior point.Through seeking all the feasible solutions consistent with these priori information,this method can ensure that the true value must belong to the feasible solution set.The credibility of the proposed method is verified by illustrative examples.Meanwhile,it is also proved that the conservatism and computational efficiency can meet the requirements of the subsequent applications.Thirdly,for the existence of bad measurement data,the concept of outliers is defined and they are divided according to their properties.For the incompatible outliers,level division is carried out in the first place,and then searching process is conducted from the lower level to the higher level.For the compatible outliers,through the definition of the consistency indicator,they are classified into effective measurements and ineffective measurements and only the effective measurements are searched so as to narrow the searching process.The test results of standard distribution networks demonstrate the credibility of this outlier identification method.