Research On Distribution System Optimal Operation And Service Restoration Methods Based On Linear Fitting Model

In recent years,the increasing penetration of distributed generations in the distribution system has changed the operation mode of distribution network and brought many challenges to the optimal operation and service restoration of distribution network.Therefore,faster and more reliable strategies are called to deal with the optimization problems of distribution network.Related researches often based on the known configuration line equipment parameters of the system,adopting the method of mathematical programming with open source or commercial solver to solve the optimization problem.The model contains operation constraints and for the purpose of calculating the optimal power injection or service restoration strategy.Due to the complexity of the three-phase asymmetric power flow model,existing researches often model the distribution network as a three-phase power flow model,under the assumption that the system network loss can be ignored and the voltage level is close to the base value,then,reform it as a linear model to improve the optimization efficiency.However,distribution network is a three-phase asymmetric system and the network loss can not be ignored,so direct linearize the model will lead to large errors.Therefore,in this thesis,based on the simulation data,we analyzes the linearity level of the key physical parameters of the distribution network,studies the linear fitting method between the parameters,and proposes to linearly fit the operation constraints of the distribution network based on multiple linear regression,and further proposes a data-driven distribution network optimal operation model and optimal service restoration model,the validity and accuracy of the proposed model are verified through a large number of simulation experiments and example tests.The main work of this thesis is organized as follows:First,the linearity of the power flow equation of the distribution network and the linear fitting method are proposed.The linearity between the power injection and the voltage magnitude of the two-node three-phase symmetrical distribution network is studied by formula derivation and simulation analysis.The power flow equation of distribution network based on linear fitting is established.The distribution network simulation software is used to generate the data set,and the undetermined linear system is obtained through the fitting method of multiple linear regression.In the normal voltage operating range,for the distribution system with low impedance level and the overall load is not excessively overloaded,parameters are approximately linear.The accuracy of the power flow model based on linear fitting is verified through cases.Then,the optimal operation method of distribution network based on linear fitting is studied.On the basis of the power flow model obtained by linear fitting,an optimization model based on linear fitting is established for the minimum power loss problem and the minimum voltage deviation problem.The performance of this model,semi-definite model and classical analytical linear model in IEEE standard cases are compared.Tests show that the optimization model proposed in this thesis can obtain the optimal power generation strategy of the distributed power sources of the distribution network through optimization decision-making under different goals.Finally,the service restoration problem in distribution network based on linear fitting is proposed.The service restoration problem aims at maximizing the weighted recovery load with integer variables.Based on linear fitting,a mixed integer linear programming model is established.The tests show that the model based on linear fitting can obtain the same restoration strategy as the mixed integer semi-definite model in most cases with a higher solution efficiency.And in different cases,the restoration strategy of this model is more accurate than the classic analytical mixed integer linear model.