Research On Planning Of Distribution Systems With Distributed Renewable Energy Clusters

Aiming at the situation that large-scale distributed renewable energy is connected to the distribution network,this thesis studies the problem of planning the distribution system with renewable energy clusters,in order to enable renewable energy to be effectively absorbed in the distribution network.In order to make the distributed renewable energy accessed in some areas of the distribution network coordinate with each other,this thesis divides the distribution network into multiple areas based on the access situations of renewable energy and the network topology.The divided areas are defined as the renewable energy clusters.Then,the indicators and methods of cluster partition are researched.Then,based on the partition of renewable energy clusters,the problems of distributed generation planning and the joint planning of distribution networks,renewable energies and storages are studied.The specific work is as follows:(1)In the context of large-scale integration of distributed renewable energy to the distribution networks,the criteria for partitioning distributed renewable energy clusters and optimization algorithms have been improved.In order to define the range of the cluster,this thesis proposes two types of indicators to divide the cluster,namely the electrical distance index and the power characteristic index.The power characteristic index includes the index of cluster net power balance and the index of node complementarity degree.Based on the electrical distance index and the cluster net power balance index,the cluster partition method based on genetic algorithm is proposed.Based on the electrical distance index and the index of node complementarity degree,the cluster partition method based on linear programming is proposed.An example system was used to verify the advantages and disadvantages of the two cluster partition methods.It was proved that the linear programming model can find the global optimal cluster partitioning scheme more efficiently and avoid the defect of genetic algorithm of falling into the local optimal solutions.(2)Based on the static cluster partition strategy,renewable energy clusters are set as the range of the power interactions,the generation planning is studied.First,for the scenario of the planning problem,the correlated scenarios of renewable energies and loads are generated based on Pearson correlation coefficient and Spearman correlation coefficient.The obtained correlated scenarios are suitable for renewable energies and loads that meet any probability distribution.Based on the obtained correlated scenarios and cluster partition strategy,the distributed generation planning model is established taking into account the investment and generation benefits of power producers and the degree of power matching within the clusters.The Benders decomposition method is used to solve the proposed distributed generation planning model.The actual system example is used to verify that the proposed planning strategy considering cluster partition and correlation of renewable energies and loads can improve the consumptions of renewable energies within the distribution networks.(3)Aiming at the planning problem of networks-resources-storages considering static cluster partition,the planning of network and renewable energies and the storage planning is conducted in two steps.Aiming at the problem of joint planning of network and resources,a multi-objective planning model that considers network investment,operation costs,reliability costs,as well as renewable energy investments and operation costs is proposed.The augmented ? constraint method is introduced to transform a multi-objective problem into a single-objective problem.According to the differences on the calculation method of reliability index,two kinds of planning methods are discussed: 1)the calculation of reliability index based on Monte Carlo simulation method and the slack planning model of network-source,the reliability evaluation and the planning is decoupled to solve the problem.The limitations of this scheme is discussed;2)Based on the analytical method,a calculation method for reliability analytical expression considering 0-1 variables of lines and island operation probabilities of clusters is proposed.The reliability analytical expression is linearized,and then is embedded into the planning model of network-source.Then,a convex optimization model is established,and the Cplex solver is used to solve the model.This method can effectively avoid the disadvantages of the decoupling method.Based on the island operation probabilities obtained by the convex optimization scheme of the network-source joint planning.The second step is to perform storage planning to ensure that each cluster can meet the reliability of power supply under the island operation mode.An example is used to verify that the proposed method of the planning of network-resource-storage can improve the reliability of system power supply and reduce power transmission losses and voltage fluctuation.(4)For the new area without network planning,this thesis proposes an iterative method for dynamic cluster partition and the planning of network-resource-storage considering the uncertainties of resources and loads.In the iterative method,due to the large number of lines to be planned in the new area,it is impossible to directly establish a planning model for all the lines of the entire network.Therefore,the planning of network-resource-storage is divided into the network planning outside the clusters and the planning of network-resource-storage within the clusters.The cluster partition and planning process take into account the uncertainties of resources and loads,and the box uncertainty set is used to describe the uncertainty of photovoltaic,wind power and load respectively.Based on the iterative method,the robust model of cluster partition,the robust model of network planning outside the cluster and the robust model of the planning of network-resource-storage within the clusters are proposed.Based on the dual theory,those two-layer robust optimization models are transformed into single-level programming models that can be directly solved by the Cplex solver.The effectiveness of the iterative method of cluster dynamic partition and the planning of network-resource-storage for the new areas is verified by examples.The economic benefits of the robust optimization model and the deterministic optimization model are compared,and the effects of the uncertainty degrees on economic benefits are also compared.The distribution system planning theory considering distributed renewable energy clusters proposed in this thesis provides theoretical support for the planning problems of distribution systems with large-scale high-permeability renewable energy.And the planning strategies also provide certain reference significance for the orderly integration of high penetrated renewable energy and the coordinated optimization for "resource-network-load-storage".