Research On Intelligent Control Technology Of High-penetration Distributed Photovoltaic Generation Cluster

Energy is the basis for human life.With the rapid economic development of the world,the global energy demand is also showing an increasing trend.Considering the deterioration of the ecological environment and the depletion of fossil energy in worldwide,the development and utilization of renewable energy has become a universal consensus and concerted action allover the world.Compared with developed countries in Europe and America,the development of distributed renewable energy in China shows a trend of regionalization and rapid development.Among these distributed renewable energies,the distributed photovoltaic power is developed fastest.The power with large-scale and high-permeability distributed photovoltaic has been formed.This kind of power grids faces the challenges of grid-connected large-scale distributed PV power generation cluster,such as,the change of power generation and load ratio,the conversion of power grid characteristics,the increase of power grid complexity and the difficulty of management and control.It has a great impact on the safety,reliability and economic of the power grid[1].At present,it is urgent to the control and manage the distributed power clusters.To solve these problems,the Long Short Term Memory(LSTM)load and PV output power forecasting methods are proposed.The advanced electrical distance is introduced considering the parameters and data of distribution power grid.The large-scale and high-permeability distributed local PV power grid is divided into clusters.Then the cuckoo search-particle swarm optimization(CS-PSO)is constructed for two-level control scheme to the power grid.The high-permeability distributed local PV power grid system is realized exemplary.The main research contents and innovations of this paper are listed as follows.1.The impact of distributed PV generation clusters to the distributed network is analized,the advantages and disadvantages of the current mainstream power flow calculation methods for distribution networks are compared,mathematical analysis and simulation calculation are carried out.The situation of voltage fluctuation and loss variation after the connection of high permeability distributed photovoltaic are compared under different capacity permeability(30%,70%,200%)with appropriate power flow calculation methods.It is found that the voltage fluctuation is restained and the line loss is reduced with the improvement of the distribution network operation,when the capacity of distributed photovoltaic power generation equipment is appropriate.The improvement is more obviously when the PV is connected with various nodes distributedly.However,with the continuous improvement of distributed PV permeability,the voltage deviation increases and the line loss will increase slowly.At this time,the advantages of distributed PV access for distribution network voltage have been lost.When large capacity distributed PV is connected,the voltage overshoot and line loss phenomenons are deteriorated in most nodes.The operating environment of the distribution network is much worse than that of the non-connected photovoltaic system.In this case,the continuous access of distributed photovoltaics should be forbidden.The analysis method in this paper is able to make a theoretical basis for the capacity and location of distributed photovoltaic access.2.The short-term forecasting methods for PV power generation considering load and generation uncertainty is used to analyze and accurately forecast the load of distribution network by combining frequency domain decomposition with Elman Neural Network(ENN)and random forest algorithm.Based on isolated forest(iForest)algorithm and LSTM Long Short Term Memory(LSTM)neural network algorithm,short-term PV power generation prediction is carried out.Load and distributed PV forecasting results can be used as the basis of distribution network control for high permeability distributed PV power generation clusters.3.The cluster partition and control technology of high permeability distributed PV power grid is studied,and an improved electrical distance based on impedance distance and sensitivity is used as the parameters of quantification and calculation,the cluster partition of high permeability renewable photovoltaic power generation is carried out with spectral clustering method.For the optimization of clustering result,Davies Bouldin Index(DBI),Silhouette Index(SI)and mean square error index are introduced as comprehensive evaluation indexes.The results of cluster partitioning are used as the basis for partitioning control of distributed photovoltaic power generation cluster.The centroid of cluster is taken as the controllable critical node.For IEEE33 node system and Jinzhai distributed photovoltaic power generation demonstration zone in Anhui Province,different clusters are effectively distinguished and critical nodes affecting each cluster can be identified,which can effectively reduce control costs.4.A new optimal planning algorithm for distribution network based on CS-PSO algorithm is proposed Voltage and line loss are taken as two optimization objectives,and the control variables of each node are reactive power and storage active power of controllable nodes.This algorithm is applied to the control of distribution network,and a double-layer strategy is proposed.The CS-PSO algorithm is applied to identify and adjust the position of the working point in a the stage of linear calculation.In each control step,based on node voltage and network loss,the global optimization calculation and precise regulation of active distribution network are carried out for nonlinear calculation.Compared with the traditional centralized dispatching control method,the proposed stritage can minimize the control cost and improve the grid-connected conditions of distributed photovoltaic cluster with the lowest operating cost.5.The construction and implementation of the photovoltaic power generation cluster control system,which is located in Jinzhai Anhui Province,in the high permeability distributed photovoltaic power generation cluster demonstration area is the field application of the theoretical research results in this paper.The effective solution for grid-connected absorption and dispatch control of high permeability distributed photovoltaic power generation cluster is provided,the power quality level of regional grid is improved,the disconnection risk and capacity of distributed photovoltaic power generation equipment is improved.The unqualified rate of bus voltage in static cluster has been reduced to less than 3%.The line loss rate in the static cluster decreased by about 30%.The income of household optical storage has been increased by about 30%.And the total harmonic current distortion rate decreased to within 4%.The whole demonstration area has the characteristics of observability,controllability and governance,which ensures the output,connection and absorbing of the high permeability distributed power generation cluster.