Font Size: a A A

Research On Distributed Optimal Control Strategy Of Active Distribution Network

Posted on:2020-01-29Degree:MasterType:Thesis
Country:ChinaCandidate:C WangFull Text:PDF
GTID:2392330590476844Subject:Power system and its automation
Abstract/Summary:PDF Full Text Request
The traditional centralized control mode of the distribution network has problems such as large calculation and communication pressure,low system reliability,and insufficient scalability of the control center.It can not meet the requirements of stable operation of the distribution network when the high-scale distributed power supply is connected.It is unable to adapt to the active demand of the active distribution network user side and the new demand of the active participation of the distributed renewable energy generation side.In recent years,the multi-agent system theory,which is structurally dissipative and autonomous and can cooperate to achieve the overall optimization control goal,has received more and more attention and research,and has been gradually applied in the power system.The research on distributed optimization control strategy can provide new ideas and methods for realizing the optimal operation of active distribution network.It is of great practical significance to promote the effective utilization of renewable energy generation and promote the construction and development of active distribution network.And practical value.Based on the multi-agent consistency theory,this paper makes an in-depth study on the multi-objective distributed optimization control strategy and control strategy performance optimization method for active distribution network.In this paper,an in-depth study on distributed optimization control strategy based on group cooperation mode is carried out.Firstly,the control structure of distribution network with distributed generation is introduced and the advantages and disadvantages are compared.Based on the distributed power group cooperation mode,the framework of multi-objective distributed optimization control system for active distribution network is designed.With distributed power utilization and power factor as the consistency variables,a distributed optimization control algorithm for group node voltage control,active distribution network active power and reactive power control is designed,and the algorithm parameters are set.In view of the fact that the resistance/reactance of the distribution network is relatively large,the distributed power supply active power is used to control the node voltage.The correctness and effectiveness of the distributed control strategy are verified by an IEEE33 node system.Secondly,this paper studies the communication weight optimization method.Firstly,the related knowledge including graph theory,matrix theory and consistency algorithm is introduced.The optimization problem of control algorithm convergence speed in control strategy is transformed into distributed fast linear iterative problem.Asymptotic convergencefactor and single step are proposed.The convergence factor is used as a measure of convergence speed and gives a mathematically rigorous proof process.Considering that the Laplacian matrix in the power communication network is generally a symmetric matrix,the optimization problem with two indicators minimized can be unified to the matrix norm optimization problem,and it is transformed into a semi-definite programming problem(SDP)by introducing a slack variable s.The optimal weight setting method for the Laplacian matrix can be obtained by Matlab LMI Toolbox.The effectiveness of the proposed Laplacian-based optimization weight setting method is verified by three test system analysis.Finally,this paper has carried out research on leading distributed power supply optimization methods.Firstly,the leader-follower consistency control problem is introduced,and the leader-follower agent individual and the system level mathematical expression are given.Secondly,the matrix theory knowledge is used to prove the stability of the leader-follower consistency control algorithm.As long as the communication network formed by the agent is connected and there is a follower agent connected to the leader,the state of all agents will eventually converge to the leader state quantity;then the leader from the leader-follower consistency control In the second chapter,the convergence rate of the control algorithm in the control strategy is proposed.The single-step convergence factor is still used as the convergence speed index.The convex relaxation variable s is introduced to convert the optimization problem into the mixed integer semi-definite programming(MISDP)solution.The heterogeneous network density to the communication network is solved by the MATLAB YALMIP toolbox based on MOSEK solver,and the binary information matrix about the optimal leader node is obtained.The optimal interconnect design based on the leader selection is proposed.Finally,taking the IEEE33 node system as an example,consider the case where the number of distributed power sources is selected as 1 and p(p?2),which verifies that the optimal selection of the distributed power supply can improve the distribution network.The convergence performance of the target distributed optimization control strategy is beneficial to its practical application in large-scale power systems.
Keywords/Search Tags:Distributed generation(DG), Distributed optimization control, Consistency algorithm, Communication matrix optimization, Leader-follower system, Multi-agent system
PDF Full Text Request
Related items