A Distributed Coordination Control based on Finite-time Consensus Algorithm for a Cluster of DC Micro grids
Abstract: A cluster of DC micro grids consisting of multiple interconnected DC micro grids has great potential for improving the reliability and reducing the cost of power generation. In this paper, a distributed coordination control strategy is proposed to overcome a series of problems of centralized control. This control strategy which adopts a hierarchical structure is based on the droop control method, then a modified control formula which achieves multi-objective is introduced to the voltage control and power generation cost control. Besides, a finite-time consensus algorithm is applied to obtain the average value obtained in finite steps. Meanwhile, an estimated connected topology is proposed for better utilization of the finite-time consensus algorithm and acceleration of the convergence speed. This proposed topology ensures all this control method is a fully distributed one, which does not require the global information previously and has high reliability. By using the Matlab/Simulink simulation platform, this proposed control strategy is verified in one DC micro grid and a cluster of those containing 4 DC micro grids. The simulation results under different scenarios all indicate the effectiveness of the proposed methods.
Existing system: Compared with the existing control methods, the control method proposed can achieve multi-objective. And for any of the distributed connected systems, this control method is suited which can ensure the convergence in finite steps. There is no need to collect information centrally update local parameters and achieve consensus. Whatever for single DC micro grid or a cluster of DC micro grids, the estimated connected topology guarantees that the control method meets the requirements of the distributed control. All distributed units do not need to obtain the global topology and only exchange information with adjacent units which can ensures the information privacy of each unit. Meanwhile, this topology can further reduce the number of iterations times and accelerating convergence with the number drop of the distinct nonzero Eigen values.
Proposed system: Centralized control is used to manage all the units in the whole system through the central controller, which is simple and convenient. However, when centralized controller fails or the communication structure changes, centralized control can’t complete the corresponding objectives, which will reduce the reliability of the system. Distributed control overcomes the drawbacks of centralized control, each unit only implements optimization control based on information of itself and its neighboring units. Therefore, under the condition that the communication is limited or the communication structure changes, the distributed control can still complete the corresponding control tasks and meet the requirements of ―plugand-play‖. Over the past years, several distributed methods have been proposed. Advantages: Centralized control is used to manage all the units in the whole system through the central controller, which is simple and convenient. However, when centralized controller fails or the communication structure changes, centralized control can’t complete the corresponding objectives, which will reduce the reliability of the system.
Constraints. It uses the estimated mismatch between demand and supply to feed back to the consensus algorithm, and then the incremental cost converges to the optimal value. In, a fully decentralized economic dispatch (ED) approach with the flooding consensus algorithm was established. Disadvantages:
Meanwhile, compared with AC micro grid, the bus voltage stability is the only standard of active power balance in DC micro grid, and there is no such problem as frequency stability, reactive power compensation and so on. In , a distributed DC-OPF approach solves local OPF problems of individual subsystems in parallel, which are coordinated via global consensus variables. A consensus-based distributed voltage control (DVC) method is proposed in, which guarantees reactive power sharing in meshed inverter-based micro grids with dominantly inductive power lines and arbitrary electrical topology. Modules: Distributed voltage control: In the flooding consensus algorithm, data in one sub-system are sent to all other sub-systems, which could exacerbate the communication bottleneck. A selfadaptive droop control strategy based on distributed consensus is proposed in, using consistency iteration to obtain the average voltage difference of the whole network, and dynamically find a virtual resistance that meets the requirements of the current sharing and voltage regulation. In, a distributed DC-OPF approach solves local OPF problems of individual subsystems in parallel, which are coordinated via global consensus variables. A consensus-based distributed voltage control (DVC) method is proposed in, which guarantees reactive power sharing in meshed inverter-based micro grids with dominantly inductive power lines and arbitrary electrical topology. For island AC micro grids, a fully distributed coordination controller including both containment-based and consensus-based controllers is proposed to achieve both bound the voltage magnitudes within a reasonable range and achieving accurate reactive power sharing. In , relying on the
mean-metropolis-based consensus algorithm, a distributed control approach is proposed to maximize social welfare of all the participants and maintain. DC Micro grid: ITH the increasing proportion of renewable energy sources in the grid, micro grid has attracted wide attention as an important form of distributed energy resource. The micro grid contains a large number of DC power sources such as photovoltaic, fuel cells, and energy storage, as well as DC loads such as LED lighting and electric vehicles. Meanwhile, compared with AC micro grid, the bus voltage stability is the only standard of active power balance in DC micro grid, and there is no such problem as frequency stability, reactive power compensation and so on. Therefore, the DC micro grid has wider research scope in the academic and industrial application. Recently, due to geographic boundaries, main power supply and functional differences, a number of DC micro grids will form a cluster within a certain power supply area. These DC micro grids are interconnected and operated in the form of clusters, and each DC micro grid achieves mutual support control through group energy scheduling and coordination control. For single DC micro grid or a cluster of DC micro grids, the operation control methods can be categorized into two types: centralized and distributed. Consensus control strategy: In this paper, a fully distributed finite-time consensus control strategy considering multi-objective for a cluster of DC micro grids is proposed. This control strategy is divided into three layers, the first layer is the traditional droop control, the second and third layers modify the original droop equation by using the finite-time consensus algorithm, realizing the bus voltage stability and the minimum cost of power generation respectively. Compared with other distributed control strategy, the proposed distributed finite-time consensus control strategy is flexible and efficiency against various operating conditions and its convergence is guaranteed through rigorous analysis. The effectiveness of the proposed finite-time convergence control strategy is demonstrated through simulation studies in the MATLAB/SIMULINK software environment. Estimated Connected Communication Topology:
For distributed control systems, each unit only knows itself and its neigh bour’s information, so the Laplacian matrix L depends on the global information of whole connected communication topology which cannot be obtained as a previous condition. Besides, the finite-time consensus algorithm also needs to calculate the average value by the Laplacian matrix L and the nonzero eigenvalues. Therefore, in order to make the control method suitable for more application scenarios, the estimated communication topology is proposed to solve this problem in this paper. The following explanation is given firstly: For each DC micro grid, there are different communication topologies. And for each communication topology, the corresponding Laplacian matrix L is unique. Let L1 represent the corresponding Laplacian matrix of the connected topology T1, for n different connected topologies, T={T1,T2,‌,Tn}, and the Laplacian matrix set is L={L1,L2,‌,Ln}