| With the further deepening of the reform of the power system,more and more stakeholders are taking part in the distribution network.The distribution network thus has many characteristics such as the increase of distributed energy resources,frequent data interaction,enhancement of time-varying characteristics of energy flow and so on.Problems like the insufficient of renewable energy consumption,power supply,the contradiction of different regions,enhancement of non convex and nonlinear characteristics of the system.The best way to solve this problem is to implement active control of distributed energy sources as well as developing active distribution network technology and realizing the coordinated optimal operation control of distribution network in various regions.Focusing on this topic,this dissertation mainly aims on the following problems:achieve the multiple source coordination optimization of active distribution network,solve the contradiction of power line schedule issued by different regional distribution networks and the inconsistent optimization objectives between different stakeholders.Based on the concept of multi-agent technology and decomposition coordination,an optimization model for coordinated management of layered and partitioned active distribution networks is established.First,a large distribution system is decomposed into multiple independent sub partitions according to the position of contact line switch and that of distributed energy resources,the regional agent optimize the resources in the management area to achieve the goal of improving the ability of the intermittent energy consumption,the quality of power supply and the operating income,reducing the operation cost and pollutant emissions and so on.Through the exchange of information and boundary variables with agent of global distribution network to coordinate the contradiction parts and last,regional agent optimize again according to the instructions issued by global agent to finally achieve the day ahead schedule optimal dispatch of ADN which takes both global and regional optimal objectives into account.For the optimization of the global distribution network,an optimization model considering the minimum active power loss of the network,the lowest operation cost and the minimum node voltage deviation taken as the objective functions,the power flow balance,the power exchange with the sub area distribution network and the output of each micro source taken as the constraints is established.For the sub distribution networks which seek for different interests,an optimization model considering the objectives on the aspects of economy and stability besides minimizing the operation and maintenance costs,maximizing the power selling incomes and benefits of energy conservation and emission reduction as well as considering the constraints of exchange power obtained from global optimization,the output of micro resources and the balance of power flow is established.Aiming at the bi-level optimization model,this paper adopts the fast non-dominated sorting genetic algorithm with elitist strategy based on the Pareto frontier concept,which can effectively handle the conflict among multiple objective functions and achieve the best comprehensive benefit.Adopting the maximum membership degree method for choosing from the Pareto-optimal solutions.On the other aspect,the discrete genetic operators are introduced to improve the algorithm so that it can be optimized in continuous space as well as in discrete space,and it can deal with the optimization problems which contains both continuous variables like the output of generators and discrete variables like number of shunt capacitors.Lastly,examples validate the analysis using the PG&E 69 case,the results show that the proposed hierarchical partition optimization strategy can improve the absorptive capacity of the distribution network of renewable energy as well as coordinating the contradiction between the regional connection power lines rationally,so as to achieve the ultimate goal of global and regional synergy. |