Optimal Allocation Of Distributed Generation In Active Distribution Network

With the increasing development of smart grid and its technology, the distributed generation(DG) attracts support from society. Since the output of DG is unstable, in order to respond to the DG’s characteristics, the traditional distribution network is being changed into the active distribution network(ADN), which has the active management capability. ADN planning is characterized by the combination of active management with planning, control and operation function of distribution network. In the stage of optimally allocating DG, active management model is fully considered to control the operation of distribution network. So it is significant for us to research about the optimal allocation problem of DG in ADN.In the beginning, the characteristics of ADN, along with the output characteristics of typical DG and the influences affected by DG accessed to the distribution network are analyzed. Then the business model of DG is studied, additionally, from the aspect of independent investor, the break-even analysis is calculated under the governmental policy.On the base of the maximum load, the maximum capacity of grid-connected DG in ADN is analyzed, which can guide the optimal allocation of DG in ADN.In order to consider about the unstable characteristics of load and DG, we refer to the time series characteristic, which is based on hour unit in 24 hours in a day. Different kinds of load and DG under the condition of four seasons multiplied by 3 kinds of weather scenes like sunny, cloudy and rainy, totally 12 scenarios, are considered. In addition, time-of-use power price system is utilized to reflect the demand side characteristic of ADN.On the basis of ADN characteristics and DG planning, the optimal allocation method of DG under active management model is proposed. A bi-level planning model is discussed, in which the objective of upper level planning is to maximize the profit of the ADN operator, while the objective of lower level planning is to minimize the curtailment of DG by active management. Additionally, the combination of adaptive multi-particle swarm shuffled optimization algorithm and the Prime-Dual Interior Point Method to solve the upper and lower planning model respectively is proposed. While in the upper planning model, the sites and sizes of the DG are given in the stage of initialization, the cite and size information of DG is transferred to the lower planning model, under 12 scenarios, the optimal power flow is calculated under the objective of minimizing the curtailment of DG by active management, then we can get the yearly profit of the ADN operator. Then by using the adaptive multi-particle swarm shuffled optimization algorithm, we can get the optimal particle, namely, the planning method of DG in ADN. Finally, the feasibility of the model and the effectiveness of the algorithm are verified with a distribution system.