Improved Reactive Power Sharing in Islanded Micro Grids using Adaptive Virtual Impedance

Nowadays distributed generation has been given considerable attention due to the increase in energy consumption and the need for using renewable energies. So, the concept of the micro grids for connecting and coordinating different distributed resources via the local power control systems has been introduced. A micro grid connects to the network at normal condition of the operation but in order to increase the system reliability and to promote the generated power quality to feed the sensitive loads, they can operate in the islanding mode too. In the islanding mode of operation in order to supply the whole load and to maintain the balance between the supply and demand, these resources use the conventional droop methods. Generally, using the conventional voltage droop control due to the difference in the impedance of the feeder connect to the resources and also the asymmetrical nature of the network, dose not lead to an appropriate reactive power sharing between the resources. In fact, using conventional droop control creates an intrinsic compromise between the accuracy of reactive power sharing and voltage regulation of micro grid buses. Therefore, in the proposed method of this paper, the coefficients of the voltage droops characteristics of the grid resources are determined by solving an optimization problem using the genetic algorithm (GA). Also in order to improve the reactive power sharing, a novel control strategy based on the adaptive virtual impedance is presented. In this method the ideal reactive power sharing is achieved by using the virtual impedance and compensating the voltage drop of the lines. The virtual impedance setting is done through the data of the voltage of PCC bus in microgrid and sending via a low bandwidth communicational link to the local controller of each resource. The proposed method is robust against the network structure changes, presence of local loads, and interruption in the communication link. Simulation results show the effectiveness of the proposed method.