Optimized settings of droop parameters using stochastic load modeling for effective DC microgrids operation

Droop control is a widely used technique for load sharing in DC microgrids (MGs); however, its performance is affected significantly by the voltage drops over coupling line impedances causing load sharing errors and voltage degradation across MG bus. Therefore, the locations of the sources in the MG and loading conditions (LCs) have considerable effect on load sharing accuracy and voltage regulation. In this paper, an optimization procedure that takes the MG topology and the load behavior into consideration is proposed to find optimal droop parameters minimizing the impact of the line impedances. In the optimization problem, the required constrains are formulated and the cost function is selected as a combination of current sharing error and the voltage degradation for various LCs. The impact of the different LCs on the cost functions are weighted based on the probabilities of their occurrences that are obtained from the load stochastic model. Particle swarm-based technique is then implemented to provide a solution for defined optimization problem. The performance of the droop controller with the optimal droop parameters is verified through a case study performed on MATLAB/Simulink environment and found to be very effective in achieving the required performance while maintaining inherent droop control such as the stability and robustness.