Active/reactive power management in islanded microgrids via multi-agent systems

•The proposed strategy aims to balance active and reactive powers to ensure proper steady state operation of islanded MGs.•A two-stage hierarchical approach using intelligent agents distributed in three levels is implemented to ensure efficient and fast responses.•The strategy relies on forecasted short-term data to calculate the optimal operating conditions of the MG for the next period.•In real time operation, expert-system rules are applied to attempt eliminating the mismatch in the forecasted data.•A guideline for the selection of the communication system is provided to avoid tripping the protective devices due to frequency or voltage deviations. This paper deals with the power management problem in islanded microgrids. The paper proposes a two-stage strategy to ensure the proper operation of islanded microgrids under different operating conditions. The first stage is an operational planning stage for optimal power scheduling of the microgrid assets. The objective is to minimize the power losses if the generation is expected to be higher than the demand, otherwise, the objective is to minimize the shed loads. This stage is based on the forecasted generation/demand data and is performed periodically. The second stage aims to operate the microgrid in real-time based on the scheduled powers and deals with the possible deviations due to the forecasting errors of the first stage. The proposed strategy is designed using a hierarchical approach to ensure close to optimal operation and fast response in real time operation. In addition, guidelines for calculating the execution time of the proposed strategy are presented in the paper to help in selecting the suitable communication system to sustain the islanded microgrid. The multi-agent system is used in this paper to implement the proposed strategy in the JADE environment while the power flow calculations are performed using Matlab platform. Several case studies are presented, and detailed results are included to demonstrate the effectiveness of the proposed strategy.