Robust power management system with generation and demand prediction and critical loads in DC microgrid

In this paper, a robust power management system (RPMS) for a DC microgrid is proposed. A novel neural network-based scheme is proposed for the PV cells’ generation prediction using ultraviolet (UV) index, temperature, and cloud coverage which gives a considerable improvement in the prediction error in comparison with the existing works. Moreover, another neural network predicts the demand for the microgrid. The proposed RPMS will make decisions under the uncertainties of these prediction errors such that the system stays robust and works near the optimal operating point. Besides, three different possible scenarios for operation of the microgrid are considered which represents all real operating conditions. Then, three corresponding optimization problems are introduced for theses scenarios. Moreover, without loss of generality, load buses are clustered in one critical load bus and three sheddable load buses. The RPMS keeps the critical load bus voltage in a standard range while feeding the maximum possible sheddable buses with 0.9 p.u. voltage or disconnecting them sequentially. The numerical simulation results show the feasibility and effectiveness of the proposed strategy.