An improved hierarchy and autonomous control for DC microgrid based on both model predictive and distributed droop control

Direct-current (dc) microgrids (MG), consisting of distributed renewable energy units and energy storage units, is expected to be the key enabling of future smart grid. The intermittent nature of renewable-energy units, coupled with the unpredictable changes in the load, requires the energy storage units compensate the fluctuating generated power and to regulate the dc-bus voltage. However, the energy storage units may not be always available, each energy unit converter should be able to switch between two different modes: current course converter to generate/consume power or voltage source converter to regulate the bus voltage. To address these two main challenges, a novel autonomous algorithm consisting of two layers of control is proposed, achieving good system dynamic, seamless transfer and decoupling performances. The primary layer control for each energy unit is based on model predictive current control, realizing free controller design and decoupled play & plug feature. Therefore, these energy units can be easily connected to the dc bus without affecting the operation of other converters. The secondary layer control based on a proposed distributed droop control determines the operation modes for each converter, either to be current source converter (CSC) or voltage source converter (VSC). The feasibility and effectiveness of the proposed control algorithm was verified under various case studies on dSPACE 1007 real-time simulation platform.