Real-time optimal dispatch for large-scale clean energy bases via hierarchical distributed model predictive control

Due to the multiple randomness from source and load sides, real-time dispatching of a large-scale clean energy base (LSCEB) with vast geographical area, numerous physical devices and complex interactive characteristics faces enormous challenges. To address this issue, this paper proposes a novel dispatching strategy named hierarchical distributed model predictive control (HDMPC). Firstly, by extracting the energy flow network of a generic LSCEB, the HDMPC strategy is finely designed, including the upper centralized manager (CM) layer and the lower distributed manager (DM) layer. The former has a three-level nested optimization problem with the day-scale, hour-scale and minute-scale cumulative objectives, respectively, for all the units in the source-side. The latter is a distributed model predictive control (DMPC) problem with one MPC controller for the optimization dispatch on the source side and the other one MPC controller for the optimization dispatch for the heating network on the grid-side. The two MPC controllers are sequentially interconnected for collaborative optimization between the source and grid side. Secondly, to provide accurate equation constraints for the above optimization problems, a multi-domain hybrid semi-mechanism modelling (MD-HSM) scheme is presented. Corresponding to the real-time dispatching task with time period of five minutes, detailed evaluation and selection of each unit's model are executed covering the source, grid and load sides. Finally, compared with the existing optimal economic dispatching strategy, simulation results show that the real-time optimal dispatch via HDMPC can achieve better operational economy, safety and flexibility and lower carbon emission, demonstrating its excellent application value.