Switched model predictive control for energy dispatching of a photovoltaic-diesel-battery hybrid power system

Abstract

A new adaptive switched model predictive control (MPC) strategy is designed in this paper for energy dispatching of a photovoltaic-diesel-battery (PDB) hybrid power system, where the battery is unpermitted to charge and discharge simultaneously. The distinguishing feature of the proposed switchedMPC is that, new switched constraints are constructed to describe the different modes (charging and discharging) of the battery, such that the burden of using a switched multiple-input-multiple-output (MIMO) state-space model could be circumvented. Parameters of the battery are unknown constants, and are estimated online with an adaptive updating law. In the switched MPC algorithm, predictive horizon and control horizon vary according to the predefined switching schedule. Based on optimization with the switched constraints, receding horizon control is utilized to obtain the dispatching strategy for the hybrid power system. Performances of the closed-loop system with the proposed switched MPC are verified by simulation results.