Saturation-Aware Model Predictive Energy Management for Droop-Controlled Islanded Microgrids (2107.02719v1)

Abstract: In this paper, we propose a minimax model predictive control (MPC)-based energy management system that is robust with respect to uncertainties in renewable infeed and load. The MPC formulation includes a model of low-level droop control with saturation at the power and energy limits of the units. Robust MPC-based energy management systems tend to under-utilize the renewable energy sources to guarantee safe operation. In order to mitigate this effect, we further consider droop control of renewable energy sources. For a microgrid with droop-controlled units, we show that enhancing droop feedback with saturation enlarges the space of feasible control actions. However, the resulting controller requires to solve a mixed-integer problem with additional variables and equations representing saturation. We derive a computationally tractable formulation for this problem. Furthermore, we investigate the performance gained by using droop with saturation, renewable droop and combination of both in a case study.