A power smoothing scheduling strategy for PV-energy storage systems based on model predictive control

The output power of distributed photovoltaic (PV) systems is highly volatile, posing significant challenges to grid dispatch and operational reliability. Although energy storage systems (ESS) offer strong regulation capabilities, conventional energy management strategies often lack joint modeling and predictive scheduling mechanisms that incorporate both future PV trends and battery states, limiting their real-time responsiveness and control efficiency. To address these limitations, this paper proposes a rolling optimization scheduling strategy for PV-ESS systems based on Model Predictive Control (MPC). The approach constructs a state-space model integrating the dynamic evolution of the state of charge (SOC) with power balance constraints. A multi-objective optimization function is formulated to account for grid-connected power smoothness, SOC boundary enforcement, and charge/discharge ramping behavior. The control problem is solved as a constrained quadratic program over a finite prediction horizon, generating optimal control sequences in real time. Under multiple weather conditions and three storage capacity configurations, four representative control strategies are systematically compared. Simulation results show that the proposed MPC method consistently outperforms similar control schemes in key metrics, including root mean square of grid power fluctuations, SOC violation frequency, and charge/discharge smoothness. Particularly in cloudy conditions with high PV volatility, the MPC strategy demonstrates superior robustness and coordination. Moreover, the method effectively suppresses SOC boundary violations and power oscillations, contributing to reduced battery degradation and lower operation and maintenance costs. These results underscore the engineering viability and practical benefits of the proposed MPC-based control strategy for distributed PV-ESS applications.