Smoothing of PV Output Power in Grid-Tied Energy Storage System with Model Predictive Control and Battery Lifetime Consideration

This paper presents a control strategy for intermittent photovoltaic (PV) system when coupled with battery energy storage (BES) connected to the grid and the load. We use model predictive control (MPC) framework to allow the BES to adapt dynamically and provide control actions according to PV output fluctuation while satisfying several constraints including meeting the grid ramp rate requirement, and addressing the maximum battery charge/discharge power constraint and state of charge (SoC) limitation. The system is represented as a discrete time state space model and convex objectives are minimized subjected to system constraints providing optimal control input. Moving average (MA) filter is used to generate the reference power and window size is selected by considering the maximum grid ramp rate and battery lifetime. Time series forecasting model has been developed based on PV data to establish novel integrated MPC method for better energy management system incorporating battery lifetime considerations. The simulation performed in Matlab/Simulink demonstrates that the proposed control method can track the reference power while maintaining the functional requirement of the BES unit and the grid efficiently as well as preserving battery lifetime significantly.