Li-Ion Batteries Remaining Useful Life Maximization through Model Predictive Control Based Optimal Charging

Abstract

The existing Battery Management Systems (BMS) use empirical equivalent circuits, not reflecting the real Lithium-Ion cell electro-chemical dynamics, thus limiting the accurate estimation of the cell state and hence, not providing a good prevision of the cell’s Remaining Useful Life (RUL). This work proposes an innovative model-based method to minimize the electrochemical degradation of the Li-Ion cells. The method is applied in a Model Predictive Control (MPC), embedded in the microcontroller of the BMS. The MPC-based algorithm is maximizing the life of the cell’s digital-twin. This optimized management of the charging and discharging current, shows a reduction of the Solid Electrolyte Interface (SEI) growth by 30% to 40% and almost suppress the lithium plating compared with classical Constant-Current Constant-Voltage (CCCV) charging. As a conclusion, the new method promises longer use of the lithium batteries, especially for the energy storage in the modern and future Power Grids and Microgrids.