Dynamic Optimization of EV Charging Based on Periodically Updated Battery RUL Algorithm

Precise battery Remaining Useful Life (RUL) algorithms are required to optimally design charging strategies for Electric Vehicles (EVs). However, the uncertainty of RUL estimations before EV commissioning is usually high, due to the scarce degradation evidence at that stage. To overcome this issue, data-driven RUL algorithms have been proposed, which can be periodically upgraded with real-life degradation evidence. These upgrades allow to regularly update any control strategy aiming to reduce battery degradation. In this context, this paper proposes a novel charging strategy for EVs, which is regularly optimized based on periodical updates of a data-driven RUL algorithm. The paper also introduces a simulation environment to demonstrate the strategy. The results demonstrate that regularly updating the charging strategy allows reaching a pre-defined battery end of life objective of 10 years with an error of 25 days, while a baseline strategy without updates may reach the end of life 2 years earlier.