Comparison and evaluation of state of charge estimation methods for a verified battery model

Battery Energy Storage Systems (BESS) emerged over the last decade due to their proven capabilities in mobile (e.g. electric vehicles) and stationary (e.g. smart grids) solutions. To manage a BESS and sustain its high performance over time, it is mandatory to monitor the battery states accurately. One of the most important states to monitor is battery State-of-Charge (SoC). SoC cannot be measured directly, therefore, its accurate estimation can improve the performance and the flexibility of the whole system. Additionally, the indications that battery SoC provides are used both for managing safe operation of the storage system and for extending the battery lifetime by preventing over-charging/discharging. Choosing an appropriate SoC estimation algorithm is typically a trade-off between algorithm complexity and accuracy. Different SoC estimation methods can be found in literature. Generally, these methods are evaluated considering a particular application (e.g. specific battery-cell technology and C-rate) that may not be suitable for other applications. This paper is focusing on simulating a verified battery Equivalent Circuit Model (ECM) and implementing different SoC estimation algorithms, using MATLAB Simulink®, to evaluate each algorithm's performance considering constant and dynamic current profiles. The results, obtained under different conditions, are compared to identify the advantages and constraints of each SoC estimation algorithm.