Estimator for a pack of lithium-ion cell

This paper propose to estimate the state of charge for a pack of lithium-ion cells used in the electric vehicles. For this purpose, experimental tests are made with Simscape-Matlab and serious effort is put on estimating the SoC of cells series lithium-ion battery with adaptive data driven. For avoiding the variations inside of a battery pack, it is necessary a filtering strategy for getting the performances of capacity, resistances conformity in battery pack. It is the intention to use Kalman filter algorithm, because of his advantages, only the last estimation value must be memorized, to obtain a clear SoC inside of battery pack. The multi-cells pack model will be simplified by the unit model. Different methods has been used to estimate the SoC of battery packs, a common way is the ampere-hour (Ah), based on current measurement and integration. Most of the time the performances are related to the measuring precision of the current. In open-loop the calculations can lead to errors because of uncertain disturbances. There are also other methods for estimating the SoC, like artificial neural networks, fuzzy logic, the extended Kalman filtering based estimators which is the most optimal method for minimize the variations. The main disadvantage of the methods mention above for battery packs is represented by the fact that inside of each individual cell there are constructive elements which cannot be ignored and mainly the purpose is to estimate the SoC for a battery cell. In this case for an optimal battery management calculation should be made for SoC and the lowest value will represent a reference point. In order to this strategy over-discharge will be avoid.