On battery State of Charge estimation: A new mixed algorithm

A new SoC estimation algorithm is presented. The estimation is performed mixing the Coulomb-Counting and the Model-Based estimation approaches. The algorithm was tested on a brand new lithium-ion cell which use a nanoscale technology able to grant a more powerful and safer lithium-ion battery than the batteries currently on the market. The cell, called A123-M1, was first characterized and then identified. The cell was identified with a equivalent electrical circuits model which is usually called 2^th Randle model: the used test-bench and the identification process are described. Then the SoC estimation problem is described. A briefly review of the state of art of SoC estimation is given and the drawbacks of the methods which are usually used are presented. Finally the new mixed estimation algorithm is described: the algorithm scheme and the results are shown.