Modeling and estimation of state of charge for Lithium-Ion batteries using ANFIS architecture

This paper presents the modeling and estimation of state of charge (SoC) for Li-ion batteries using ANFIS architecture. The system consists of two phases of operation. The phase 1 is the SoC modeling process. The phase 2 is the real-time estimation process. Firstly, on the phase-1 operation, a brand-new Li-ion battery is used for a completely discharge cycle which consists of 355 cycles of discharge/charge command profile to collect the data of extracted charge, internal resistance, and no-load voltage for training the ANFIS. On the phase-2 operation, the trained parameters of the system are then used to construct the estimator by using MATLAB/SIMULINK. The estimator is then used for the estimation the SoC of a Li-ion battery under test by getting the data using only one cycle of the command profile. Finally, four Li-ion batteries are tested and the result shows the brand-new batteries have higher SoC value than the used batteries.