Abstract
This paper presents an adaptive controller for a battery equalization system (BES) for serially connected Li-ion battery packs. The proposed equalization scheme consists of software and hardware parts to implement an adaptive neuro-fuzzy algorithm. The proposed combined software and hardware implementation of the adaptive neuro-fuzzy algorithm provides an offline learning ability to track the dynamic reactions on battery packs and a high-speed response for equalizing currents in the individual cell equalizers (ICEs). The output currents driving pulsewidth-modulated (PWM) signals are generated from the proposed hardware analog controllers. A feedback line is utilized to observe these output currents for the training process. The adaptive neuro-fuzzy algorithm is implemented in the main processor to provide adaptive parameters for the hardware. The proposed BES has an adaptability and tracking ability to deal with dynamic reactions of serially connected battery cells. The hardware controllers are implemented in a 0.13- μm CMOS technology with a supply voltage of 2.5 V. The results demonstrate that the proposed scheme has the ability to learn from previous stages and to provide a precise model of the battery cell voltages and currents. The proposed system achieved learning accuracy error of 1.8 × e -5 .
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have