Adaptive Battery Management and Parameter Estimation Through Physics-Based Modeling and Experimental Verification

In this paper, adaptive battery monitoring, health, and performance analysis techniques are proposed and implemented for use in a hybrid energy storage management system. Developed through physics-based models of a lead acid and lithium ion battery cell, a chemistry detection and equivalent circuit estimation technique was accomplished using a low-frequency C/10 pulsed load. Once in operation, an adaptive coulomb counting algorithm accounts for shifts in the state of health from cycle-to-cycle using two assessment methods: estimating equivalent circuit parameters and updating the usable capacity represented by a capacitive energy model. The proposed system has the following novelties: 1) determination of the battery chemistry type and cell configuration through the use of a single standardized pulse; 2) the use of a fixed C-rate pulsed load in order to obtain a basic set of equivalent circuit parameters; 3) new voltage and temperature-based initial state of charge mechanisms for both a lead acid and a lithium ion battery; and 4) the implementation of a final control platform with chemistry detection, cell configuration, refined initial state of charge estimation, and production of a Randles equivalent circuit, regardless of the battery state of health.