Estimation and balancing of multi-state differences between lithium-ion cells within a battery pack

This article develops a combined estimation and control strategy for the balancing of cell-to-cell differences within lithium-ion battery packs. Heterogeneity in state of charge, state of health, and temperature reduces both pack lifespan and real-time performance capabilities. Balancing algorithms, typically designed for specific battery models, pack sizes, and heterogeneity types, mitigate these issues by removing differences through manipulation of cell currents. This article provides a generalized approach to balancing by introducing a framework that models state heterogeneity through a linear time-varying (LTV) model. The framework facilitates the development of a state estimation and balancing algorithm based on the Kalman filter (KF) and linear quadratic regulator (LQR). This balancing strategy contains three main benefits derived from the LTV model: (1) The modeling strategy explicitly expresses heterogeneity. Using this, (2) the combined estimator and controller is applicable for a large subset of heterogeneity types and cell models. (3) The form of the LTV heterogeneity model allows for reduction in computing costs of both the estimation and control algorithms, nearly decoupling runtime from pack size. The capabilities of the novel estimation-LQR strategy are evaluated using a realistic electro-thermal pack model with charge, temperature, and electrochemical state heterogeneity. Three case studies are performed to evaluate the accuracy of the heterogeneity estimator, effectiveness of the balancing strategy, and reduction in computation runtime as compared to a baseline strategy.