Adaptive Estimation of Thermal Dynamics and Charge Imbalance in Battery Strings.

Battery;State Estimation;Thermal Dynamics;Charge Imbalance;Mechanical Engineering;Engineering;Mechanical Engineering

Lin, XinfanGrizzle, Jessy W. ;

University of Michigan

关键词: Battery;    State Estimation;    Thermal Dynamics;    Charge Imbalance;    Mechanical Engineering;    Engineering;    Mechanical Engineering;   

Others  :  https://deepblue.lib.umich.edu/bitstream/handle/2027.42/108789/xflin_1.pdf?sequence=1&isAllowed=y

瑞士|英语

来源: The Illinois Digital Environment for Access to Learning and Scholarship

【 摘 要 】

Effective battery management relies on accurate monitoring of battery states, including temperature, state of charge, and voltage among others. The large number of cells used in battery packs for vehicle applications require expensive monitoring hardware, which includes sensors, wiring, data acquisition and computation capacity.Due to the cost and complexity of the hardware, reduced sensing with limited and non-intrusive measurements is pursued by all manufacturers.In this dissertation, first, the monitoring of battery thermal dynamics based on only a limited number of sensors mounted on the surface of few cells is considered. Such scheme is augmented with model-based estimation techniques to capture the temperature gradient both across a single cell and among cells in the battery pack.Second, for lithium ion battery, the voltage of every single cell is currently measured to prevent overcharge and overdischarge. This dissertation develops nonlinear estimation techniques for reducing the individual cell voltage sensing requirement. Specifically, in the first part of this dissertation, a model-based estimator using surface temperature measurement and continuously identified parameters is designed for adaptive prediction of the cell core temperature.The model-based estimation is then extended for the thermal network of cells inside a pack.Based on the battery string thermal model, the number of sensors and their location required for full observability is investigated,followed by an optimal observer design under the frugal sensor allocation and cell-to-cell variability.In the second part of this dissertation, reduced voltage sensing, which relies on measuring the total voltage of multiple cells,is considered to replace the existing single-cell voltage sensing system.The feasibility of state of charge estimation under reduced voltage sensing is first investigated based on observability analysis. Nonlinear observers are then designed for SOC estimation and validated by experiments. The results are later extended to the case when both SOC and capacity imbalance exist in the battery stringdue to non-uniform cell self-discharge rates, cell degradation, and manufacturing variability.The developed estimation technique provides the potential of reducing the voltage sensing in battery packs by half.