| The battery is commonly used in various electronics devices ranging from laptop computers to large-scale energy storage. Due to the limited battery energy capacity, battery management is very critical to system performance. In order to maximize the battery’s performance, it is important to derive a mathematical battery model to quantitatively characterize major battery nonlinear capacity effects such as rate-dependent capacity, recovery effect, temperature effect, and capacity fading. Aslo it is important to characterize nonlinear circuit characteristics such as open-circuit voltage, internal resistant, and output voltage. Such a model will enable us to gain a thorough understanding of battery behaviors under various operation conditions. It will be useful for circuit simulation, multi-cell battery design and analysis, battery maintenance, and battery performance prediction and optimization.;In literature, various battery models have been developed, but there are no models, which consider all major nonlinearities. Also, these battery models don’t consider cell connection and cell-to-cell variation.;In this research, we will focus on addressing the fundamental problem by developing an accurate battery model based on electrical circuit analysis to consider cell connection, cell-to-cell variation, and all battery nonlinear effects. The proposed model will be validated with experimental data. (1) A comprehensive, circuit-based single-cell battery model is designed to accurately capture the performance of the battery under both constant and variable load profiles. This model considers nonlinear capacity effects and nonlinear circuit characteristics of battery. It is a comprehensive battery model. (2) A comprehensive, circuit based multi-cell battery model is developed. This model accurately estimates performance with consideration of all nonlinearities found in a battery, cell connection, and cell-to-cell variation. A novel algorithm is proposed to accurately derive the current distribution of cells in the cell string in parallel connection. The computational complexity of this algorithm increases linearly with the number of cells. This makes the multi-cell battery model ready to model a large-scale battery system. (3) A cell interaction model is studied to evaluate the effectiveness of parallel connection for cells with cell-to-cell SOC variation. Based on the proposed cell interaction model, a theoretical bound of cell-to-cell variation for a given number of cells and a required load is derived. The maximum number of battery cells for a given cell-to-cell variation distribution and a required load is also obtained. Both derived bounds enable optimization of battery performance down to the cell level. |
