锂离子电池寿命预测技术研究成都电子科技大学工学硕士学位论文重点.doc

摘 要 随着电动汽车产业化的兴起，锂离子电池在电动汽车领域得到广泛应用，锂离子电池组的有效管理亦成为保障电动汽车性能稳定，安全可靠的关键。荷电状态（, ）估计电池管理技术的核心受各种非线性因素的影响，且实时性要求高，成为电池管理的难点。剩余寿命（, ）预测作为电池管理的另一重要指标，为保障电池的及时更换提供必要依据。但预测是近年来才开始研究的课题，方法有限且缺乏深入分析比较。基于上述考虑，估计和预测的方法研究，及监测系统的设计，主要完成了以下工作： 1. 锂离子电池荷电状态估计。由于电池内外部非线性因素的影响，同时考虑到磷酸铁锂电池充放电电压相对变化较小的特点，选择具有自适应学习能力的神经网络作为训练算法。稳态放电实验，探讨了电池容量与充放电倍率等因素之间的关系。基于对电池各项参数与内在联系的分析，选择电流和电压作为网络输入，完成神经网络的训练及估计。实验数据表明，提出的算法能达到较好的估计精度。 预测，根据预测结果比较各自优缺点。进而提出粒子群（）优化的自回归（）模型，该模型以均方根误差（）作为模型的定阶准则，基于算法对模型的定阶方法进行优化。进而结合新陈代谢数据处理技术实现阶数的自适应改变，有效避免了传统模型定阶流程复杂的缺陷。最终数据验证得到了准确的预测趋势并有效提高了预测精度。 3. 锂离子电池监测系统的设计开发。基于和的研究结果，设计锂离子电池监测系统，以满足锂离子电池管理的需求。该系统有效实现了电池信号采集，数据传输上位机显示估计及预测的动态链接库程序对电池参数进行分析处理，完成了电池的估计和预测。估计，神经网络，预测，模型，监测系统ABSTRACT With the rapid rise of the industrialized electric vehicle (EV), the energy issue has been drawing wide attention from both academia and industry. The efficient management of the lithium-ion battery pack becomes a key to ensure reliability and security of EVs. As the core of battery management, the state of charge (SOC) estimation is one of the most difficult problems due to the nonlinear effects and real-time requirement. On the other hand, the remaining useful life (RUL) is another critical performance indicator of battery management, which guarantees the timely replacement of batteries. However, the research on RUL prediction has not been started until recent years, therefore there is little existing work can be found. Based on the above considerations, this thesis conducts comprehensive study on the SOC and RUL problems and battery monitoring system design, major contributions are summarized as follows: 1. Lithium-ion batteries SOC estimation. Considering the external and internal nonlinear factors as well as the observation that the voltage remains mostly constant as SOC changes during the charge/discharge process of LiFePO4 battery, the back propagation (BP) neural network model is chosen as the training algorithm due to its adaptive learning ability. Through the constant curre