空间锂离子电池组在线退化特征识别和寿命预测方法研究

This work focuses on the on-line state monitoring optimization, degradation identification, state-of-health (SOH) estimation and remaining useful life prediction for the lithium-ion batter pack for aerospace application. Firstly, the measured parameters of lithium-ion battery pack are not enough under on-line in-orbit conditions, and the battery cell is difficult to be monitored in the lithium-ion battery pack. Thus, the approximated equivalent monitoring of the high precision coulombic efficiency measurement will be achieved to obtain the dynamic parameters. Secondly, the degradation feature identification will be conducted with the importance measure theory. Moreover, the inconsistency between batteries as well as the balanced control will be modeled to influence the degradation. The quantifiable model for the degradation representation will be optimized with the improved information entropy measure. Thirdly, the SOH estimation method will be developed considering the dynamic operating conditions and the improved on-line feature parameters. The lebesgue state sampling based as well as the Phase type based remaining useful life prediction algorithms will be the proposed for battery pack. The hybrid model of the fused algorithms above will also be focused. Finally, the verification and evaluation will be demonstrated with the simulated data, open-source battery data, actual satellite battery data, as well as satellite in-orbit monitoring battery data. The in-orbit optimized operation and management and the on-ground experiments will be benefited with the proposed methodology and the framework to the lithium-ion battery pack in aerospace. The main contributions and novelty of this work will be on the state monitoring optimization, quantified model of the degradation identification and remaining useful life prediction model for the battery pack.

针对空间锂离子电池组在线状态评估和寿命预测问题，开展在线监测优化、退化状态识别、健康状态估计及寿命预测方法等研究。首先，针对电池组在线工况下可测参量少、模组条件下单体状态难于监测和估计等挑战，研究基于高精度库伦效率测试的近似状态监测方法，实现电池组动态参量获取；其次，基于重要性测度评价识别锂电池组退化特征参数，并构建不一致性及均衡控制等因素的影响模型，基于改进信息熵测度在参数数据域构建优化的电池组定量表征模型；再次，研究面向动态工况和优化融合特征参数的电池组健康状态估计方法，重点研究基于勒贝格状态采样和基于相型分布多状态转移概率预测的电池组寿命预测方法，并实现二者的关联建模及优化融合；最后，通过真实卫星电池组试验样本及实际卫星在轨监测数据进行验证，构建面向空间在轨优化运行管理和地面试验评估的方法体系。课题预期在电池组状态监测优化、退化特征定量建模和寿命预测建模三方面开展创新性研究。

针对空间锂离子电池组在线状态评估和寿命预测问题，在明确锂离子电池单体和电池组性能退化机理的基础上，开展航天器锂离子电池组在轨性能退化表征、电池组多维多时间尺度状态联合估计预测方法研究，并面向地面多参数、多应力测试需求和在轨计算系统功耗、体积制约限制，设计地面多参数测试系统和多尺度状态估计智能计算单元。首先，基于电池组性能退化与电池单体性能退化的耦合关系，建立航天器锂离子电池组性能退化数字孪生体，实现电池组性能退化过程的反演。然后，基于在轨可监测的电池组电压、电流参数，开展在轨可监测的性能退化表征参数提取及优化方法研究，实现航天器锂离子电池组在轨性能退化识别。最后，研究基于异构模型融合的锂离子电池组性能退化评估及预测方法，重点关注面向多尺度、多参数的联合状态估计以及多工况等效的电池组性能退化预测方法，结合研制的电池组多参数地面测试平台和高性能状态评估系统，从机理分析、模型构建、测试和计算平台实现电池组性能退化的评估和预测，为锂离子电池组的地面性能测试和在轨状态评估提供切实了用的理论体系和测试方法支撑。