【作者】 张齐；

【导师】 王大方；

【作者基本信息】 哈尔滨工业大学 ， 机械工程， 2023， 博士

【摘要】 随着新能源汽车保有量的增加,由锂离子动力电池热失效引发的新能源汽车安全事故问题也日渐突出。从事故分析来看,动力电池的热失效与其使用过程中的老化密不可分。锂离子电池在老化过程中其内部会发生多种副反应过程和失效形式,从而导致其热失效的概率增加。因此,锂离子动力电池的老化状态估计和老化诊断对保障新能源汽车的安全性具有重要意义。电池模型的搭建是新能源汽车进行电池管理的必要手段,为适应愈加严格的电池安全管理要求,更准确更能反映电池内部机理的电化学模型的建立也至关重要。在当前对于锂离子动力电池的研究中,仍存在电化学模型难以实现对电池外特性与内部电化学行为的闭环验证、电化学模型参数辨识效率低、电池老化状态估计和老化诊断方法不统一且老化诊断的实验结果与仿真结果难以对应的问题。故本文的工作从模型建立、模型参数辨识、模型的老化应用三个方面展开。为解决电化学模型难以实现对电池外特性与内部电化学行为的闭环验证的问题,提出时域与频域结合的电化学建模方法。首先,将双电层和钝化层的模型描述P2D模型,形成扩展P2D模型,实现了宽频段下时域电化学模型的时域仿真与频域仿真。随后,为提高阻抗谱进的仿真效率,提出直接建立频域电化学模型仿真阻抗谱的方法。所建立频域电化学模型包括全电池电化学阻抗模型和单电极电化学阻抗模型。最后,结合弛豫时间分析的思想,搭建了频域电化学模型的孪生等效电路模型,实现了频域模型的时域仿真。为提高电化学模型参数辨识的效率,提出基于电池特征与电化学参数对应关系的参数辨识方法。首先,将电化学模型的参数分为尺寸参数、电池状态相关参数和动力学相关参数三类。随后,重点对动力学相关参数的辨识方法进行优化。对于时域模型的动力学相关参数辨识,提出了基于电压特征的参数预辨识方法;对于频域模型的动力学相关参数辨识,提出了基于阻抗特征的参数辨识方法。为统一锂离子动力电池的老化状态估计和老化诊断,提出了基于电化学阻抗谱的老化状态估计和老化诊断方法。对于电池的老化状态估计,通过弛豫时间分析的手段,解耦并提取了电池内部各电化学过程对应的阻抗及其参数,基于高斯过程回归,实现了电池老化状态的估计。对于电池的老化诊断,首先借助CT扫描、光学照片观察、SEM观察和EDS扫描四种技术手段对电池内部在宏观层面和微观层面的形态变化和元素组成变化进行了全面诊断;随后,结合单电极电化学阻抗模型及其参数,在不拆解电池的前提下,对电池卷绕结构变形、正极活性材料损失和正极活性材料溶解三个可引起电池急剧失效的电池内部行为进行了诊断,诊断结果也能与基于实验方法的诊断结果对应。更多还原

【Abstract】 With the increase in the number of new energy vehicles,the safety accidents of new energy vehicles caused by the thermal runaway of lithium-ion power batteries have become increasingly prominent.According to the accident analysis,the thermal runaway of power batteries is closely related to their degradation in the working process.A variety of side reaction processes and failure forms will occur inside lithium-ion batteries in the aging process,resulting in an increase in the probability of their thermal runaway.Therefore,the aging state estimation and the aging diagnosis of lithium-ion power batteries are of great significance to ensure the safety of new energy vehicles.The establishment of battery models is necessary for the battery management of new energy vehicles.In order to meet the increasingly strict requirements of the battery safety management,it is also very important to establish an electrochemical model that has higher accuracy and is more capable of reflecting the internal mechanisms of batteries.Some problems are still exist in the current research on lithium-ion power battery,including the difficulty of electrochemical model to achieve closed-loop verification of battery external characteristics and internal electrochemical behavior,low efficiency of electrochemical model parameter identification,the inconsistency between battery aging state estimation and aging diagnosis methods and the difficulty of experimental results of aging diagnosis to correspond with their simulation results.Therefore,the work of this thesis is carried out from three aspects: the establishment of models,the identification of model parameters and aging applications of models.In order to solve the problem that the electrochemical model is difficult to achieve closed-loop verification of battery external characteristics and internal electrochemical behavior,an electrochemical modeling method based on the combination of time domain and frequency domain is proposed.Firstly,the model descriptions of double electric layer and passivation layer is introduced into P2 D model to form an extended P2 D model,which achieves the time-domain simulation and the frequency-domain simulation at the high frequency range.Then,in order to improve the simulation efficiency of impedance spectrum,the method of establishing the frequency-domain electrochemical model to directly simulate the impedance spectrum is proposed.The established frequency domain electrochemical models include full-cell electrochemical impedance model and single-electrode electrochemical impedance model.Then,combined with the idea of relaxation time analysis,a twin equivalent circuit model of frequency-domain electrochemical model is established to achieve the time-domain simulation using the frequency-domain model.In order to improve the efficiency of electrochemical model parameter identification,parameter identification methods based on the corresponding relationship between battery characteristics and electrochemical parameters are proposed.Firstly,parameters of electrochemical models are divided into three categories: size parameters,battery state parameters and kinetic parameters.Then,identification methods of kinetic parameters are mainly optimized.For the kinetic parameter identification of the time-domain model,a parameter pre-identification scheme based on voltage characteristics is proposed.For the kinetic parameter identification of the frequency-domain model,a parameter identification scheme based on impedance characteristics is proposed.In order to unify the aging state estimation and aging diagnosis of lithium-ion power battery,aging state estimation and aging diagnosis method based on electrochemical impedance spectroscopy are proposed.For the estimation of the battery aging state,the impedance of each electrochemical process inside the battery and their corresponding parameters are decoupled and extracted by the distribution of relaxation time method,and the estimation of the battery aging state is realized based on Gaussian process regression.For the aging diagnosis of batteries,changes of morphologies at the macro and micro levels and element composition of batteries at different aging stages are first diagnosed completely by methods of CT scanning,optical photograph observation,SEM observation and EDS scanning.Then,combined with the single-electrode electrochemical impedance model and its parameters,three battery internal behaviors which can cause rapid battery failure,including the deformation of battery winding structure,the loss of positive active material and the dissolution of positive active material,are diagnosed without disassembling the battery.The diagnosis results can also correspond to diagnosis results based on the experimental method.更多还原