Electrochemical Modeling and Parameterization Towards Co - 2022 - Control Engine

Control Engineering Practice 124 (2022) 105176
Contents lists available at ScienceDirect
Control Engineering Practice

journal homepage: www.elsevier.com/locate/conengprac
Electrochemical modeling and parameterization towards control-oriented

management of lithium-ion batteries
Kailong Liu a , Yizhao Gao b ,∗, Chong Zhu b , Kang Li c ,∗, Minrui Fei d , Chen Peng d , Xi Zhang b ,
Qing-Long Han e
a
WMG, The University of Warwick, Coventry, United Kingdom
b
School of Mechanical Engineering, Shanghai Jiao Tong University, Shanghai, China
c School of Electronic and Electric Engineering, University of Leeds, United Kingdom
d School of Mechatronics Engineering and Automation, Shanhai University, China
e School of Science, Computing and Engineering Technologies, Swinburne University of Technology, Australia
ARTICLE INFO ABSTRACT

Keywords: Battery management systems based on electrochemical models could achieve more accurate state estimations
Lithium-ion battery and efficient battery controls with access to cell unmeasurable physical variables. As battery electrochemical
Control-oriented management models are governed by first-principle partial differential equation sets, model complexity and multiple
Energy storage
parameter determination are bottlenecks for their wider applications. This paper gives a systematical review
Electrochemical model
of recent advancements in electrochemical model development and parameterization. Specifically, classic
Model reduction
Parameter identification
pseudo-two-dimensional model and related model order reduction methodologies are first summarized and
analyzed. Given that the homogenization hypothesis of the pseudo-two-dimensional model could lead to
significant model mismatch under some operational conditions, enhanced models considering cell internal
inhomogeneity with multi-particles, multi-scales, aging and thermal dynamics are examined. To facilitate
model portability, parameter identification techniques of these models are classified, and solutions for
optimizing the parameterization procedure are explored. Finally, current research gaps in the literature and
remaining challenges are discussed and highlighted with some suggestions. This review will therefore inform
the engineers of battery management and control engineering, whilst boosting the research, design and
operation of control-oriented electrochemical models for smarter battery management at different readiness
levels.
1. Introduction divided into two categories: equivalent circuit models (ECMs) and
electrochemical models.
Clean, renewable energy sources are needed to help create a sus- For the ECMs (Jin et al., 2018; Lekshmi & PS, 2019), one or
tainable society. Due to the superiorities in terms of energy density, more parallel resistor–capacitor sub-circuits are applied to perform
efficiency, low discharge rate, and environmental friendliness (Wang phenomenological emulations of cell external responses, where battery
et al., 2020; Wei et al., 2021), lithium-ion (Li-ion) batteries have current, voltage and aging behavior could be also well matched (Li,
become one of mainstream energy storage components in numerous Jiao, & Li, 2021; Ouyang, Wang, Liu, Xu, & Li, 2019). Due to the sim-
sustainable applications such as electric vehicles (EVs) (Klintberg, Zou, ple structure and low computational burden, ECMs have been widely
Fridholm, & Wik, 2019; Shafikhani et al., 2021), renewable energy adopted in BMSs for control-oriented management (He, Ye, He, Song,
storage, and smart grid (Ferro, Laureri, Minciardi, & Robba, 2019). & Shi, 2020; Li, Li, Xiao, & Wong, 2019; Liu, Li, & Zhang, 2017).
Battery management systems (BMSs) generally implement model-based In contrast, electrochemical models allow the enhanced monitoring
algorithms to protect battery from abuse, prolong battery service life, and prediction of battery individual mechanisms (Liu, Li, Peng, &
and maintain battery operated in an efficient status (Fang, Wang, Zhang, 2019). Given the capability of capturing cell-intrinsic phys-
Sahinoglu, Wada, & Hara, 2014; Li, Li, Liu, Wang, & Zhang, 2021). As a ical states, electrochemical model-based BMSs enable health-aware
key functional part for achieving control-oriented battery management, control of Li-ion batteries to improve their operational safety, reliabil-
Li-ion battery modeling techniques with a specific form can be mainly ity, and efficiency (Gao, Liu, Zhu, Zhang, & Zhang, 2021; Hu et al.,
∗ Corresponding authors.
E-mail addresses: kailong.liu@warwick.ac.uk (K. Liu), gaoyizhao@sjtu.edu.cn (Y. Gao), chong.zhu@sjtu.edu.cn (C. Zhu), k.li1@leeds.ac.uk (K. Li),
mrfei@staff.shu.edu.cn (M. Fei), c.peng@i.shu.edu.cn (C. Peng), braver1980@sjtu.edu.cn (X. Zhang), qhan@swin.edu.au (Q.-L. Han).
https://doi.org/10.1016/j.conengprac.2022.105176
Received 18 November 2021; Received in revised form 4 March 2022; Accepted 3 April 2022
Available online 23 April 2022
0967-0661/© 2022 Elsevier Ltd. All rights reserved.
K. Liu, Y. Gao, C. Zhu et al. Control Engineering Practice 124 (2022) 105176
model (Lee, Chemistruck, & Plett, 2012a). Some researchers also use
Abbreviations Padé approximation to match the frequency responses of PDEs (Yuan,
EV Electric vehicle Jiang, Yin, Wu, & Zhang, 2017b). Furthermore, the polynomial profiles
P2D Pseudo-two-dimensional (PP) are fast approximations of electrolyte concentration and potential
along the 𝑥 direction (Sharma et al., 2019), while the solid-phase
BMS Battery management system
diffusion along the r direction (Subramanian, Tapriyal, & White, 2004).
PDE Partial differential equation
The PP model poses simple structure and converts PDEs into the Differ-
SPM Single particle model
ential Algebraic Equations (DAEs). The spatial discretization methods
SPMe Single particle model with electrolyte dy- including the finite difference (Xiong, Li, Li, Yu, & Mu, 2018) and finite
namics volume method (Ashwin, McGordon, & Jennings, 2017), are effective
CCCV Constant-current constant-voltage tools for numerical solutions of PDEs (Wang, Ma, & Zhang, 2017). The
DAE Differential algebraic equation discretization methods could achieve high accuracy because they divide
MPM Many particle model the feasible region into multiple cells (Dickinson, Ekström, & Fontes,
DFN Doyle–Fuller–Newman 2014). Some researchers combine the ECM and electrochemical models
FVM Finite volume method to make best of the two approaches, e.g. high computational speed with
LFP LiFePO4 physical meaning (Zhang, Lu, Yuan, Yang, & Zhou, 2017). A recent
SOC State of charge trend of research is to use the asymptotic techniques for achieving
the order reduction of the electrochemical model (Marquis, Timms,
ECM Equivalent circuit model
Sulzer, Please, & Chapman, 2020). The asymptotic reduction approach
IMA Integral method approximation
is based on exploiting the vastly different time scales of various physical
DRA Discrete-time realization algorithm
processes, leading to a more organic simplification (Tranter et al.,
PSD Particle size distribution 2020).
ROM Reduced order model On the other hand, P2D model is formulated under the homo-
SEI Solid electrolyte interphase geneity assumptions. For example, the P2D model assumes that the
EIS Electrochemical impedance spectroscopy active material particles are spherical and distribute uniformly with
GITT Galvanostatic intermittent titration tech- identical radius (Forouzan, Mazzeo, & Wheeler, 2018). However, prac-
nique tical electrodes have non-spherical particle shapes that exhibit large
dispersity (Sacci et al., 2015). This indicates that the P2D model
would be limited in the cases where cells have dispersed particle
shapes. Therefore, the accuracy of P2D model would be degraded
2019; Zou, Hu, Wei, Wik, & Egardt, 2017). However, the electro- in predicting cell voltage under extreme circumstances, such as low
chemical models generally consist of a series of partial differential SOC range, resting period (Kirk, Please, & Jon Chapman, 2021), and
equations (PDEs), which would inevitably increase the computational high temperatures (Kraft et al., 2015). In addition, volume average
complexity in battery practical applications (Feng et al., 2020). As approach does not consider the actual topology and electrode morphol-
the most widely-adopted electrochemical model, battery pseudo-two- ogy (Pereira et al., 2020). Porosity and active volume fraction are not
dimensional (P2D) model is derived from the porous electrode theory, constant throughout battery lifespan due to the volume changes, cracks,
concentrated solution theory, and kinetics equations (Eapen, Suresh, lithium plating and stress effects occurring in the electrodes during cell
Patil, & Rengaswamy, 2021). The P2D model is established using the charge and discharge (Avdeev et al., 2017). The lithiation/delithiation
volume averaging technique and treats the electrode as a homogeneous mechanisms describing the intra-particle phase transformation of P2D
medium (Li, Cui, Wang, & Zhang, 2021), where the full order P2D model fail to analyze the coexistence of Li-poor phase and Li-rich
model is computationally expensive and difficult to be implemented phase within LFP crystals (Koltypin, Aurbach, Nazar, & Ellis, 2007).
on the hardware. Therefore, it is necessary to simplify the P2D model The coexistence of two phases is the major cause of the flat open-
before the control-oriented applications can be realized. circuit-potential within LFP electrodes (Anseán et al., 2016). Therefore,
To reduce computational complexity of electrochemical models, it can be inferred that the limitations of the P2D model arise from
various simplified versions of the P2D model have been investigated. the model simplification process. Many researchers have extended P2D
To be specific, by simplifying each electrode as a single solid spher- model to consider the inhomogeneity of particle and electrode (Bistri
ical particle, single particle model (SPM) becomes the most popular & Di Leo, 2021). A simplified electrochemical multi-particle model for
simplified P2D model (Tian, Wang, & Chen, 2021a). However, due LFP cathodes in Li-ion batteries was developed (Mastali Majdabadi,
to electrolyte concentration and potential are considered to be con- Farhad, Farkhondeh, Fraser, & Fowler, 2015). Therein, the active ma-
stant for the cell, the performance of SPM can only match as well terial particles of non-uniform properties (e.g., size, contact resistance,
as the rigorous P2D model in the cases where the current rate is material chemistry etc.) were incorporated. The multi-particle model
less than 1C. To overcome this limitation, both electrolyte are tem- could accurately simulate the cell charge/discharge current rates up to
perature dynamics are integrated with SPM (Han, Ouyang, Lu, & Li, 5C. The realistic 3D anode structures using a stochastic microstructure
2015), where the acceptable accuracy and low-cost computational model was generated, which is parameterized using tomography data.
effort enable these SPMs become competent for developing battery The results show that the deviations between uniform and realistic
state estimation strategies (Li, Fan, Ringbeck, J?st, Han, Ouyang, & electrodes are pronounced at high current rates (Hein et al., 2016). Mul-
Sauer, 2020) and optimal control approaches (Pramanik & Anwar, tiple cases of heterogeneities, such as non-uniform ionic resistance and
2016). Another promising approach is to linearize the coupled PDEs active material loading, are studied at different charge and discharge
of rigorous P2D porous electrode and develop analytic Laplace-domain current rates (Fang, Ge, Wang, Li, & Zhang, 2020). The results indicate
transfer functions from the linearized model (Yuan, Jiang, Yin, Wu, & that higher current rates increase the non-uniformities of temperature,
Zhang, 2017a). For example, the transfer functions of reaction flux, current density, positive and negative electrode SOC, especially in the
solid-electrolyte potential difference, overpotential, and solid particle case of charging (Forouzan et al., 2018).
surface concentration versus input current were derived (Smith, Rahn, Accurate model parameterization is also crucial for efficient and
& Wang, 2007). Then a low-order model was generated through a reliable electrochemical model-based prediction, state estimation, and
residue-grouping approach. A novel discrete-time realization algorithm control (Xia et al., 2016). Generally, there are two main solutions for
(DRA) was used to produce a reduced-order discrete-time state–space identifying parameters of electrochemical models: direct measurements
2
Table 1
An overview of the published literature related to battery electrochemical models.
Topic Reference Content
Jokar, Rajabloo, Désilets, and Lacroix (2016) Simplified P2D models.
P2D model simplification
Li, Ralahamilage et al. (2021) P2D model order reduction techniques.
Li and Monroe (2020) Multiscale Lithium-ion battery modeling from materials to cells.
Enhanced electrochemical model
Franco (2013) Multiscale modeling of rechargeable batteries.
Waldmann et al. (2018) Post-mortem analysis of lithium-ion batteries.
Miguel et al. (2021b) Computational parameter estimation methods.
Parameter identification method
Dos Reis, Strange, Yadav, and Li (2021) Battery testing and experimental datasets.
Wang et al. (2021) Summarization of methods to measure or infer parameters.
and invasive parameter estimations. For the direct measurement, after • Current research challenges are discussed in Section 5. Feasible
dismantling battery cell, its internal structures, compositions, and prop- solutions to address these challenges are suggested as future
erties would be analyzed with specific instruments (Sturm et al., 2019). research direction towards the improvement of battery control-
This type of solution is robust as the physical parameters of battery oriented electrochemical modeling technologies.
can be directly measured. A detailed review for the post-mortem tech-
niques and subsequent electrochemical analysis of cell performance and
2. P2D model and simplifications
physical–chemical properties can be found in Stiaszny, Ziegler, Krau?,
Schmidt, and Ivers-Tiffée (2014) and Waldmann, Wilka, Kasper, Fleis-
chhammer, and Wohlfahrt-Mehrens (2014), respectively. The invasive The pseudo-2-dimension (P2D) model pioneered by Doyle et al.
parameter estimation refers to the methods that estimate model pa- is one of the most popular electrochemical models for Li-ion bat-
rameters by fitting electrochemical model output to the measured cell teries (Chen, Liu, Su, & Zhuang, 2021; Doyle, Fuller, & Newman,
quantities (e.g., terminal voltage, temperature) (Zhang et al., 2015). 1993). This type of model consists of numerous PDEs and is compu-
Due to most of the parameters are weakly correlated with the bat- tationally complex for control-oriented applications (Reddy, Scharrer,
tery terminal voltage, complete set of parameters is difficult to be Pichler, Watzenig, & Dulikravich, 2019), where various model reduc-
identified simultaneously (Namor, Torregrossa, Cherkaoui, & Paolone, tion techniques are required to simplify P2D model. In this section,
2017). Therefore, a stepwise solution is generally adopted to identify the development of rigorous P2D model is first introduced, followed
these parameters (Zhang et al., 2014). Both these two parameteriza- by the summary and comparison of various P2D model order reduction
tion methods have their own merits and drawbacks. Cell postmortem methods and their control-oriented applications.
analysis could provide visualized insights into material properties.
However, this method is time-consuming and the related equipment 2.1. Definitions and derivations of the basic P2D model
used in the analysis is often expensive (Waldmann, Hogg, & Wohlfahrt-
Mehrens, 2018). Parameter estimation method could save identification
P2D model is a classical electrochemical model and has been widely
time but the robustness as well as validity of the estimated parame-
investigated. For the macroscopic model of Li-ion batteries, concen-
ters need to be further verified (López C, Wozny, Flores-Tlacuahuac,
trated solution theory is generally utilized to describe the transport
Vasquez-Medrano, & Zavala, 2016).
properties in the solution phase, while porous electrode theory is
To date, a few review articles regarding the battery P2D model
simplification, enhanced electrochemical model, and parameter iden- adopted to analyze the composite electrodes (Doyle, Newman, Gozdz,
tification method are presented, as summarized in Table 1. It should Schmutz, & Tarascon, 1996). Here the concentrated solution theory
be known that all these reviews mainly focus on one aspect alone. As assumes that electrolyte consists of a binary salt system (Li+ and 𝑋 − )
battery electrochemical model towards control-oriented management in a single solvent. For the porous electrode theory, a solution phase,
generally requires to conduct all of the following steps, namely model a solid active-material phase, and any conductive filler or binding
simplification, improvement, and parameter identification, where these additives are considered as superimposed continua, hence all points of
aspects are heavily correlated. In this context, a review article covering the electrode are perfectly connected. The active material is supposed
all these aspects is urgently required especially for promoting battery to be comprised of multiple spherical particles with a diameter 𝑅𝑠,𝑘
electrochemical model-based control applications. Here, a systematical (𝑘 ∈ 𝑛 = 𝑛𝑒𝑔𝑎𝑡𝑖𝑣𝑒𝑒𝑙𝑒𝑐𝑡𝑟𝑜𝑑𝑒, 𝑝 = 𝑝𝑜𝑠𝑖𝑡𝑖𝑣𝑒𝑒𝑙𝑒𝑐𝑡𝑟𝑜𝑑𝑒). Based upon above
review of electrochemical modeling and parameterization methods to- assumptions, the P2D model could be established, as shown in Fig. 1. To
wards control-oriented management of Li-ion batteries is given, while be specific, the spherical electrode particles are defined in the domain
their challenges and prospects are also discussed. The review is targeted 0 ≤ 𝑟 ≤ 𝑅𝑠,𝑘 . The electrolyte traverses across the two porous electrodes
to inform control-oriented technology choices and academic research and separator (0 ≤ 𝑥 ≤ 𝐿), where the anode, separator and cathode are
agendas alike, thus advancing battery electrochemical model develop- defined in 0 ≤ 𝑥 ≤ 𝐿𝑛 , 𝐿𝑛 ≤ 𝑥 ≤ 𝐿𝑛 +𝐿𝑚 , and 𝐿𝑛 +𝐿𝑚 ≤ 𝑥 ≤ 𝐿𝑛 +𝐿𝑚 +𝐿𝑝 ,
ment at different technology readiness levels. The following topics are respectively. Here the subscripts 𝑛, 𝑚, and 𝑝 denote the variables in
covered as: negative electrode, separator, and positive electrode, respectively. The
Li+ diffusion in the solid particles with boundary conditions follows
• In Section 2, typical battery P2D model with different simplifi-
Fick’s second law as (Khalik, Donkers, Sturm, & Bergveld, 2021):
cation solutions are comprehensively reviewed. Their merits and
drawbacks are also thoroughly compared and discussed. 𝜕𝑐𝑠 𝐷 𝜕 𝜕𝑐
= 𝑠 (𝑟2 𝑠 ) (1)
• For the enhanced electrochemical modeling technology, battery 𝜕𝑡 𝑟2 𝜕𝑟 𝜕𝑟
multi-particle model, electrochemical model with multiple scales, 𝜕𝑐𝑠 || 𝜕𝑐 | 𝑗
= 0, 𝐷𝑠 𝑠 || =− (2)
𝜕𝑟 ||𝑟=0
electrochemical model with aging and thermal effects are re-
𝜕𝑟 |𝑟=𝑅𝑠 𝑎𝑠 𝐹
viewed in Section 3, and their advantages and disadvantages are
discussed. where 𝑐𝑠 is the solid concentration, 𝑗 denotes the lithium flux across the
• For the parameterization of battery electrochemical model, tech- boundary of solid phase, 𝐷𝑠 is the solid diffusion coefficient, 𝑎𝑠 is the
nologies including physical parameter measurements and non- specific interfacial area, 𝑅𝑠 is the particle radius, and 𝐹 is the Faraday
invasive parameter estimation are comprehensively surveyed in constant. The lithium concentration at particle surface is denoted as
Section 4. 𝑐𝑠,𝑒 .
3
where 𝑖0 is the exchange current density, 𝑖0 = 𝐹 𝑘(𝑐𝑒 )𝛼𝑎 (𝑐𝑠,𝑚𝑎𝑥 − 𝑐𝑠,𝑒 )𝛼𝑎
(𝑐𝑠,𝑒 )𝛼𝑐 , 𝑘 is a kinetic rate constant. 𝜂 is overpotential, which is an extra
force required to overcome the surface reaction as :
𝜂𝑘 = 𝜙𝑠,𝑘 − 𝜙𝑒,𝑘 − 𝑈𝑜𝑐𝑝,𝑘 , 𝑘 ∈ {𝑛, 𝑝} (9)
where 𝑈𝑜𝑐𝑝 is the electrode open circuit potential and can be de-
scribed by lithium surface concentration 𝑐𝑠,𝑒 . Cell voltage could be
described by the potential difference between the positive and negative
current-collectors, and voltage drop due to film resistance as (Farag,
Fleckenstein, & Habibi, 2017):
𝑉 (𝑡) = 𝜙𝑠 (𝐿, 𝑡) − 𝜙𝑠 (0, 𝑡) − 𝑅𝑓 𝐼 (10)
where 𝑅𝑓 is the film resistance.

It should be known that the coupled PDEs (1)–(7) can be utilized to
investigate battery electrochemical mechanisms and optimize battery
designs without numerous experimental data. These PDE models are
infinite-orders since there are infinite number of 𝑥− and 𝑟− dimensional
variables to be calculated at each point in time 𝑡. A complete analytical
solution of the governing PDEs is generally difficult to obtain. As the
P2D models are too complicated to be implemented in real-time appli-
cations particular for battery control-oriented management, a great deal
of efforts have been done to reduce the finite order of PDE models into
small order while retaining fidelity, further resulting in the PDEs can be
Fig. 1. Schematic diagram of Li-ion battery P2D model.
transformed into ordinary differential equations (ODEs) with relatively
lower computational complexity.
The aim of developing simplified P2D models is to approximate
The solid electrode potential 𝜙𝑠 is described by the Ohm’s law the rigorous P2D models at a high accuracy but with a relatively
as (Safari & Delacourt, 2011): low computational burden (Subramaniam, Kolluri, Santhanagopalan,
𝜕 2 𝜙𝑠 & Subramanian, 2020). Battery management systems with the sim-
𝜎 𝑒𝑓 𝑓 − 𝑎𝑠 𝑗 = 0 (3) plified P2D models have significant advantages over ECMs as this
𝜕𝑥2
type of control-oriented models could provide richer electrochemical
where 𝜎 𝑒𝑓 𝑓 is the effective electrode conductivity, 𝑎𝑠 is the electrode
information, further benefitting cell optimal control, online estimation,
specific surface area. The concentration of electrolyte is determined
optimization, and fault diagnostics (Gao et al., 2020).
by Safari and Delacourt (2011):
( )
𝜕𝑐 𝜕 2 𝑐𝑒 𝑎𝑠 1 − 𝑡0+ 2.1.1. Single particle model based
𝜀𝑒 𝑒 = 𝐷𝑒𝑒𝑓 𝑓 + 𝑗 (4)
𝜕𝑡 𝜕𝑥2 𝐹 Single particle model (SPM) is a classical simplified version of the
with the following boundary conditions: P2D model based on three critical hypotheses: First, the negative and
positive electrodes are represented with only two spherical particles.
⎡ 𝜕𝑐𝑒 || 𝜕𝑐𝑒 || ⎤ Second, the electrolyte concentration and potentials are neglected.
⎢ 𝜕𝑥 ||𝑥=0 = 0, 𝜕𝑥 ||𝑥=𝐿 = 0 ⎥
⎢ − +
⎥ Third, the lithium flux is proportional to the input current. The SPM
⎢𝑐𝑒 (𝐿𝑛 ) = 𝑐𝑒 (𝐿𝑛 ) ⎥ is computational efficient and could be adopted in control-oriented
⎢ − + ⎥
𝑐 ((𝐿
⎢ 𝑒 𝑛 + 𝐿 𝑚 ) ) = 𝑐 𝑒 ((𝐿 𝑛 + 𝐿 𝑚 ) ) ⎥ (5) applications (Ren et al., 2021), as shown in Fig. 2. However, SPM
⎢ 𝜕𝑐 (𝐿− ) 𝜕𝑐 (𝐿+ ) ⎥ would fail under the high current rates (> 1𝐶) (Khalik, Donkers, &
⎢𝐷𝑒𝑓 𝑓 𝑒 𝑛 = 𝐷𝑒𝑓 𝑓 𝑒 𝑛 ⎥
⎢ 𝑒,𝑛 𝜕𝑥 𝑒,𝑚 𝜕𝑥 ⎥ Bergveld, 2021). To further handle this limitation, some efforts have
⎢ 𝜕𝑐 ((𝐿 + 𝐿 ) −
) 𝜕𝑐 ((𝐿 + 𝐿 )+ ⎥
) been made to improve SPM by considering the electrolyte (Moura,
⎢𝐷 𝑒𝑓 𝑓 𝑒 𝑛 𝑚
=𝐷 𝑒𝑓 𝑓 𝑒 𝑛 𝑚 ⎥
⎣ 𝑒,𝑚 𝜕𝑥 𝑒,𝑝 𝜕𝑥 ⎦ Argomedo, Klein, Mirtabatabaei, & Krstic, 2017) and thermal dynam-
ics (Guo, Sikha, & White, 2011). As a result, these enhanced SPMs
where 𝑐𝑒 denotes the electrolyte concentration, 𝐷𝑒𝑒𝑓 𝑓 is the effective are able to provide sufficiently accurate results under high current
electrolyte diffusion coefficient as 𝐷𝑒𝑒𝑓 𝑓 = 𝐷𝑒 𝜀𝑏𝑟𝑢𝑔
𝑒 . 𝜀𝑒 represents the rates and achieve a substantially lower computational burden than P2D
volume fraction of electrolyte, and 𝑡0+ is the transference number of Li+ . model (Tian, Wang, & Chen, 2021b).
Bruggeman coefficient is denoted as 𝑏𝑟𝑢𝑔. The charge conservation in
the electrolyte can be described by Saw, Ye, and Tay (2013):
2.1.2. Transfer-functions type
𝑒𝑓 𝑓 𝜕 2 𝜙𝑒 𝑒𝑓 𝑓 𝜕 2 ln 𝑐𝑒 Besides, establishing transfer functions of all electrochemical vari-
𝜅 + 𝜅𝐷 + 𝑎𝑠 𝐹 𝑗 = 0 (6)
𝜕𝑥2 𝜕𝑥2 ables and converting these transfer functions into low-order high-
where 𝜙𝑒 is the electrolyte phase potential, 𝜅 𝑒𝑓 𝑓 is the effective ionic fidelity state–space approximate models is also becoming a promising
conductivity as 𝜅 𝑒𝑓 𝑓 = 𝜅𝜀𝑒𝑓 𝑓 𝑒𝑓 𝑓 𝑒𝑓 𝑓 solution to simplify the P2D model. This transfer-function-based so-
𝑒 , and 𝜅𝐷 is a short term for 𝜅𝐷 =
+
2𝑅𝑇 𝜅(𝑡0 −1)∕𝐹 . Here the boundary conditions at two current collectors lution can solve any subset of the internal electrochemical variables
are: desired at any selection of internal cell locations. Even though it is
mathematically complex to derive transfer functions, the final compu-
𝜕𝜙𝑒 || 𝜕𝜙𝑒 ||
= 0, =0 (7) tational complexity could become as significantly low as the battery
𝜕𝑥 ||𝑥=0 𝜕𝑥 ||𝑥=𝐿
ECM (Rodríguez & Plett, 2017).
The lithium insertion process in the porous electrode is described The transfer functions for solid surface concentration, lithium flux
by Butler–Volmer equation as: and phase potential difference were first derived (Smith & Wang,
( ( ) ( )) 2006).With these transfer functions, model reduction techniques, in-
𝛼𝑎 𝐹 𝛼 𝐹
𝑗 = 𝑖0 exp 𝜂 − exp − 𝑐 𝜂 (8) cluding the truncated Taylor-series expansion, residue grouping, and
𝑅𝑇 𝑅𝑇
4
The Padé approximation of the transcendental transfer functions for

the solid-state diffusion were applied (Moura, Chaturvedi, & Krsti?,
2014), where the first-order, second-order, and third-order Padé ap-
proximations were presented and compared in the frequency domain. It
turned out that the Padé approximates could also well capture the low-
frequency dynamics, where the model would become more accurate at
high-frequency as the Padé order increases.
√
𝐶̃𝑠𝑠 (𝑠) 𝑅𝑠 tanh(𝑅𝑠 𝑠 ∕𝐷𝑠 )
= √ √ (13)
𝐽 (𝑠) 𝐹 𝐷𝑠 tanh(𝑅 𝑠 ∕ − 𝑅 𝑠 ∕ )
𝑠 𝐷𝑠 𝑠 𝐷𝑠
where 𝑅𝑠 is the particle radius, 𝐹 is the Faraday constant, 𝐷𝑠 is the

solid-phase diffusion coefficient.
A model-order reduction procedure based on the Padé approxima-
tion was presented (Marcicki, Canova, Conlisk, & Rizzoni, 2013). The
frequency distribution of current profiles concerning two regulatory
driving cycles including the US06 highway cycle and Federal Urban
Fig. 2. Schematic diagram of single-particle model.
Dynamometer Schedule (FUDS) was investigated. As 90% of these
cycles are within the frequency domain less than 2.5 Hz, a first-order
Padé approximation is therefore chosen to model the liquid diffu-
sion dynamics, and a third-order truncation is selected for the solid-
nonlinear optimization were applied to generate a low-order electro-
diffusion dynamics. Based upon the Lee’s derivation of transcendental
chemical model. Here the terminal voltage from a 12th order state
transfer functions on lithium flux 𝑗, solid surface concentration 𝑐𝑠,𝑒 ,
variable model could reach less than 1% error for pulse and constant
electrolyte concentration 𝑐𝑒 , and electrolyte potential, a reduced-order
current profiles at the current rates up to 50C. The state equation was
model was presented where Padé approximates were used to handle the
constructed with constant negative real eigenvalues distributed in the
complicated transcendental transfer functions. The obtained reduced-
frequency range from 0 to 10 Hz (15). This type of method was further
order model is amenable with rapid computation for control-oriented
extended to derive transfer functions for electrolyte concentration,
applications (Tran et al., 2018).
solid and electrolyte potentials (Lee, Chemistruck, & Plett, 2012b). The
Integral method approximation (IMA) is another effective solution
reaction flux transfer function was given by:
to solve the transfer functions, where a distribution of the electro-
𝐽 (𝑧, 𝑠) 𝜎 𝑒𝑓 𝑓 cosh(𝜐𝑘 (𝑠)𝑧) + 𝜅𝑘𝑒𝑓 𝑓 cosh(𝜐𝑘 (𝑠)(𝑧 − 1)) chemical variables across battery cell is assumed and the governing
= ±𝜐𝑘 (𝑠) 𝑘 (11) equations are integrated. The IMA was used to solve Li+ conservation
𝐼(𝑠) 𝑎𝑘𝑠 𝐹 𝐿𝑘 𝐴(𝜅𝑘𝑒𝑓 𝑓 + 𝜎𝑘𝑒𝑓 𝑓 ) sinh(𝜐𝑘 (𝑠))
across the anode, cathode, and separator of cell (Tanim, Rahn, & Wang,
where 𝑘 ∈ [𝑛 = 𝑛𝑒𝑔𝑎𝑡𝑖𝑣𝑒 𝑒𝑙𝑒𝑐𝑡𝑟𝑜𝑑𝑒, 𝑝 = 𝑝𝑜𝑠𝑖𝑡𝑖𝑣𝑒 𝑒𝑙𝑒𝑐𝑡𝑟𝑜𝑑𝑒]. 𝑧 = 𝑥∕𝐿𝑘 . 𝜈 is 2015). A parabolic distribution of the electrolyte concentration was
the unitless impedance ratio. Afterwards, the discrete-time realization assumed and substituted into Li+ conservation equations, where the
algorithm (DRA) was used to convert these transfer functions into an boundary conditions are integrated and simplified to obtain the third
optimal low-order discrete-time state-space approximate model. order transfer functions for electrolyte concentrations.
It should be noted that the following critical assumptions are made
for both Smith and Plett’s work when solving the transfer functions 2.1.3. Polynomial expression
from nonlinear PDEs: First, the nonlinear equations could be linearized The polynomial profiles are simple approximations for solid-state
by using Taylor series. This assumption is reasonable as the transfer concentration function (Liu, Tang, & Gao, 2020) and electrolyte concen-
functions only exist for linear systems. Second, the electrolyte potential tration distributions (Deng, Yang, Deng, Cai, & Li, 2018). The efficient
𝜙𝑒 is predominantly a function of lithium flux 𝑗, further indicating that microscale diffusion inside porous electrodes was developed, assuming
the effect of electrolyte concentration on the electrolyte potential was that the solid-state concentration inside spherical particle could be
ignored. However, the second assumption could fail when the battery expressed as a polynomial in the spatial direction. The two-parameter
cell is operated under constant-current conditions, where a large con- and three-parameter parabolic models presented the following forms
centration gradient occurs in the electrolyte. Through eliminating the as:
requirement of using second assumption, a more complicated transfer
𝐶(𝑧, 𝜏) = 𝑎(𝜏) + 𝑏(𝜏)𝑧2 ;
functions for PDEs (Rodríguez, Plett, & Trimboli, 2018) was observed. (14)
Similarly, DRA method was employed to obtain the final reduced-order 𝐶(𝑧, 𝜏) = 𝑎(𝜏) + 𝑏(𝜏)𝑧2 + 𝑐(𝜏)𝑧4 ;
model whose computational complexity is similar to battery ECM. The approximate models were tested for arbitrary functions of pore
The Padé approximation method is an effective tool for converting wall flux and could save computational effort by over 80% without
transcendental transfer functions (the exact solutions of diffusion equa- compromising the accuracy (Subramanian et al., 2004). The parabolic
tions) into polynomial transfer functions that can be easily deployed in models was extended for higher charge–discharge rates
control-oriented applications (Xu, Wang, Pei, Mao, & Zhu, 2020). The (Khaleghi Rahimian, Rayman, & White, 2013). Specifically, Li-ion
𝑁th-order Padé approximation of a transcendental transfer function concentration and potential profiles in the electrolyte phase were
𝐺(𝑠) is a ratio of two polynomials in 𝑠 where the denominator is of order approximated by a polynomial function. A cubic polynomial was used
𝑁. The numerator is of order 𝑁 or less. Normally, the computational for electrolyte concentration and potential inside electrodes, while
speed is determined by the denominator order as: separator liquid phase potential and electrolyte concentration were
∑𝑀 𝑚 calculated by parabolas. Diffusion inside solid particles was simplified
𝑚=0 𝑏𝑚 𝑠 𝑛𝑢𝑚(𝑠)
𝐺(𝑠) = ∑𝑁 (𝑀 ≤ 𝑁) = (12) using an approximate solution based on the analytical solution for solid
1+ 𝑎 𝑠𝑛 𝑑𝑒𝑛(𝑠)
𝑛=1 𝑛 concentration. This type of reduced model could accurately predict the
where 𝑀 is the order of numerator and 𝑁 is the order of denominator. battery cell voltage with less than 1% error for discharging/charging
𝑏𝑚 and 𝑎𝑛 are coefficients for the numerator and the denominator, current rates up to 5C, while the average computation time was reduced
respectively. by a factor of 5.
5
2.1.4. Discretization-based method electric potentials are spatially homogeneous after asymptotic reduc-
For nonlinear PDE models, analytical solutions would become dif- tion and cell voltage behavior can be understood through a sequence
ficult to be derived. In this context, numerical methods could be of asymptotic regimes, which elucidates simple underlying physical
employed to discretize the governing equations of PDE (Qadir, Li, & processes. The asymptotic methods were used to reduce a thermo-
Chen, 2022). Finite-Element Method (FEM) is a numerical technique electrochemical model and obtain solutions for common battery opera-
for discovering approximate solutions to PDEs. The FEM discretized the tion modes (Hennessy & Moyles, 2020). Asymptotic techniques were
domain 𝑥 ∈ (0, 𝐿) into 𝑁 − 1 subdomains. The 𝑁th-order FEM approx- used to derive a thermal-electrochemical model systematically. The
imation has 𝑁 nodes. A FEM was used to develop three-dimensional physical parameters and variables were nondimensionalized as:
thermal abuse model on lithium-ion batteries. A thermal model for 𝛷0 𝐹
a cylindrical battery was developed based on the FEM (Wang et al., 𝜆= , 𝛿 = 𝜆−1 (15)
𝑅𝑇
2017). A particle-resolved 3D finite element model was implemented
where 𝑅 is the gas constant, 𝑇 is the temperature, 𝐹 is Faraday
to study the effect of cathode microstructure on the electrochemical
constant. All the variables could be expanded in powers of 𝛿, using the
and mechanical responses of Li-ion batteries (Mai, Yang, & Soghrati,
notation as:
2019).
Finite Difference method (FDM) is the simplest and widely-utilized 𝛷𝑘 = 𝛷𝑘0 + 𝛿𝛷𝑘1 + 𝑜(𝛿 2 )
(16)
approach to handle the PDEs of battery electrochemical models. In the 𝑖𝑘 = 𝑖𝑘0 + 𝛿𝑖𝑘1 + 𝑜(𝛿 2 )
FDM, spatial domain is discretized into 𝑁 nodes at 𝑥 = 0, ℎ, … , (𝑁 −1)ℎ, The leading order expansion was taken for model approximation.
where ℎ = 𝐿∕(𝑁 − 1) is assumed constant for simplicity. Spatial deriva- The reduced model was validated against a thermal P2D model and
tives are approximated by forward−dif ference, backward−dif ference, or demonstrated very high accuracy with a computational cost reduced
central-difference. The mixed FDM was selected to discretize the spatial by over forty times (Brosa Planella, Sheikh, & Widanage, 2021).
𝑟, where the node points are unevenly spaced (Sharma et al., 2019).
The mixed FDM can yield higher resolution with fewer node points in
2.2. Comparison and discussion
contrast to equally-spaced node points. The FDM was used to mesh the
active particles along radial direction (Xiong et al., 2018). Then the
Table 2 summarizes various types of reduction methods for P2D
complex PDE were discretized into ODEs and the state space expression
model in terms of their reference, advantages and drawbacks. It should
of system was obtained. The governing equations were discretized with
be known that the SPM based methods are the most widely utilized
central difference scheme for the first and second derivatives (Kong,
solution to simplify the P2D model. This method requires minimum
Wetton, & Gopaluni, 2019).
computation resources and is able to provide sufficient model accuracy
at low current rates. As an improved version of SPM, the SPM with
2.1.5. Physics-based equivalent circuit models electrolyte dynamics (SPMe) could predict the cell voltage at a high
Due to the involvement of electric circuit elements to describe cell C-rate with lower computation burden than P2D model. However, due
responses, ECMs can be identified rapidly. Here simplicity and practi- to the electrode width is much larger than particle radius, while only
cability are two obvious merits for the application of ECMs (Carnevale, one particle for one electrode is insufficient, this SPMe would be limited
Possieri, & Sassano, 2019). However, ECMs also present the limited ca- for cells with thick electrodes. For the transfer-function based solutions,
pability in terms of electrochemical mechanisms. Recently, P2D models although they are much more complicated than SPM based methods
have been simplified into the form of ECMs, which could significantly especially for transcendental transfer functions, this type of solutions
reduce the computational burden and enrich the physical meanings of could provide high accuracy in both frequency range and time domain.
ECMs. Finite volume method (FVM) was applied to a P2D model of Li- Specifically, residue grouping methods reduce the model order by op-
ion batteries (Li, Vilathgamuwa, Farrell, Choi, Tran, & Teague, 2019). timizing the eigenvalues and residue vectors, while the residues would
The standard passive components were used to construct ECMs, indicat- be grouped with similar eigenvalues. However, the electrochemical
ing the energy storage nature of Li-ion battery. The parametric values parameters must be given before residue grouping. DRA could also
of circuit elements were expressed as the functions of Li-ion concentra- simplify P2D model to a low-order approximation and accurately match
tions and temperature. Results illustrated that the developed ECM could the frequency range. However, the battery electrochemical parameters
achieve close agreement with P2D model under a wide range of applied must be known a priori, further making it difficult to parameterize DRA
current rates, but occupies a much reduced computational resource. in an iteration process. The Padé approximation is an effective tool to
A novel implementation for P2D model was presented (Geng, Wang, convert transcendental transfer functions into polynomial expressions.
Lacey, Brandell, & Thiringer, 2021). This model had a circuit-based The accuracy of the reduced model relies on the order selection of
structure without any simplification of the physical origins. A classic Padé approximant. A high-order approximation could provide sufficient
transmission line structure was used to replace the governing equations accuracy at high frequency but at the cost of increased computational
of current distribution within electrode. The concentration distributions efficiency. Under the assumption of Li-ion concentration distributions,
were solved with the FDM. A new empirically parameterized physics- the integral approximation method is a much simplified approach to
informed ECM was derived, as depicted in Fig. 3 (Merla, Wu, Yufit, derive the transfer functions of reduced models.
Martinez-Botas, & Offer, 2018). The P2D physics is discretized and For the polynomial expression-based solution, the polynomial ex-
represented as circuit elements in a network, requiring only 3 low-cost pressions of battery electrochemical variables are capable of provid-
in-situ experiments: slow discharge, pulse discharge, and electrochem- ing accurate prediction with significantly reduced computational ef-
ical impedance spectroscopy (EIS) under load. The model showed an fort. However, in general, the coefficients of parabolic functions are
accuracy of 19.6 mV compared to the multiple C-rate pulse loading empirically fitted, hence lacking physical meanings.
experiments. Spatial discretization-based solutions belong to the commonly uti-
lized numerical approaches for simplifying P2D models. Here the FEM
2.1.6. Asymptotic reduction methods discretizes the domain into multiple subdomains. Due to the way of
Asymptotic techniques provide an effective means for model reduc- approximating solutions with numerous basic elements, FEM becomes
tion and solution construction with clear ranges of validity. For exam- the most accurate one with a heavy computational burden. For the
ple, the asymptotic reduction of P2D model was performed by Moyle FDM, it can also simplify P2D model with high accuracy. However, the
et al. based on the fact that reaction kinetics dominate electrical ef- spaces and number of discretized nodes for all numerical methods must
fects (Moyles, Hennessy, Myers, & Wetton, 2019). They showed that the be carefully determined to ensure reasonable convergence.
6
Fig. 3. Cell level model schematic diagram for a single particle model (Merla et al., 2018).
Table 2
Different P2D model reduction techniques and their control-oriented applications.
Category Simplification technique Refs Strengths Weaknesses
SPM Guo et al. (2011) Simple; Low Limited capability at
SPM Guo and White (2012) computation burden; high C-rates.
based Ning and Popov (2004)
Santhanagopalan, Guo, Ramadass,
and White (2006)
SPMe (Improved with Li, Adewuyi, Lotfi, Landers, and Low computation Limited capability for
electrolyte dynamics) Park (2018) Malik, Li, and Chen burden; Enhanced cells with thick
(2021) Moura et al. (2017) Han performances at high electrodes.
et al. (2015) Mehta and Gupta C-rates, cell aging, and
(2021) Perez, Dey, Hu, and thermal behavior; Easy
Moura (2017) for control-oriented
implementations
Residue Smith et al. (2007) Baba, Low-order; Accurate in Requiring pole/residue
Transfer- grouping Yoshida, Nagaoka, Okuda, and partial frequency range; optimization;
function Kawauchi (2014) Computation efficient; Prior-known
type parameters.
Discrete-Time Lee et al. (2012a) Ortiz-Ricardez, Low-order; Accurate in Prior-known
Realization Romero-Becerril, and overall frequency parameters; Hard-to
Algorithm (DRA) Alvarez-Icaza (2021) range; Computation parameterize.
efficient; Easy for
control-oriented
implementations
Padé Tran et al. (2018) Yuan et al. Low-order; Accurate in Trade-off between order
approximation (2017a) Yuan et al. (2017b) Xu partial frequency range; selection(accuracy) and
et al. (2020) Computation efficient; computation burden.
Easy for
control-oriented
implementations.
Integral method Gopalakrishnan, Zhang, and Offer Low-order; Assumed Li-ion
approximation (2017) Subramanian et al. (2004) Easy-to-derive; concentration profiles.
(IMA) Computation efficient
Polynomial Parabolic functions Khaleghi Rahimian et al. (2013) Low-order; Lacking physical
expression Deng et al. (2018) Tanim and Easy-to-derive; meanings of polynomial
Rahn (2015) Subramanian, Computation efficient; coefficients.
Diwakar, and Tapriyal (2005)
Spatial Finite-Element method Li, Cui et al. (2021) Dickinson High accuracy Heavy computation
Discretiza- (FEM) et al. (2014) Guo et al. (2010) burden. Convergence
tion Wang et al. (2017) requirement. Difficult
for control-oriented
implementations.
Finite Difference Kong et al. (2019) Xiong et al. High accuracy Sensitive to spaces and
method (FDM) (2018) Sung and Shin (2015) number of discretized
nodes. Convergence
requirement.
Physics-based Physical analogies and Li, Vilathgamuwa et al. (2019) Low-order; Lumped parameters.
ECMs interpretation Geng et al. (2021) Ashwin et al. Easy-to-parameterize. Sensitive to the number
(2017) Merla et al. (2018) Computation efficient; of elementary sections;
Asymptotic Leading-order Brosa Planella et al. (2021) High accuracy; Nondimensionalized
reduction approximation Hennessy and Moyles (2020) Computation efficient; parameters;
Leading-order Moyles et al. (2019) Labor-intensive model
approximation derivation; Hard for
control-oriented
implementations
The physics-based ECMs are essentially physical interpretation of reduced by lumping the parameters, while its performance highly relies
electrochemical mechanisms. Similar to ECMs, this type of solutions on the selection of the number of elementary sections.
can be implemented easily but also reflects the electrochemical process The asymptotic reduction simplifies the P2D model by using a
to some degree. For the physics-based ECM, its model parameters are leading-order approximation. Here the electrochemical reactions at
7
different time scales are considered. All the known electrochemical accelerate the side reaction rate and contribute to lithium plating. In
parameters are nondimensionalized. The derivation of the asymptotic order to minimize degradation rate and reduce charging time simul-
reduction is relatively complicated. taneously, a charging method considering different limiting factors,
The electrochemical model could enhance the battery control per- including surface ion concentrations and side reaction rate, are crucial
formances on cell state estimation, fault diagnosis, and health-aware for cell health-aware charging protocol. Based on SPMe, the dynamic
optimal charging. The physical model parameters reflect the inherent programming (DP) technique was employed to find the trade-off be-
characteristics of the cell. For example, the volume fraction of active tween charging duration, solid electrolyte interphase (SEI) growth, and
materials and maximum lithium-ion concentration determine the cell lithium plating. The experimental results showed that the optimized
capacity. Thus, the cell degradation mechanisms including the loss of fast charging strategy could reduce the capacity fade significantly com-
recyclable lithium ions and active materials can be identified by observ- pared with the widely used constant-current constant-voltage (CCCV)
ing these parameters. Similarly, the cell fault can be diagnosed through charging method without sacrificing battery health (Lin, Hao, Liu, &
the estimation of the variation of the physical parameters. In addition, Jia, 2018). An optimal charging current at different SOCs was found
the electrochemical model could provide cell internal states such as using nonlinear model predictive control, which reduces side reaction
the lithium-ion concentrations in the electrolyte and solid phase, side rate and lithium plating rate. Pulse discharging current was added to
reaction overpotential, and reaction flux across the electrodes. These promote the lithium stripping. With a comparable capacity fade, the
electrochemical states are related to the cell aging process. Hence, cell proposed charging method was tested to reduce 39% of charging time
health-aware control could be realized by optimizing these electro- compared with 1C CCCV charging method (Yin & Choe, 2020).
chemical states. To sum up, the electrochemical model represents the Similarly, the fast charging protocol considering side reaction rate
cell interior dynamics. This could be helpful to achieve accurate cell and ion concentration was proved to reduce more than 40% in charging
state estimation, fault diagnosis, and efficient charging control. time compared with normal charging method but with less capacity
To be specific, the SPM model has better controllability and observ- and power fade (Yin, Hu, Choe, Cho, & Joe, 2019). To prevent the
ability than the P2D model (Muddappa & Anwar, 2014). An adaptive internal short circuit caused by lithium plating at the negative elec-
unscented Kalman filter based on SPM was proposed to estimate the trode, a fast charging control algorithm with two closed loops were
SOC, lithium-ion concentrations, and potentials (Li, Fan et al., 2020). presented. One loop was an anode over-potential observer, whereas
The estimated internal information is valuable to avoid lithium plating. the second loop includes a feedback structure that adjusted the cur-
All estimated states were verified to demonstrate fast convergence, rent based on estimated status of lithium plating. The results showed
robustness, and high accuracy even with a 20% initial SOC error. By as- that the cell can be fully charged within 52 min without lithium
suming that the total amount of lithium in the cell is known, the dynam- deposition (Chu et al., 2017). The electrochemical model parameters
ical properties of the SPMe observer, including the marginal stability, have rich physical meanings and were considered as the referenced
local invertibility, and conservation of lithium can be exploited (Moura values for a healthy battery. However, the battery fault conditions
et al., 2017). such as aging, overcharge/discharge cause significant variations of
Besides, the convergent estimates of solid and electrolyte phase parameters from nominal values. An output error injection-based par-
concentrations on high C-rate cycles and transient electric vehicle tial differential–algebraic equation observer was applied to produce
charge/discharge cycles were also illustrated. For phase-transition ma- voltage error signals. These imposed error values could be used to
terials during the battery’s normal charge and discharge operation detect the ongoing fault conditions of the battery (Rahman, Anwar, &
such as LiFePO4 , a novel boundary observer to estimate the concen- Izadian, 2015). The fuzzy logic system was used to detect any variation
tration of lithium ions together with a moving boundary radius from in battery cell parameters and diagnose battery faults accurately in
the SPM via the backstepping method for PDEs was derived (Koga, real-time (Muddappa & Anwar, 2014).
Camacho-Solorio, & Krstic, 2021). During the lithium ions intercalation Table 3 summarizes different P2D model-order reduction techniques
and deintercalation, the stress generation contributes to the electrode and their control-oriented applications to cell state estimation, optimal
particle fracture. A coupled SPM-mechanical stress model was used to control, and fault diagnostics. It can be noted that SPM based models
design an adaptive observer for cell SOC, electrode particle stress, and are suitable for developing cell state estimation, optimal control strat-
solid-phase diffusivity. The internal stress and SOH-related parameters egy, and fault diagnosis algorithms. The spatial discretization method
are predicted from real-time electric current and terminal voltage is the most accurate technique. Commercial software such as COMSOL
measurements (Zhang, Dey, Couto, & Moura, 2019). A model from the uses the FEM to solve the PDEs and is often used as the benchmark
pre-compute set of physics-based reduced-order components that span for evaluating the simplified models (Dickinson et al., 2014). However,
the expected dynamics of the cell over its lifetime was presented. Two FEM is difficult to be implemented for control-oriented applications.
significant aging mechanisms were considered: solid-electrolyte inter- The polynomial profiles and physics-based ECMs could significantly
phase layer formation in the negative electrode and material dissolution reduce the computational complexity. The transfer-function type and
in the positive electrode. An interacting multiple-model Kalman filter asymptotic reduction models could achieve high accuracy as well.
was used to select the pre-computed model for accurate estimations However, the computational complexity of the asymptotic reduction
of cell internal electrochemical variables and output terminal volt- approach is relatively high, further limiting its applications in real
age (Smiley & Plett, 2018). A composite LiMn2 O4 −LiNi1∕3 Mn1∕3 Co1∕3 O2 battery control-oriented management.
electrode battery model was used in dual-nonlinear observers to esti-
mate the cell SOC and loss of cyclable lithium over time (Bartlett et al., 3. Extension of P2D model
2016). The surface and bulk lithium concentration of each material
as well as the current split between each material were predicted. As discussed in Section 2, battery P2D model is mainly based
The reduced-order model was incorporated in a sigma-point Kalman on the porous electrode theory. The P2D model would simplify the
filter to enable measurement feedback to improve voltage and internal- modeling of complex electrode microstructure with the volume-average
variable estimates. The system was implemented in a microcontroller method at a macroscopic scale where the ion transport dynamics in the
and proved to be valid for BMSs (Miguel et al., 2021a). An electro- electrolyte and electronic conduction process in the solid phase are all
chemical model-based solution provided a basis for an output-injection captured, and a microscopic scale on which the ionic diffusion inside
observer to estimate the SOC over a wide range of operations, especially the electroactive particles is also considered. Due to the computational
at high discharge rates (Lotfi, Landers, Li, & Park, 2017). burden and accuracy, the P2D model is mainly adopted for battery
Battery fast charging is one of the most effective techniques that electrode structure design. However, the P2D model has several issues
affect the acceptance of EVs. Increasing the charging current could when it is used for the development of battery simulation technologies:
8
Table 3
Different P2D model reduction techniques and their control-oriented applications.
Category Cell state estimation Optimal charge control Fault diagnosis
SPM based Li, Fan et al. (2020) Moura et al. (2017) Lin et al. (2018) Yin et al. (2019) Yin Rahman et al.
Koga et al. (2021) Zhang et al. (2019) Bi and Choe (2020) (2015)
and Choe (2020) Li, Landers, and Park
(2020) Dey, Ayalew, and Pisu (2015)
Transfer-function Smiley and Plett (2018) Stetzel, Aldrich, Perez, Shahmohammadhamedani, and
type Trimboli, and Plett (2015) Bartlett et al. Moura (2015) Xavier, de Souza, Karami,
(2016) Miguel et al. (2021a) Plett, and Trimboli (2021) Lin, Wang,
and Kim (2019)
Polynomial Sturm et al. (2018) Zou, Hu, Wei, Wik, and Egardt (2018) Muddappa and
expression Lotfi et al. (2017) Hu, Cao, and Egardt Gao et al. (2020) Chu et al. (2017) Anwar (2014)
(2018)
Spatial Ashwin, Chung, and Wang (2016) Pozzi, Torchio, Braatz, and Raimondo
Discretization (2020)
Physics-based ECMs Li, Vilathgamuwa et al. (2019) Fleischer, Li, Vilathgamuwa et al. (2021)
Waag, Heyn, and Sauer (2014)
Asymptotic None None None
reduction
First, the particle size is assumed to be uniform in a P2D model, active cathode particle based on multi-particle theory was derived. The
which does not accurately represent the electrode topology such as dimensionality reduction is based on the volume averaging of dynamic
the actual size distribution of particles, the agglomerate structure of equations (Zelič & Katrašnik, 2019). The obtained 0D model has an ex-
particle in the cathode. Second, the positive electrode is regarded as plicit advantage of short computation time and high level of accuracy.
a single component without considering the blended cathode active A simplified physics-based model was derived combining the porous-
materials. Third, the Bruggeman equation is used to model the effects of electrode theory and multi-particles to predict a LFP-based electrode
porous structure on the ionic transport and taken as constant, ignoring performance (Mastali Majdabadi et al., 2015). As shown in Fig. 5,
the microstructure variability. The Fick’s diffusion law oversimplifies four-particle bins were considered to represent the apparent particle
the intra-particle ionic transport. To address these aforementioned size distribution (PSD) of the cathode. A polynomial approximation
limitations, some attempts including the development of multi-particle method was incorporated to model the effects of electrolyte. A funda-
model, multiple-scale model, coupled aging and thermal model have mental analysis of the effect of particle size distribution on graphite
been made and summarized in this section. electrodes and their performances was presented. The effects of parti-
cle heterogeneity on surface overpotentials were investigated (Röder,
3.1. Multi-particle model Sonntag, Schröder, & Krewer, 2016). It was concluded that for a
graphite electrode, the surface-area- and volume-based approximations
Many intercalation materials exhibit phase changes during bat- are sufficiently accurate for large- and small-scale PSDs.
tery charging/discharging process. Phase transitions strongly affect the The commercially available positive materials such as lithium fer-
lithium distribution inside particles and change the particle potentials. rous phosphate (LFP), lithium cobalt oxides (LCO), lithium nickel man-
LFP is a typical cathode chemistry with phase change (Sharma et al., ganese cobalt oxides (NMCs), lithium manganese spinel (LMO), and
2019). The coexistence of Li-poor and Li-rich phases should be ac- lithium nickel cobalt aluminum oxide (NCA), have their own charac-
counted for and the boundaries between these two phases should be teristics in terms of power and energy costs. To obtain an acceptable
tracked, while the P2D model fails to analyze this phenomenon (Biz- trade-off between these characteristics, battery manufacturers would
eray, Zhao, Duncan, & Howey, 2015). A many-particle model was mix two or more positive materials to form a blend electrode. However,
introduced where particles within a porous electrode are allowed to the P2D model assumes the single composition of the active material.
randomly exchange Li+ ions and electrons through the electrolyte and A Li-ion mathematical model was developed to treat two positive-
conductive matrix, respectively (Dreyer et al., 2010). A general expla- electrode materials (Li𝑦 Ni0.80 Co0.15 Al0.05 O2 and Li𝑦 Mn2 O4 ) (Albertus,
nation of the occurrence of non-monotonic chemical-potential behavior Christensen, & Newman, 2009). To match the behavior of high rate
in LFP electrodes was presented through two possible scenarios of new discharge and impedance data, multiple types of electronic connections
phase formation in a many-particle system, as shown in Fig. 4. At between particles and particle-conductive matrix were required. An ex-
location A, all particles behave as a single phase. At location B, two tension of P2D model, accounting for the agglomerate of active material
alternative scenarios are available to the particle ensemble. The bottom particles, LiNi1∕3 Mn1∕3 Co1∕3 O2 (NCM) and LiNi1∕3 Co1∕3 Al1∕3 O2 (NCA),
path is that a single particle will hose a two-phase region, the top was derived by Lueth, Sauter, and Bessler (2015). The morphology of
alternative is that the other particles will contribute lithium to the agglomerates can be seen in Fig. 6a. Micro-sized spherical secondary
filling particle so that the relaxed result will be either fully lithiated or particles are the agglomerate of nanometer-sized primary particles
delithiated particles. At location C, particles can internally host phase which would be held together by binder (Huang, Ge, Li, & Zhang, 2014)
boundaries, or the particles can either be lithiated or delithiated to the as shown in Fig. 6b. The model assumes that the liquid-transport is
stable locations (B and D). faster than the solid phase transport, while the electronic conductiv-
To develop an adequate modeling representation of the topology of ity within agglomerate is sufficiently high. Therefore, the concentra-
the anode and cathode particles, Farkhondeh et al. used a multi-particle tion polarization is significantly smaller than compact particles, which
model with variable solid-state diffusivity (Farkhondeh & Delacourt, alleviates the rate-limiting effect of the solid diffusion on compact
2011). The memory effect of LFP electrodes was also analyzed. Sim- particles.
ulation/experiment comparison for cells with the current rates up to A LiNi1∕3 Mn1∕3 Co1∕3 O2 − LiMn2 O4 (NMC-LMO) blended cathode ob-
1C demonstrated the robust particle-size distribution estimation for tained from a commercial Li-ion battery was considered (Mao, Farkhon-
galvanostatic charges/discharges. A 0D model of a phase separating deh, Pritzker, Fowler, & Chen, 2015). As NMC forms agglomerates, it
9
Fig. 4. Chemical potential of LFP cathode evolution with two possible scenarios.
Fig. 5. Schematic diagram of the LFP cathode that is assumed to contain four particle sizes.
was divided into two classes of particles, i.e., primary particles and these reduced-order multi-particle models are promising for control-
secondary particles. The secondary particles are assumed to be large oriented applications (Santos-Mendoza, Henquín, Vazquez-Arenas, &
non-porous NMC agglomerates. Three distributions from LMO particles, Aguirre, 2020).
NMC primary particles, and NMC secondary particles were accounted
for the multiple particle sizes of the active materials. The model could 3.2. Electrochemical model with multiple scales
accurately capture the galvanostatic discharge at various current rates.
The P2D model was modified to accommodate a composite electrode The P2D model was derived by using volume averaging method
of LiMn2 O4 (LMO) and Li1∕3 Ni1∕3 Mn1∕3 Co1∕3 O2 (NMC), as shown in (VAM), where the local microstructure features such as pore shape and
Fig. 6c (Bartlett et al., 2016). A composite electrode is viewed as two localized current density distribution are neglected. This causes the P2D
parallel particles, one representing each active material. The negative model become insufficient for applications involving active material
electrode is a single material graphite electrode. The presented model utilization. The cell multi-scale is depicted in Fig. 7. The macroscopic
structure was applied to battery SOC/SOH co-estimation. A high-power models from microscopic processes with one smaller length scale to
cell with lithium manganese spinel and lithium cobalt oxide at the pos- account for geometric microstructure features such as pore shape was
itive while lithium titanate at the negative with 2 particle populations derived by mathematical homogenization theory derives.
was studied with an SPMe model (Petit, Calas, & Bernard, 2020). A The multi-scaled homogenized model is developed using asymptotic
simplified electrochemical model was developed to describe and maxi- analysis of the microscopic model. The homogenization assumes that
mize the specific capacity of battery cells through the variation of active the porous electrode is composed of periodic micro-units. Each micro
(LFP)/inactive (carbon conductor and binder) ratios. The computation unit is composed of solid active materials 𝑆, electrolyte 𝐿, and solid-
time was significantly reduced with these simplified models. Therefore, electrolyte interface 𝑇 , as shown in Fig. 8. The ratio of micro unit
10
Fig. 6. Multi-particle agglomerate coupling with P2D model(a), agglomerate model(b), and simplified multi-particle model(c).
length 𝑙 to macro length 𝐿 is denoted as spatial multiscale factor So both the concentration and current density are dominated by the
𝛿. 𝛿 = 𝑙∕𝐿. Microdomain coordinate 𝑦 and macrodomain coordinate variations of battery length scale with relatively smaller variations on
𝑥 is converted by 𝑦 = 𝑥∕𝛿. The basic idea of the homogenization the electrode-particle length scale. The current density 𝑗 flows around
method is to expand the system variables at multiple scales, and then the electrode particles and hence displays significant variations on the
substitute these expansions into microscopic equations to obtain PDEs particle scale.
with different scale components. Then, the expanded PDE is analyzed A comparative analysis of the full homogenized model and P2D
by 𝛿 at different orders (Taralova & Lai, 2011). The central ansatz for model was conducted (Arunachalam & Onori, 2019). The results illus-
the expansions of the variables is: trate that the predictability of P2D model deteriorates when predicting
battery voltage response at low SOC for high operational tempera-
𝑐 = 𝑐0 (𝑥, 𝑡) + 𝛿𝑐1 (𝑥, 𝑦, 𝑡) + 𝛿 2 𝑐2 (𝑥, 𝑦, 𝑡) + ⋯ , ture under 1C discharging case. The P2D model failed to capture the
(17)
𝑗 = 𝑗0 (𝑥, 𝑡) + 𝛿𝑗1 (𝑥, 𝑦, 𝑡) + 𝛿 2 𝑗2 (𝑥, 𝑦, 𝑡) + ⋯ , polarization of electrolyte under these scenarios, as shown in Fig. 9.
11
Fig. 7. Li-ion battery structures at multiple length scales.
Therefore, the homogenized macroscopic model presents a wider range and Lai (2011). The results indicated that the classical Doyle–Fuller–
of applications. A homogenized model was developed, which considers Newman (DFN) model is only the 0th order terms of the expansion.
lithium diffusion within particles, lithium transfer from particles to the More specifically, the DFN-based approach is only valid under the
electrolyte and transport within the electrolyte. It was found that for assumption that 𝛿 ≪ 1. Ciucci’s analysis was extended by computing
the asymptotic order of interface exchange current densities, which is
low discharging currents, battery acts almost uniformly (Richardson,
an important factor in the homogenization study (Taralova, Iliev, &
Denuault, & Please, 2011). When the current is increased to a critical Efendiev, 2016). As the commonly used Bruggeman’s formula may in
value, Li-ions in some regions of the battery are depleted, leading to the fact violate the physical bounds in some regimes, the exact values of
spatially nonuniform use of electrode. The conversion of homogenized the different effective transport coefficients were investigated (Gully
equations to P2D model was carried out by Ciucci et al. Taralova et al., 2014). An extended homogenized porous electrode model was
12
Fig. 8. Representation of the porous electrode of a lithium-ion battery in the form of spatially periodical unit. Every unit cell y is composed of active particles S and electrolyte
solution L, that are separately by an interface T.
one classical solution is to extend P2D or simplified 1D electrochemical

models to 2D or 3D thermal models. In this way, two sub-models
including an electrochemical model and a thermal model would be
operated simultaneously. As illustrated in Fig. 10, a 1D electrochemical
model would predict the heat generation rate (𝑄ℎ) and voltage (𝑉 ) of
battery cell through various loading cycles. The voltage at a point from
1D model which corresponds to the tab of battery cell is then mapped
to the tab of 2D thermal-electric model, and this battery 2D model
allows the consideration of both temperature and current distributions
across the surface of battery cell (Tourani, White, & Ivey, 2014). A 1D
electrochemical model was adopted to calculate the heat generation
rate, as depicted in Fig. 10 (Jiang, Zhuang, Hu, Liu, & Huang, 2020).
Then the heat source is coupled into a 3D thermal model for calculating
the temperature distribution inside a battery. The obtained average
temperature would affect multiple parameters within the electrochemi-
cal model. As a result, the heat generation rate of 1D model would vary
based on different temperature-dependent parameters. In the thermal
model, the energy conversion equation based on the basic principle of
heat transfer can be given as:
𝜕𝑇 𝜕2 𝑇 𝜕2 𝑇 𝜕2 𝑇
Fig. 9. Performance of the P2D model and full homogenized model (FHM) against
𝜌𝐶𝑝 = 𝜆𝑥 + 𝜆𝑦 + 𝜆𝑧 + 𝑞̇ (18)
𝜕𝑡 𝜕𝑥2 𝜕𝑦2 𝜕𝑧2
measured voltage from 1C-rate discharge test conducted at 52 ◦ C (Arunachalam &
Onori, 2019). where 𝜆𝑥 , 𝜆𝑦 , and 𝜆𝑧 are thermal conductivities in the x, y, and z
directions of the battery, respectively. The value of 𝑞̇ refers to heat
generation in the cell, which includes reversible heat and irreversible
presented considering the electrochemical coupling between particles heat. The reversible heat is mainly related to the entropy change of the
of a real particle size distribution obtained from tomography data. electrode active material. The irreversible heat consists of the active
The vanishing plateaus in the graphite potential curve with increasing polarization heat arising from the electrochemical reaction and ohmic
charge or discharge rates was explained (Traskunov, 2015). The P2D heat.
model is difficult to predict the heterogeneous reactions within the Due to the increased complexity of coupled multiscale model, the
porous media with complex microstructures. An accurate standard underlying physical models need to be simplified for battery control-
homogenization of the battery was presented to analyze how the fluctu- oriented applications. Li, Choe, and Joe (2015) reported a reduced-
ations occur and estimate them theoretically (Traskunov & Latz, 2021). order electrochemical thermal model of a pouch cell. The ion con-
However, it should be noted that the control-oriented applications of centration within an electrode is approximated using the polynomial
the homogenized model are rarely-researched. approach, while the ion concentration within electrolyte is simplified
using the state space method. Both potentials and electrochemical ki-
3.3. Electrochemical model with aging and thermal effects netics are linearized. Cell temperature is calculated based on the energy
conversion formula to further determine the temperature-dependent
Besides, apart from electrochemical behaviors, battery thermal and diffusion coefficients. Then the model performance is evaluated by ana-
aging (degradation) dynamics also plays a vital role in determining lyzing appropriate experimental data at multiple cycles under different
battery performance. In this context, battery electrochemical model is operational conditions. Illustrative results confirm that the computa-
worth being expanded to consider these two aspects. To achieve this, tional effort of the reduced-order model (ROM) would be reduced
13
Fig. 10. The coupled 1D electrochemical model and thermal model.
to the fifteenth, while its accuracy can be still maintained. It should models. Illustrative results indicate that the cracking and fracture of SEI
be known that various simplified versions of coupled models have layer would be significantly affected by cycling operating conditions
been proposed. For example, the accurate electrochemical and ther- and play an important role in accelerating the electrochemical mecha-
mal response of Li-ion cells were explored (Gambhire et al., 2015). nisms. A growth rate of around 0.64 nm/hr at 1C for the SEI film on
Here the local heat generation and spatial temperature variation are the negative electrode was observed. In addition, for a convective heat
volume averaged to derive the reduced-order coupled PDEs of battery transfer coefficient (ℎ) larger than 1𝑊 ∕(𝑚2 𝐾), there is limited effects
electrochemical-thermal model. This ROM structure could be easily of battery aging model on surface cell temperature (Lamorgese, Mauri,
utilized for parameter identification of controlling processes at high & Tellini, 2018).
current rates. A control-oriented thermal-electrochemical model was In theory, mechanical strain would also play a pivotal role in af-
developed and well validated for the commercial prismatic battery (Li, fecting battery durability. In this context, the coupling between battery
Yang, & Li, 2021). Here the multilayer thermal model and polynomial electrochemical and mechanical dynamics can also provide significant
approximation method are applied to describe battery thermal behavior insights into cell degradation. The thermo-mechanical behavior within
and electrochemical process, respectively. This 3D battery model is a multilayer section of lithium 18650 cell during discharge (Valentin,
divided into dozens of thermal nodes connecting by the thermal resis- Thivel, Kareemulla, Cadiou, & Bultel, 2017) was evaluated.
tances. The multiplayer thermal model is constructed by using thermal The equation of heat transfer in each part of the cell is expressed
balance equations for all thermal nodes. by:
In addition, the coupled electrochemical-thermal model also allows 𝑑
the investigation of aging effects concurrently. It should be known (𝜌𝐶𝑝 𝑇 ) = −𝑑𝑖𝑣(−𝜆𝑔𝑟𝑎𝑑(𝑇 )) + 𝑞̇ (19)
𝑑𝑡
that battery aging significantly depends on temperature and could
where 𝑇 denotes temperature, 𝜌 density, 𝐶𝑝 specific heat capacity, 𝜆
lead to significant heat release. Aging would generally occur at the
thermal conductivity and 𝑞̇ heat sources.
active material particle scale, while heat transport occurs at the cell
The heat generation 𝑞̇ per unit volume of the cell is given as:
scale. Kupper et al. investigated the battery multiscale heat, mass
transport and complex electrochemical reaction mechanisms based on 1 𝑑𝐸𝑂𝐶𝑉
𝑞̇ = [𝑖(𝑉 − 𝐸𝑂𝐶𝑉 ) + 𝑖(𝑇 )] (20)
a coupled model (Kupper & Bessler, 2016). Here this coupled model is 𝑉𝑏𝑎𝑡 𝑑𝑇
derived by considering the homogenization of three 1D models: one for where 𝑖 is cell current,𝑉 is battery potential, 𝐸𝑂𝐶𝑉 open-circuit voltage
intra-particle lithium diffusion, one for electrode-pair mass and charge and 𝑉𝑏𝑎𝑡𝑡 battery volume. The partition of strains leads to write the total
transport, and one for cell-level heat transport. The main and side strain as the sum of all possible strains:
reactions are described with the flexible multi-phase electrochemistry.
Besides, the thermal model is able to provide a feedback between 𝜀𝑡𝑜𝑡𝑎𝑙 = 𝜀𝑒 + 𝜀𝑇 + 𝜀𝐿 (21)
temperature and battery electric dynamics as well as aging. One of the where 𝜀𝑒 the elastic strain tensor, 𝜀𝑇 is the thermal strain tensor and 𝜀𝐿
challenges for anode side is that dendritic and mossy lithium surface is the lithiation/delithiation expansion of electrodes. The stress–strain
can form solid electrolyte interphase (SEI) easily during cycling, as relationship is:
shown in Fig. 11 (Louli, Ellis, & Dahn, 2019). For fresh anode surface,
the particles are covered with an SEI layer. When the anode is lithiated 𝐸 𝜐
𝜀𝑒 = 𝜎 − 𝑡𝑟(𝜎)𝐼 (22)
1+𝜐 𝐸
during charge process, a large volume expansion occurs as shown in
Fig. 11B. The SEI layer could be damaged, leading to the exposure E being the Young’s Modulus, 𝜐 the Poisson ratio and I the identity
of the particles to electrolyte. Hence, a thicker SEI layer would form tensor. The thermal strain is expressed by:
consuming the lithium inventory as depicted in Fig. 11C. After many 𝜀𝑇 = 𝛼(𝑇 − 𝑇0 ) (23)
cycles, a much thickened SEI evolves as shown in Fig. 11D. For the
cathode side, higher mechanical stress would accelerate the cracks and where 𝛼 is the coefficient of thermal expansion and 𝑇0 the initial
de-cohesion of particles. Furthermore, the cathode-induced oxidation temperature and 𝑇 is the actual temperature.
of solvents will cause gas generation (Mai et al., 2019). As shown in Fig. 12, lithium diffuses in the lattice of electrode struc-
A simplified electrochemical and thermal aging model (Prada et al., ture at the electrode scale, further leading to structural changes and
2013) was presented. Here the mechanism of porosity modification mechanical stresses. At the same time, the macroscopic electrochemical
caused by the growth of solid electrolyte interphase (SEI) film at the model calculates the heat generation caused by the ohmic losses, charge
negative electrode is incorporated into the electrochemical and thermal transfer over-potentials at the interface, and mass transport limitations.
14
Fig. 11. Scheme of negative electrode degradation mechanisms (Louli et al., 2019).
Fig. 12. Main coupling phenomena.
These factors contribute to the mechanical stresses. It is illustrated at low SOCs and high temperatures. The aging and thermal effects
that the gradients of lithium concentration determine the stress level, are critical aspects of long-time battery modeling (Zhao, Liu, & Ji,
and large stresses are generated at the electrode-separator interface 2018). In general, the aging and thermal effects are coupled with the
when high current rates are applied (Ai, Kraft, Sturm, Jossen, & Wu, electrochemistry phenomenon. The combined electrochemical-aging-
2019). A coupled electrochemical-thermal-mechanical model was used thermal models are powerful tools to depict the multi-physics cells.
to investigate Li-ion battery degradation. It was found that high temper- In summary, these extended electrochemical models are promising for
ature can accelerate battery SEI growth, while low temperature would achieving effective performance under control-oriented applications
cause severe lithium-plating. Current with high rates would increase where the typical P2D model fails.
the diffusion-induced stress significantly and result in severe cathode
active material loss (Yang et al., 2019). These reduced electrochemical- 4. P2D model parameterization
thermal models could benefit the accurate cell state estimation and
optimal charge control, especially for all-climate electric vehicles. Parameter identification of electrochemical model is also important
for accurate model-based battery state estimation and optimal control
3.4. Discussion design. As the PDEs within P2D model contain many physical param-
eters, it is a key but challenging task to identify all these parameters.
This section summarizes the extended variants of the P2D model, First, battery manufacturers generally do not disclose this information
including the multi-particle models, multi-scale models, aging, and in their manufactured battery specification sheet. Second, the measur-
thermal models. These models make up for the deficiency of the P2D able signals including voltage, current, and temperature present the
model. Specifically, the multi-particle models are suitable for cathode complicated nonlinear relations with these parameters. The parameter
materials with significant phase-transition such as LFP, where the identifiability would vary for different operating conditions, further
particle interactions dominate the electrode performance. The inho- complicating the parameterization goal. Third, various characterization
mogeneity of the particle distributions can be captured by the multi- and expensive instruments are required to measure these parameters.
particle models. Thus, they are appropriate for composite electrodes, Furthermore, the parameters are specific to each cell design (including
where two or three types of active materials are mixed, and electrodes geometry and chemistry), leading to the fact that not all parameters
composed of agglomerated particles. The multi-scale models are of could be transferable from one cell design to another. One common-
great interest when the electrode microstructures are concerned. They utilized approach is to fit the parameters of electrochemical model for
are derived with asymptotic methods and could outperform P2D model matching the measured cell terminal voltage, which is sensitive to the
15
initial set of these parameters (Fan, 2020). In many related research, 4.2. Noninvasive parameter estimation
these parameter sets are generally adopted from literatures, while
their origins are seldom traced. Another typical solution is to measure On the other hand, noninvasive parameter identification methods
these parameters experimentally. In this section, two widely adopted have been investigated widely without the need of tearing down battery
methods, including direct measurements and noninvasive parameter cells, where related parameters are estimated by fitting the model
estimation, are summarized and discussed. responses to the experimentally measured voltage (Lu, Trimboli, Fan,
Zhang, & Plett, 2021b). Parameter identifiability belongs to a model
basic property to describe whether the parameters of model can be
4.1. Physical parameter measurements
identified from its output for a specific input. In general, parameters
with high sensitivity could be identified accurately with designed ex-
For the physical parameter measurements, parameters of battery periments (Baker & Verbrugge, 2018). For instance, an effective way
electrochemical model would be obtained directly through cell post- through combining parameter analysis and identification to determine
mortem analysis (Oca et al., 2021). For example,a physio-chemical a battery electrochemical model (Schmidt, Bitzer, Imre, & Guzzella,
model was fully parameterized for a high-energy pouch battery cell 2010) was presented. Here the Fisher-information matrix approach in
(Ecker, Tran, Dechent, Käbitz, Warnecke, & Sauer, 2015). Specifi- combination with a sensitivity analysis was derived to estimate the
cally, battery cell was opened under argon atmosphere for the mea- identifiability of each parameter. Then the number of experiments
surements of geometrical data. The porosity, particle radius, and tor- would be reduced to a relatively small one, while 33 electrochemical
tuosity of the electrodes and the separator were determined with parameters could be fully parameterized based on this sensitivity in-
Hg-porosimetry. Electrolyte conductivity and diffusion constants, as formation. The full set of P2D model parameters were identified from
well as electronic conductivity were detected by using the voltage cycling data based on a genetic algorithm, where the accuracy and
response to a dc current. The electrode open circuit voltage curves, dif- identifiability of the resulting full parameter set were assessed using
fusion coefficient, and charge transfer kinetics, were measured on the Fisher information (Forman, Moura, Stein, & Fathy, 2012). The most
reconstructed coin cells. The parameterized physico-electrochemical identifiable condition for each parameter was proposed based on the
model was validated against the commercial cells with same chem- sensitivity analysis and clustering analysis. Then the highly sensitive
istry (Ecker, Käbitz, Laresgoiti, & Sauer, 2015). Johannes extracted the parameters were identified under such condition (Zhang et al., 2014).
physical parameters from a high-power prismatic cell. The composition Some researchers determined the specific subsets of parameters rather
as well as the porous structure were measured using optical emission than identifying all the parameters, while other parameters were taken
spectroscopy and Hg-porosimetry. The electrochemical properties of from the literature (Marcicki et al., 2013).
electrode materials were determined using coin cells with lithium as The adopted optimization algorithms, number of estimated param-
counter electrode (Schmalstieg, Rahe, Ecker, & Sauer, 2018). Relative eters, optimization objectives, and experiments for estimating param-
validation showed that the behavior of full cell can be reproduced eters of battery electrochemical model are summarized in Table 5. As
from the parameterization of individual material parameters with small shown in Table 5, the number of estimated parameters would vary sig-
errors (Schmalstieg & Sauer, 2018). The experimental methodologies nificantly on different publications. This difference is mainly caused by
for cell teardown and subsequent chemical, physical, electrochemical the model complexity, estimation procedure, and optimization objec-
kinetics, and thermodynamic analysis was presented. The cell consists tive (e.g., whether temperatures and EIS-based frequency responses are
of a NMC positive electrode and graphite-SiOx negative electrode. The modeled). Here the most commonly used optimization algorithm is the
electrode open-circuit-voltages (OCVs) and lithium stoichiometry were gradient-based algorithm, such as the least-square algorithm (Schmidt
obtained using galvanostatic intermittent titration technique(GITT) in et al., 2010), Levenberg–Marquardt algorithm (Park et al., 2018), and
half cells. The electrochemical impedance spectroscopy (EIS) measure- Genetic algorithm (Marcicki et al., 2013). Recently, the neural network
ments were used to determine the activation energy and exchange based algorithms have been also used to estimate the parameters of
electrochemical model. For example, a deep Bayesian neural network
current coefficient (Chen et al., 2020).
was used to identify optimal parameters (Kim et al., 2019). Results
Table 4 summarizes several physical parameters of electrochemical
illustrated that the neural network based method require less calcu-
model and their related measurements. The geometric configuration
lation time than genetic algorithm, particle swarm optimization, and
of a cell, as well as the domain morphologies within the cell, are
Levenberg–Marquardt algorithm. Apart from time-domain signals of
determined by manufacturing processes. The electrode plating area,
battery such as terminal voltage and temperature, the frequency test
current collector/electrode/separator thickness are straightforward to
via EIS was incorporated in the parameterization (Chu et al., 2020). The
be understood and are relatively easy to be measured through us-
identification process was divided into four tests and eight steps, where
ing calipers or a micrometer gauge. Scanning electron microscopy
the number of parameters to be identified in each step is significantly
(SEM) is an effective tool for 2D image acquisition. These obtained reduced (Chu et al., 2020).
2D images can be then processed to extract measurements for layer
thickness, particle radius, and particle size distributions (Schmalstieg 4.3. Discussion
et al., 2018). SEM was combined with FIB to investigate the 3D mi-
crostructure of electrode, while the porosity was calculated based on This section summarizes two practical model parameterization meth-
these images (Chen et al., 2020). The tortuosity can also be calculated ods: direct parameter measurements and noninvasive parameter iden-
from these images. Mercury porosimetry is a destructive technique to tification. Each method has its strengths and weaknesses. The direct
measure porosity, pore volume, and pore size distribution. The mercury parameter measurements could provide the exact parameter values
is forced into small void spaces in the porous medium, and the volume through post-mortem analysis, which is robust and accurate. However,
of intruded mercury was recorded (Ecker, Tran et al., 2015). GITT this process is costly and time-consuming, since the cell teardown and
is a method that can measure the open circuit potential and solid- subsequent chemical, physical, electrochemical kinetics and thermody-
phase diffusion coefficients (Birkl, McTurk, Roberts, Bruce, & Howey, namic analysis should be carefully conducted. The specific instruments
2015). The activation energy and exchange current density can be for conducting these studies are expensive. For noninvasive parameter
determined by using EIS (Ecker, Käbitz et al., 2015). In terms of the estimation, this method is highly dependent on the identifiability
chemical properties, elemental composition of the active materials can and sensitivity of these parameters. The quality of training data and
be analyzed by using energy-dispersive X-ray spectroscopy (EDS) and preset parameter variation range are critical to achieving ideal results.
inductively coupled plasma optical emission spectroscopy (ICP-OES). This method indeed could save much parameterization efforts if these
16
Table 4
Physical parameters and their characterization method.
Physical parameters Characterization method
Electrode plating area Disassembly and direct measurements (Chen et al., 2020;
Ecker, Tran et al., 2015; Schmalstieg & Sauer, 2018)
Geometrical Current collector thickness Disassembly and direct measurements (Chen et al., 2020);
ICP
spectrocopy (Ecker, Tran et al., 2015); SEM images
(Schmalstieg et al., 2018)
Electrode/separator thickness Disassembly and direct measurements (Chen et al., 2020;
Ecker, Tran et al., 2015);
SEM images (Schmalstieg & Sauer, 2018)
Particle radius SEM images (Chen et al., 2020); Hg-porosimetry (Ecker, Tran
et al., 2015; Schmalstieg & Sauer, 2018)
Porosity SEM-FIB (Chen et al., 2020); Mercury porosimetry (Ecker,
Tran et al., 2015; Schmalstieg & Sauer, 2018);
liquid absorption (Venugopal, Moore, Howard, & Pendalwar,
1999); X-ray micro CT (Channagiri, Nagpure, Babu, Noble, &
Hart, 2013)
Tortuosity SEM-FIB (Chen et al., 2020); Hg-porosimetry (Ecker, Tran
et al., 2015; Schmalstieg et al., 2018)
Solid-phase lithium diffusivity GITT (Chen et al., 2020) (Schmalstieg et al., 2018) (Cabanero
et al., 2018) (Geng et al., 2021; Tang et al., 2019); EIS
(Ecker, Tran et al., 2015); CV (Tang, Lai, & Lu, 2008);
Electrochemical
PITT (Xia, Lu, & Ceder, 2006)
Solid-phase electronic conductivity Four-point probe (Chen et al., 2020)
electrolyte electronic conductivity EIS (Sauerteig et al., 2018)
Open circuit potential GITT (Bizeray et al., 2015; Chen et al., 2020); Slow charge
and
discharge (Schmalstieg & Sauer, 2018) (Verbrugge, Baker,
Koch, Xiao, & Gu, 2017) (Lu, Trimboli, Fan, Zhang, & Plett,
2021a); Three-electrode method (Chen et al., 2020)
Activation energy EIS (Chen et al., 2020); GITT& EIS (Ecker, Tran et al.,
2015); GITT (Schmalstieg et al., 2018);
Reaction rate constant EIS (Chen et al., 2020; Ecker, Tran et al., 2015); CV
(Buschmann, Berendts, Mogwitz, & Janek, 2012; Streeter,
Wildgoose, Shao, & Compton, 2008);
Active material composition EDS (Chen et al., 2020); ICP-OES (Schmalstieg et al., 2018)
Chemical
Electrolyte composition Gas chromatography (Schmalstieg et al., 2018); Ion
chromatography (Kraft et al., 2015)
requirements are met. The parameter identification results determine details of electrode microstructure and describes them as homogeneous
the efficiency of the electrochemical-model-based applications. There- media in which electrolyte and solid materials coexist at each point.
fore, the researchers should pay attention to the detailed parameter This assumption is invalid when it is necessary to capture local mi-
identification method and select the appropriate protocols. croscopic effects, especially the interaction between microstructure,
degradation, and heat. Thus, macroscopic P2D model cannot fully
5. Challenges and perspectives depict the cell dynamics at extreme operating conditions, such as low
SOCs and low temperatures. Furthermore, P2D model employs Fick’s
In this section, the challenges for tackling the limitations of clas- law to determine the lithium diffusion in active materials. For phase-
sical battery P2D model, incapability of model-order reduction tech- change materials, such as LFP, the P2D model fails to provide accurate
niques, and bottlenecks of model parameter identification are dis- predictions on intra-particle diffusions. Fast charging remains the bot-
tleneck technique to increase the willingness of consumers to purchase
cussed. Then some perspectives to further handle these issues towards
EVs. The physical model-based optimal charge protocol is promising
battery electrochemical model-based control-oriented management are
to accelerate charging speed and alleviate cell degradation (Yin &
given.
Choe, 2020). However, most of the published works on battery elec-
trochemical modeling were validated to achieve accurate predictions
5.1. Challenges
with constant discharge scenarios (Khaleghi Rahimian et al., 2013). The
evaluation of model performance with charging condition, especially at
Limitations of P2D model: Battery electrochemical model presents high C-rates is rarely-seen and worth exploring.
great promise for understanding the fundamental mechanisms of ex- The incapability of model-order reduction techniques: The rig-
isting battery materials and components, while the dynamics and per- orous P2D model should be properly simplified and tailored for control-
formance of new battery materials and components could be also oriented applications. SPM based algorithms are among the most prac-
predicted. The P2D-based battery modeling framework has become tical applications (Li, Ren et al., 2021). In general, ROMs are lin-
an ubiquitous tool to perform cell-scale simulations of Li-ion batter- earized around preset equilibrium points. For example, the Butler–
ies. This accelerates the wide deployments of electrochemical model- Volmer equation was linearized around a set-point that includes zero
based control algorithms for onboard battery state estimation, optimal reaction flux (i.e., 𝑗0 = 0) (16). However, as the input current and
charge/discharge control, and fault diagnostics. Despite the widespread reaction flux increase, ROM derived from 𝑗0 = 0 would become no
usage of battery electrochemical model, its challenges towards control- longer accurate. In this context, model performance far from equilib-
oriented battery management still remain. P2D model ignores the rium conditions need further investigation. Prior knowledge of battery
17
Table 5
Parameter estimation algorithm, number of estimated parameters, used optimization objectives, and required experiments for model identification.
References Optimization Number of Optimized Experiment
algorithm estimated objectives
parameters
Schmidt et al. (2010) Least-square 33 Voltage C-rate and pulse tests
Temperature
Forman et al. (2012) Genetic algorithm 88 Voltage Drive cycles
Edouard, Petit, Forgez, Bernard, and Tuned 41 Voltage C-rate tests
Revel (2016) Temperature
Park, Kato, Gima, Klein, and Moura Levenberg–Marquardt 21 Voltage C-rate and pulse tests
(2018) algorithm
Dangwal and Canova (2021) Particle swarm 4 Voltage C-rate and pulse tests
optimization
Kostetzer, Nebl, Stich, Bund, and Evolutionary 11 Voltage C-rate tests
Schweiger (2020) algorithm
Namor et al. (2017) Least-square 7 Voltage Pulse tests
Reddy et al. (2019) Hybrid optimization 44 Voltage C-rate and pulse tests
algorithm
Kim et al. (2019) Deep Bayesian neural 6 Voltage C-rate and pulse tests
network
Chu et al. (2020) Particle swarm 14 Voltage, C-rate and pulse tests,
optimization EIS data EIS test.
Baker, Verbrugge, and Gu (2019) Generalized reduced 20 Voltage Linear-sweep voltammetry
gradient
optimization
Li et al. (2022) Cuckoo search algorithm 26 Voltage C-rate and drive cycles
network
Lu, Trimboli, Fan, Zhang, and Plett Particle swarm 28 Voltage Pulse test
(2021c) optimization
characteristics, including battery chemistry, rate capability, and param- the interactions of multiple particles, the charge–discharge mecha-
eter sets of electrochemical models, is necessary to develop ROMs. Each nism inside active particles has not been fully understood. Besides,
model order reduction technique has its strengths and weaknesses. It typical P2D model cannot describe the microstructure of battery elec-
is an ultimate goal of control-oriented model simplification to reduce trode, while the relevant developed microscopic model solution is
the computational burden while maintaining the predictability of in- still computationally more demanding than the P2D model. In this
tegrated model. Unfortunately, no such a simplified model could meet context, a reduced-order 3D microscopic model is urgently required.
all requirements to our best knowledge. Alleviating the computational The aging effects such as lithium plating are distributed within an
burden while maintaining the comprehensive model predictability is inhomogeneous electrode structure. Lithium electroplating would cause
an ultimate goal for control-oriented model simplification, where none the growth of lithium dendrites, which presents the risk of causing
such simplified model could solve all current issues. thermal runaway. This demands the development of coupled elec-
Model parameter identification bottlenecks: The parameter iden- trochemical models allowing the combination of local temperature,
tification of battery electrochemical model lays the foundation for the concentration fluctuations, and aging processes. Similarly, the coupling
development of model-based control algorithms. However, as some of stress development from both the particle and electrode levels needs
parameters are difficult to be obtained through battery disassembly to be considered. For battery pack applications, the inevitable cell
analysis, it becomes challenging to obtain a suitable set of parameters inconsistencies would lead to electrical imbalances. In this context,
for battery electrochemical model. Currently, many parameters are still deriving efficient pack-level electrochemical model considering cell
difficult to be fully identified from direct current and voltage measure- inconsistency becomes necessary.
ments (Bizeray, Kim, Duncan, & Howey, 2019). On the other hand, Model parameterization: Model parameterization is still a bottle-
when using computational methods to estimate the model parameters, neck for wider applications of battery electrochemical model in control-
oriented battery management. The consistency of assumptions for many
there are always some redundant parameters that cannot be identified
parameters is still unclear. For example, numerous assumptions are
directly. In this context, the lumped parameter reformulated models
typically made to parameterize a model (e.g., whether insertion mate-
that minimize the set of parameters to be estimated by normalizing the
rials are treated as spherical or planar for solid-state diffusivity inferred
geometric coordinates and grouping parameters that always appear to-
by EIS). Identifiability issue further complicates this parameterization
gether on the PDEs are recommended (Chu, Plett, Trimboli, & Ouyang,
challenge. In this context, more robust and efficient parameter identifi-
2019).
cation methods are required (Ali, Mukhopadhyay, Rehman, & Khurram,
2017).
5.2. Perspectives
6. Conclusions
The existing research rarely brings significant improvements on the
aforementioned aspects. To further improve the performance of battery With the rapid development of battery electrochemical models,
control-oriented electrochemical model so as to promote the develop- the advanced electrochemical model-based BMS could achieve safe,
ment of next-generation BMS, the following aspects are suggested to be efficient, and reliable control strategies for onboard Li-ion batteries.
considered in future research, as illustrated in Fig. 13. As a classical type of electrochemical models, the P2D model is dif-
Model extension and improvements: For cases of great practi- ficult to be adopted directly for control-oriented applications such as
cal interest, such as LFP active material in Li-ion batteries, due to cell state estimation, charging control, and fault diagnostics. Model
18
Fig. 13. Future trends of control-oriented modeling.
simplification is thus crucial for converting P2D model governed by 19500712300), 111 Project (No. D18003), the International Corpo-
numerous PDEs to low-order linear models for real control applications. ration Project of Shanghai Science and Technology Commission un-
In this review, different P2D model-order reduction methods are first der Grant 21190780300, and the High Value Manufacturing Catapult
surveyed, where their advantages and shortcomings are discussed and project under grant No. 8248 CORE.
compared. The survey has shown that the SPM-based protocols are
the most popular approach for P2D model simplification. After ana- References
lyzing the limitations of P2D model, enhanced solutions including the
multi-particle model, multi-scale model, model with thermal and aging Ai, W., Kraft, L., Sturm, J., Jossen, A., & Wu, B. (2019). Electrochemical thermal-
elements are reviewed. Then the corresponding reduced versions of mechanical modelling of stress inhomogeneity in Lithium-Ion pouch cells. Journal
these enhanced electrochemical models are summarized as well to fur- of The Electrochemical Society, 167(1), http://dx.doi.org/10.1149/2.0122001jes.
ther benefit battery control-oriented management. As accurate model Albertus, P., Christensen, J., & Newman, J. (2009). Experiments on and modeling of
positive electrodes with multiple active materials for lithium-ion batteries. Journal
parameterization is also the key to the effective operation and control of
of The Electrochemical Society, 156(7), A606.
Li-ion batteries, model parameterization methods including the direct
Ali, D., Mukhopadhyay, S., Rehman, H., & Khurram, A. (2017). UAS based Li-ion battery
parameter measurements and invasive parameter estimation are also model parameters estimation. Control Engineering Practice, 66, 126–145.
summarized and compared. Based on these discussions, challenges, and Anseán, D., Dubarry, M., Devie, A., Liaw, B. Y., García, V. M., Viera, J. C., et al.
perspectives on the development of control-oriented electrochemical (2016). Fast charging technique for high power LiFePO 4 batteries: A mechanistic
models are finally presented. analysis of aging. Journal of Power Sources, 321, 201–209. http://dx.doi.org/10.
1016/j.jpowsour.2016.04.140.
In a nutshell, reliable control-oriented management of battery elec-
Arunachalam, H., & Onori, S. (2019). Full homogenized macroscale model and pseudo-
trochemical model requires to be implemented for online application, 2-dimensional model for Lithium-Ion battery dynamics: Comparative analysis,
but many corresponding technologies are immature. None of the elec- experimental verification and sensitivity analysis. Journal of The Electrochemical
trochemical modeling solution is a one-size-fits-all strategy; instead Society, 166(8), A1380–A1392. http://dx.doi.org/10.1149/2.0051908jes.
there should be an inherent trade-off between the model complexity Ashwin, T. R., Chung, Y. M., & Wang, J. (2016). Capacity fade modelling of lithium-ion
battery under cyclic loading conditions. Journal of Power Sources, 328, 586–598.
and performance. This review would provide useful reference points to
http://dx.doi.org/10.1016/j.jpowsour.2016.08.054.
support the design and operation of battery electrochemical model and Ashwin, T. R., McGordon, A., & Jennings, P. A. (2017). Electrochemical modelling of
parameterization, whilst also informing the agenda of the battery and Li-ion battery pack with constant voltage cycling. Journal of Power Sources, 341,
control engineering communities. 327–339. http://dx.doi.org/10.1016/j.jpowsour.2016.11.092.
Avdeev, M. V., Rulev, A. A., Bodnarchuk, V. I., Ushakova, E. E., Petrenko, V. I.,
Gapon, I. V., et al. (2017). Monitoring of lithium plating by neutron reflectometry.
Declaration of competing interest
Applications of Surface Science, 424, 378–382. http://dx.doi.org/10.1016/j.apsusc.
2017.01.290.
The authors declare that they have no known competing finan- Baba, N., Yoshida, H., Nagaoka, M., Okuda, C., & Kawauchi, S. (2014). Numerical
cial interests or personal relationships that could have appeared to simulation of thermal behavior of lithium-ion secondary batteries using the
enhanced single particle model. Journal of Power Sources, 252, 214–228. http:
influence the work reported in this paper.
//dx.doi.org/10.1016/j.jpowsour.2013.11.111.
Baker, D. R., & Verbrugge, M. W. (2018). Multi-species, multi-reaction model for porous
Acknowledgments intercalation electrodes: Part I. Model formulation and a perturbation solution
for low-scan-rate, linear-sweep voltammetry of a spinel lithium manganese oxide
electrode. Journal of The Electrochemical Society, 165(16), A3952.
This work was financially supported by the EPSRC under grant
Baker, D. R., Verbrugge, M. W., & Gu, W. (2019). Multi-species, multi-reaction model
EP/R030243/1, the National Natural Science Foundation of China for porous intercalation electrodes: Part II. Model-experiment comparisons for
(NSFC) under Grant number 62173218, 52007119, the Key Project of linear-sweep voltammetry of spinel lithium manganese oxide electrodes. Journal
Science and Technology Commission of Shanghai Municipality (No. of The Electrochemical Society, 166(4), A521.
19
Bartlett, A., Marcicki, J., Onori, S., Rizzoni, G., Yang, X. G., & Miller, T. (2016). Doyle, M., Newman, J., Gozdz, A. S., Schmutz, C. N., & Tarascon, J.-M. (1996).
Electrochemical model-based state of charge and capacity estimation for a compos- Comparison of modeling predictions with experimental data from plastic lithium
ite electrode Lithium-Ion battery. IEEE Transactions on Control Systems Technology, ion cells. Journal of The Electrochemical Society, 143(6), 1890.
24(2), 384–399. http://dx.doi.org/10.1109/TCST.2015.2446947. Dreyer, W., Jamnik, J., Guhlke, C., Huth, R., Moskon, J., & Gaberscek, M. (2010).
Bi, Y., & Choe, S.-Y. (2020). An adaptive sigma-point Kalman filter with state equality The thermodynamic origin of hysteresis in insertion batteries. Nature Materials,
constraints for online state-of-charge estimation of a Li(NiMnCo)O2/Carbon battery 9(5), 448–453. http://dx.doi.org/10.1038/nmat2730, URL https://www.ncbi.nlm.
using a reduced-order electrochemical model. Applied Energy, 258, http://dx.doi. nih.gov/pubmed/20383130.
org/10.1016/j.apenergy.2019.113925. Eapen, D. E., Suresh, R., Patil, S., & Rengaswamy, R. (2021). A systems engineering
Birkl, C. R., McTurk, E., Roberts, M. R., Bruce, P. G., & Howey, D. A. (2015). perspective on electrochemical energy technologies and a framework for application
A parametric open circuit voltage model for Lithium Ion batteries. Journal of driven choice of technology. Renewable and Sustainable Energy Reviews, 147, http:
The Electrochemical Society, 162(12), A2271–A2280. http://dx.doi.org/10.1149/2. //dx.doi.org/10.1016/j.rser.2021.111165.
0331512jes. Ecker, M., Käbitz, S., Laresgoiti, I., & Sauer, D. U. (2015). Parameterization of a
Bistri, D., & Di Leo, C. V. (2021). Modeling of chemo-mechanical multi-particle physico-chemical model of a lithium-ion battery: II. Model validation. Journal of
interactions in composite electrodes for liquid and solid-state Li-Ion batteries. The Electrochemical Society, 162(9), A1849.
Journal of The Electrochemical Society, 168(3), http://dx.doi.org/10.1149/1945- Ecker, M., Tran, T. K. D., Dechent, P., Käbitz, S., Warnecke, A., & Sauer, D. U.
7111/abe8ea. (2015). Parameterization of a physico-chemical model of a lithium-ion battery: I.
Bizeray, A. M., Kim, J.-H., Duncan, S. R., & Howey, D. A. (2019). Identifiability Determination of parameters. Journal of The Electrochemical Society, 162(9), A1836.
and parameter estimation of the single particle Lithium-Ion battery model. IEEE Edouard, C., Petit, M., Forgez, C., Bernard, J., & Revel, R. (2016). Parameter sensitivity
Transactions on Control Systems Technology, 27(5), 1862–1877. http://dx.doi.org/ analysis of a simplified electrochemical and thermal model for Li-ion batteries ag-
10.1109/tcst.2018.2838097. ing. Journal of Power Sources, 325, 482–494. http://dx.doi.org/10.1016/j.jpowsour.
Bizeray, A. M., Zhao, S., Duncan, S. R., & Howey, D. A. (2015). Lithium-ion battery 2016.06.030.
thermal-electrochemical model-based state estimation using orthogonal collocation Fan, G. (2020). Systematic parameter identification of a control-oriented electro-
and a modified extended Kalman filter. Journal of Power Sources, 296, 400–412. chemical battery model and its application for state of charge estimation at
http://dx.doi.org/10.1016/j.jpowsour.2015.07.019. various operating conditions. Journal of Power Sources, http://dx.doi.org/10.1016/
Brosa Planella, F., Sheikh, M., & Widanage, W. D. (2021). Systematic derivation and j.jpowsour.2020.228153.
validation of a reduced thermal-electrochemical model for lithium-ion batteries Fang, R., Ge, H., Wang, Z., Li, Z., & Zhang, J. (2020). A two-dimensional heterogeneous
using asymptotic methods. Electrochimica Acta, 388, http://dx.doi.org/10.1016/j. model of Lithium-Ion battery and application on designing electrode with non-
electacta.2021.138524. uniform porosity. Journal of The Electrochemical Society, 167(13), http://dx.doi.org/
Buschmann, H., Berendts, S., Mogwitz, B., & Janek, J. (2012). Lithium metal electrode 10.1149/1945-7111/abb83a.
kinetics and ionic conductivity of the solid lithium ion conductors ‘‘Li7La3Zr2O12’’ Fang, H., Wang, Y., Sahinoglu, Z., Wada, T., & Hara, S. (2014). State of charge
and li7?xLa3Zr2?xtaxo12 with garnet-type structure. Journal of Power Sources, 206, estimation for lithium-ion batteries: An adaptive approach. Control Engineering
236–244. http://dx.doi.org/10.1016/j.jpowsour.2012.01.094. Practice, 25, 45–54. http://dx.doi.org/10.1016/j.conengprac.2013.12.006.
Farag, M., Fleckenstein, M., & Habibi, S. (2017). Continuous piecewise-linear, reduced-
Cabanero, M. A., Boaretto, N., Roeder, M., Mueller, J., Kallo, J., & Latz, A. (2018).
order electrochemical model for lithium-ion batteries in real-time applications.
Direct determination of diffusion coefficients in commercial Li-ion batteries. Journal
Journal of Power Sources, 342, 351–362. http://dx.doi.org/10.1016/j.jpowsour.
of The Electrochemical Society, 165(5), A847.
2016.12.044.
Carnevale, D., Possieri, C., & Sassano, M. (2019). State-of-charge estimation for lead–
Farkhondeh, M., & Delacourt, C. (2011). Mathematical modeling of commer-
acid batteries via embeddings and observers. Control Engineering Practice, 85,
cial LiFePO4Electrodes based on variable solid-state diffusivity. Journal of
132–137.
The Electrochemical Society, 159(2), A177–A192. http://dx.doi.org/10.1149/2.
Channagiri, S. A., Nagpure, S. C., Babu, S. S., Noble, G. J., & Hart, R. T. (2013).
073202jes.
Porosity and phase fraction evolution with aging in lithium iron phosphate battery
Feng, F., Teng, S., Liu, K., Xie, J., Xie, Y., Liu, B., et al. (2020). Co-estimation
cathodes. Journal of Power Sources, 243, 750–757. http://dx.doi.org/10.1016/j.
of lithium-ion battery state of charge and state of temperature based on a
jpowsour.2013.06.023.
hybrid electrochemical-thermal-neural-network model. Journal of Power Sources,
Chen, C.-H., Brosa Planella, F., O’Regan, K., Gastol, D., Widanage, W. D., & Kendrick, E.
455, Article 227935.
(2020). Development of experimental techniques for parameterization of multi-
Ferro, G., Laureri, F., Minciardi, R., & Robba, M. (2019). A predictive discrete event
scale Lithium-ion battery models. Journal of The Electrochemical Society, 167(8),
approach for the optimal charging of electric vehicles in microgrids. Control
http://dx.doi.org/10.1149/1945-7111/ab9050.
Engineering Practice, 86, 11–23.
Chen, G., Liu, Z., Su, H., & Zhuang, W. (2021). Electrochemical-distributed thermal
Fleischer, C., Waag, W., Heyn, H.-M., & Sauer, D. U. (2014). On-line adaptive
coupled model-based state of charge estimation for cylindrical lithium-ion batteries.
battery impedance parameter and state estimation considering physical principles
Control Engineering Practice, 109, Article 104734.
in reduced order equivalent circuit battery models. Journal of Power Sources, 260,
Chu, Z., Feng, X., Lu, L., Li, J., Han, X., & Ouyang, M. (2017). Non-destructive
276–291. http://dx.doi.org/10.1016/j.jpowsour.2014.01.129.
fast charging algorithm of lithium-ion batteries based on the control-oriented Forman, J. C., Moura, S. J., Stein, J. L., & Fathy, H. K. (2012). Genetic identification and
electrochemical model. Applied Energy, 204, 1240–1250. http://dx.doi.org/10.1016/ fisher identifiability analysis of the Doyle–Fuller–Newman model from experimental
j.apenergy.2017.03.111. cycling of a LiFePO4 cell. Journal of Power Sources, 210, 263–275. http://dx.doi.
Chu, Z., Jobman, R., Rodríguez, A., Plett, G. L., Trimboli, M. S., Feng, X., et al. org/10.1016/j.jpowsour.2012.03.009.
(2020). A control-oriented electrochemical model for lithium-ion battery. Part II: Forouzan, M. M., Mazzeo, B. A., & Wheeler, D. R. (2018). Modeling the effects of
Parameter identification based on reference electrode. Journal of Energy Storage, 27, electrode microstructural heterogeneities on li-ion battery performance and lifetime.
http://dx.doi.org/10.1016/j.est.2019.101101. Journal of The Electrochemical Society, 165(10), A2127–A2144. http://dx.doi.org/10.
Chu, Z., Plett, G. L., Trimboli, M. S., & Ouyang, M. (2019). A control-oriented 1149/2.1281809jes.
electrochemical model for lithium-ion battery, Part I: Lumped-parameter reduced- Franco, A. A. (2013). Multiscale modelling and numerical simulation of rechargeable
order model with constant phase element. Journal of Energy Storage, 25, Article lithium ion batteries: concepts, methods and challenges. RSC Advances, 3(32),
100828. http://dx.doi.org/10.1039/c3ra23502e.
Dangwal, C., & Canova, M. (2021). Parameter identification for electrochemical models Gambhire, P., Ganesan, N., Basu, S., Hariharan, K. S., Kolake, S. M., Song, T., et
of Lithium Ion batteries using sensitivity analysis. ASME Letters in Dynamic Systems al. (2015). A reduced order electrochemical thermal model for lithium ion cells.
and Control, 1–33. http://dx.doi.org/10.1115/1.4050794. Journal of Power Sources, 290, 87–101. http://dx.doi.org/10.1016/j.jpowsour.2015.
Deng, Z., Yang, L., Deng, H., Cai, Y., & Li, D. (2018). Polynomial approximation 04.179.
pseudo-two-dimensional battery model for online application in embedded battery Gao, Y., Liu, K., Zhu, C., Zhang, X., & Zhang, D. (2021). Co-estimation of state-of-charge
management system. Energy, 142, 838–850. http://dx.doi.org/10.1016/j.energy. and state-of- health for lithium-ion batteries using an enhanced electrochemical
2017.10.097. model. IEEE Transactions on Industrial Electronics, 1. http://dx.doi.org/10.1109/TIE.
Dey, S., Ayalew, B., & Pisu, P. (2015). Nonlinear robust observers for state-of-charge 2021.3066946.
estimation of Lithium-Ion cells based on a reduced electrochemical model. IEEE Gao, Y., Zhang, X., Guo, B., Zhu, C., Wiedemann, J., Wang, L., et al. (2020). Health-
Transactions on Control Systems Technology, 23(5), 1935–1942. http://dx.doi.org/ aware multiobjective optimal charging strategy with coupled electrochemical-
10.1109/tcst.2014.2382635. thermal-aging model for Lithium-Ion battery. IEEE Transactions on Industrial
Dickinson, E. J., Ekström, H., & Fontes, E. (2014). COMSOL multiphysics® : Finite Informatics, 16(5), 3417–3429. http://dx.doi.org/10.1109/tii.2019.2935326.
element software for electrochemical analysis. a mini-review. Energy Conversion and Geng, Z., Wang, S., Lacey, M. J., Brandell, D., & Thiringer, T. (2021). Bridging physics-
Management, 40, 71–74. based and equivalent circuit models for lithium-ion batteries. Electrochimica Acta,
Dos Reis, G., Strange, C., Yadav, M., & Li, S. (2021). Lithium-ion battery data and 372, http://dx.doi.org/10.1016/j.electacta.2021.137829.
where to find it. Energy and AI, 5, Article 100081. Gopalakrishnan, K., Zhang, T., & Offer, G. J. (2017). A fast, memory-efficient discrete-
Doyle, M., Fuller, T. F., & Newman, J. (1993). Modeling of galvanostatic charge time realization algorithm for reduced-order Li-Ion battery models. Journal of
and discharge of the lithium/polymer/insertion cell. Journal of The Electrochemical Electrochemical Energy Conversion and Storage, 14(1), http://dx.doi.org/10.1115/1.
Society, 140(6), 1526. 4035526.
20
Gully, A., Liu, H., Srinivasan, S., Sethurajan, A. K., Schougaard, S., & Protas, B. (2014). Kostetzer, L., Nebl, C., Stich, M., Bund, A., & Schweiger, H.-G. (2020). Physics-
Effective transport properties of porous electrochemical materials — a homoge- based modeling and parameter identification for Lithium Ion batteries under high
nization approach. Journal of The Electrochemical Society, 161(8), E3066–E3077. current discharge conditions. Journal of The Electrochemical Society, 167(14), http:
http://dx.doi.org/10.1149/2.011408jes. //dx.doi.org/10.1149/1945-7111/abc726.
Guo, G., Long, B., Cheng, B., Zhou, S., Xu, P., & Cao, B. (2010). Three-dimensional Kraft, V., Grutzke, M., Weber, W., Menzel, J., Wiemers-Meyer, S., Winter, M.,
thermal finite element modeling of lithium-ion battery in thermal abuse application. et al. (2015). Two-dimensional ion chromatography for the separation
Journal of Power Sources, 195(8), 2393–2398. http://dx.doi.org/10.1016/j.jpowsour. of ionic organophosphates generated in thermally decomposed lithium
2009.10.090. hexafluorophosphate-based lithium ion battery electrolytes. Journal of
Guo, M., Sikha, G., & White, R. E. (2011). Single-particle model for a Lithium- Chromatography (Amsterdam), 1409, 9–201. http://dx.doi.org/10.1016/j.chroma.
Ion cell: Thermal behavior. Journal of The Electrochemical Society, 158(2), A122. 2015.07.054, URL https://www.ncbi.nlm.nih.gov/pubmed/26209196.
http://dx.doi.org/10.1149/1.3521314. Kupper, C., & Bessler, W. G. (2016). Multi-scale thermo-electrochemical model-
Guo, M., & White, R. E. (2012). An approximate solution for solid-phase diffusion in ing of performance and aging of a LiFePO4/Graphite Lithium-Ion cell. Journal
a spherical particle in physics-based li-ion cell models. Journal of Power Sources, of The Electrochemical Society, 164(2), A304–A320. http://dx.doi.org/10.1149/2.
198, 322–328. http://dx.doi.org/10.1016/j.jpowsour.2011.08.096. 0761702jes.
Han, X., Ouyang, M., Lu, L., & Li, J. (2015). Simplification of physics-based elec- Lamorgese, A., Mauri, R., & Tellini, B. (2018). Electrochemical-thermal P2D aging
trochemical model for lithium ion battery on electric vehicle. Part I: Diffusion model of a LiCoO2/graphite cell: Capacity fade simulations. Journal of Energy
simplification and single particle model. Journal of Power Sources, 278, 802–813. Storage, 20, 289–297. http://dx.doi.org/10.1016/j.est.2018.08.011.
http://dx.doi.org/10.1016/j.jpowsour.2014.12.101. Lee, J. L., Chemistruck, A., & Plett, G. L. (2012a). Discrete-time realization of
transcendental impedance models, with application to modeling spherical solid
He, L., Ye, W., He, Z., Song, K., & Shi, Q. (2020). A combining sliding mode control
diffusion. Journal of Power Sources, 206, 367–377. http://dx.doi.org/10.1016/j.
approach for electric motor anti-lock braking system of battery electric vehicle.
jpowsour.2012.01.134.
Lee, J. L., Chemistruck, A., & Plett, G. L. (2012b). One-dimensional physics-based
Hein, S., Feinauer, J., Westhoff, D., Manke, I., Schmidt, V., & Latz, A. (2016).
reduced-order model of lithium-ion dynamics. Journal of Power Sources, 220,
Stochastic microstructure modeling and electrochemical simulation of lithium-ion
430–448. http://dx.doi.org/10.1016/j.jpowsour.2012.07.075.
cell anodes in 3D. Journal of Power Sources, 336, 161–171. http://dx.doi.org/10.
Lekshmi, S., & PS, L. P. (2019). Mathematical modeling of electric vehicles-a survey.
1016/j.jpowsour.2016.10.057.
Hennessy, M. G., & Moyles, I. R. (2020). Asymptotic reduction and homogenization Li, J., Adewuyi, K., Lotfi, N., Landers, R. G., & Park, J. (2018). A single particle
of a thermo-electrochemical model for a lithium-ion battery. Applied Mathematical model with chemical/mechanical degradation physics for lithium ion battery state
Modelling, 80, 724–754. http://dx.doi.org/10.1016/j.apm.2019.11.018. of health (SOH) estimation. Applied Energy, 212, 1178–1190. http://dx.doi.org/10.
Hu, X., Cao, D., & Egardt, B. (2018). Condition monitoring in advanced battery 1016/j.apenergy.2018.01.011.
management systems: Moving horizon estimation using a reduced electrochemical Li, X., Choe, S.-Y., & Joe, W. T. (2015). A reduced order electrochemical and thermal
model. IEEE/ASME Transactions on Mechatronics, 23(1), 167–178. http://dx.doi.org/ model for a pouch type lithium ion polymer battery with LiNixMnyCo1- x-
10.1109/tmech.2017.2675920. yO2/LiFePO4 blended cathode. Journal of Power Sources, 294, 545–555.
Hu, X., Feng, F., Liu, K., Zhang, L., Xie, J., & Liu, B. (2019). State estimation for Li, C., Cui, N., Wang, C., & Zhang, C. (2021). Reduced-order electrochemical model
advanced battery management: Key challenges and future trends. Renewable and for lithium-ion battery with domain decomposition and polynomial approximation
Sustainable Energy Reviews, 114, http://dx.doi.org/10.1016/j.rser.2019.109334. methods. Energy, 221, http://dx.doi.org/10.1016/j.energy.2020.119662.
Huang, J., Ge, H., Li, Z., & Zhang, J. (2014). An agglomerate model for the impedance Li, W., Demir, I., Cao, D., Jöst, D., Ringbeck, F., Junker, M., et al. (2022). Data-driven
of secondary particle in Lithium-Ion battery electrode. Journal of The Electrochemical systematic parameter identification of an electrochemical model for lithium-ion
Society, 161(8), E3202–E3215. http://dx.doi.org/10.1149/2.027408jes. batteries with artificial intelligence. Energy Storage Materials, 44, 557–570.
Jiang, G., Zhuang, L., Hu, Q., Liu, Z., & Huang, J. (2020). An investigation of heat Li, W., Fan, Y., Ringbeck, F., J?st, D., Han, X., Ouyang, M., et al. (2020). Electrochem-
transfer and capacity fade in a prismatic Li-ion battery based on an electrochemical- ical model-based state estimation for lithium-ion batteries with adaptive unscented
thermal coupling model. Applied Thermal Engineering, 171, http://dx.doi.org/10. Kalman filter. Journal of Power Sources, 476, http://dx.doi.org/10.1016/j.jpowsour.
1016/j.applthermaleng.2020.115080. 2020.228534.
Jin, X., Vora, A., Hoshing, V., Saha, T., Shaver, G., Wasynczuk, O., et al. (2018). Li, P., Jiao, X., & Li, Y. (2021). Adaptive real-time energy management control
Applicability of available li-ion battery degradation models for system and control strategy based on fuzzy inference system for plug-in hybrid electric vehicles. Control
algorithm design. Control Engineering Practice, 71, 1–9. Engineering Practice, 107, Article 104703.
Jokar, A., Rajabloo, B., Désilets, M., & Lacroix, M. (2016). Review of simplified pseudo- Li, J., Landers, R. G., & Park, J. (2020). A comprehensive single-particle-degradation
two-dimensional models of lithium-ion batteries. Journal of Power Sources, 327, model for battery state-of-health prediction. Journal of Power Sources, 456, http:
44–55. http://dx.doi.org/10.1016/j.jpowsour.2016.07.036. //dx.doi.org/10.1016/j.jpowsour.2020.227950.
Khaleghi Rahimian, S., Rayman, S., & White, R. E. (2013). Extension of physics-based Li, Y., Li, K., Liu, X., Wang, Y., & Zhang, L. (2021). Lithium-ion battery capacity
single particle model for higher charge–discharge rates. Journal of Power Sources, estimation—A pruned convolutional neural network approach assisted with transfer
224, 180–194. http://dx.doi.org/10.1016/j.jpowsour.2012.09.084. learning. Applied Energy, 285, Article 116410.
Khalik, Z., Donkers, M. C. F., & Bergveld, H. J. (2021). Model simplifications and Li, S., Li, K., Xiao, E., & Wong, C.-K. (2019). Joint SoC and SoH estimation for
their impact on computational complexity for an electrochemistry-based battery Zinc–Nickel single-flow batteries. IEEE Transactions on Industrial Electronics, 67(10),
modeling toolbox. Journal of Power Sources, 488, http://dx.doi.org/10.1016/j. 8484–8494.
Li, G., & Monroe, C. W. (2020). Multiscale Lithium-battery modeling from materials to
jpowsour.2020.229427.
cells. Annual Review of Chemical and Biomolecular Engineering, 11, 277–310. http://
Khalik, Z., Donkers, M. C. F., Sturm, J., & Bergveld, H. J. (2021). Parameter estima-
dx.doi.org/10.1146/annurev-chembioeng-012120-083016, URL https://www.ncbi.
tion of the Doyle–Fuller–Newman model for Lithium-ion batteries by parameter
nlm.nih.gov/pubmed/32212821.
normalization, grouping, and sensitivity analysis. Journal of Power Sources, 499,
Li, Y., Ralahamilage, D., Vilathgamuwa, M., Mishra, Y., Farrell, T., Choi, S. S., et al.
http://dx.doi.org/10.1016/j.jpowsour.2021.229901.
(2021). Model order reduction techniques for physics-based Lithium-Ion battery
Kim, M., Chun, H., Kim, J., Kim, K., Yu, J., Kim, T., et al. (2019). Data-efficient parame-
management: A survey. IEEE Industrial Electronics Magazine, 2–18. http://dx.doi.
ter identification of electrochemical lithium-ion battery model using deep Bayesian
org/10.1109/mie.2021.3100318.
harmony search. Applied Energy, 254, http://dx.doi.org/10.1016/j.apenergy.2019.
Li, L., Ren, Y., O’Regan, K., Koleti, U. R., Kendrick, E., Widanage, W. D., et al. (2021).
113644.
Lithium-ion battery cathode and anode potential observer based on reduced-order
Kirk, T. L., Please, C. P., & Jon Chapman, S. (2021). Physical modelling of the electrochemical single particle model. Journal of Energy Storage, 44, http://dx.doi.
slow voltage relaxation phenomenon in Lithium-Ion batteries. Journal of The org/10.1016/j.est.2021.103324.
Electrochemical Society, 168(6), http://dx.doi.org/10.1149/1945-7111/ac0bf7. Li, Y., Vilathgamuwa, M., Farrell, T., Choi, S. S., Tran, N. T., & Teague, J. (2019).
Klintberg, A., Zou, C., Fridholm, B., & Wik, T. (2019). Kalman filter for adaptive A physics-based distributed-parameter equivalent circuit model for lithium-ion
learning of two-dimensional look-up tables applied to OCV-curves for aged battery batteries. Electrochimica Acta, 299, 451–469. http://dx.doi.org/10.1016/j.electacta.
cells. Control Engineering Practice, 84, 230–237. 2018.12.167.
Koga, S., Camacho-Solorio, L., & Krstic, M. (2021). State estimation for Lithium-Ion Li, Y., Vilathgamuwa, M., Wikner, E., Wei, Z., Zhang, X., Thiringer, T., et al. (2021).
batteries with phase transition materials via boundary observers. Journal of Dynamic Electrochemical model-based fast charging: Physical constraint-triggered PI control.
Systems, Measurement, and Control, 143(4), http://dx.doi.org/10.1115/1.4048779. IEEE Transactions on Energy Conversion, 1. http://dx.doi.org/10.1109/tec.2021.
Koltypin, M., Aurbach, D., Nazar, L., & Ellis, B. (2007). More on the performance 3065983.
of LiFePO4 electrodes—The effect of synthesis route, solution composition, aging, Li, D., Yang, L., & Li, C. (2021). Control-oriented thermal-electrochemical modeling
and temperature. Journal of Power Sources, 174(2), 1241–1250. http://dx.doi.org/ and validation of large size prismatic lithium battery for commercial applications.
10.1016/j.jpowsour.2007.06.045. Energy, 214, http://dx.doi.org/10.1016/j.energy.2020.119057.
Kong, X., Wetton, B., & Gopaluni, B. (2019). Assessment of simplifications to a Lin, X., Hao, X., Liu, Z., & Jia, W. (2018). Health conscious fast charging of Li-ion
pseudo–2D electrochemical model of Li-ion batteries. IFAC-PapersOnLine, 52(1), batteries via a single particle model with aging mechanisms. Journal of Power
946–951. Sources, 400, 305–316. http://dx.doi.org/10.1016/j.jpowsour.2018.08.030.
21
Lin, X., Wang, S., & Kim, Y. (2019). A framework for charging strategy optimization Moyles, I. R., Hennessy, M. G., Myers, T. G., & Wetton, B. R. (2019). Asymptotic
using a physics-based battery model. Journal of Applied Electrochemistry, 49(8), reduction of a porous electrode model for Lithium-Ion batteries. SIAM Journal of
779–793. http://dx.doi.org/10.1007/s10800-019-01322-1. Applied Mathematics, 79(4), 1528–1549. http://dx.doi.org/10.1137/18m1189579.
Liu, K., Li, K., Peng, Q., & Zhang, C. (2019). A brief review on key technologies in the Muddappa, V. K., & Anwar, S. (2014). Electrochemical model based fault diagnosis
battery management system of electric vehicles. Frontiers of Mechanical Engineering, of li-ion battery using fuzzy logic. In ASME international mechanical engineering
14(1), 47–64. congress and exposition, Vol. 46483. American Society of Mechanical Engineers,
Liu, K., Li, K., & Zhang, C. (2017). Constrained generalized predictive control of battery Article V04BT04A048.
charging process based on a coupled thermoelectric model. Journal of Power Sources, Namor, E., Torregrossa, D., Cherkaoui, R., & Paolone, M. (2017). Parameter identifica-
347, 145–158. tion of a lithium-ion cell single-particle model through non-invasive testing. Journal
Liu, B., Tang, X., & Gao, F. (2020). Joint estimation of battery state-of-charge and of Energy Storage, 12, 138–148. http://dx.doi.org/10.1016/j.est.2017.04.008.
state-of-health based on a simplified pseudo-two-dimensional model. Electrochimica Ning, G., & Popov, B. N. (2004). Cycle life modeling of lithium-ion batteries. Journal
Acta, 344, http://dx.doi.org/10.1016/j.electacta.2020.136098. of The Electrochemical Society, 151(10), A1584.
López C, D. C., Wozny, G., Flores-Tlacuahuac, A., Vasquez-Medrano, R., & Zavala, V. M. Oca, L., Miguel, E., Agirrezabala, E., Herran, A., Gucciardi, E., Otaegui, L., et al. (2021).
(2016). A computational framework for identifiability and ill-conditioning analysis Physico-chemical parameter measurement and model response evaluation for a
of Lithium-Ion battery models. Industrial and Engineering Chemistry Research, 55(11), pseudo-two-dimensional model of a commercial lithium-ion battery. Electrochimica
3026–3042. http://dx.doi.org/10.1021/acs.iecr.5b03910. Acta, 382, Article 138287.
Lotfi, N., Landers, R. G., Li, J., & Park, J. (2017). Reduced-order electrochemical model- Ortiz-Ricardez, F. A., Romero-Becerril, A., & Alvarez-Icaza, L. (2021). Residue grouping
based SOC observer with output model uncertainty estimation. IEEE Transactions order reduction method in solid-phase lithium-ion battery models. Journal of Applied
on Control Systems Technology, 25(4), 1217–1230. http://dx.doi.org/10.1109/tcst. Electrochemistry, http://dx.doi.org/10.1007/s10800-021-01599-1.
2016.2598764. Ouyang, Q., Wang, Z., Liu, K., Xu, G., & Li, Y. (2019). Optimal charging control
Louli, A. J., Ellis, L. D., & Dahn, J. R. (2019). Operando pressure measurements reveal for lithium-ion battery packs: A distributed average tracking approach. IEEE
solid electrolyte interphase growth to rank Li-Ion cell performance. Joule, 3(3), Transactions on Industrial Informatics, 16(5), 3430–3438.
745–761. http://dx.doi.org/10.1016/j.joule.2018.12.009. Park, S., Kato, D., Gima, Z., Klein, R., & Moura, S. (2018). Optimal experimental design
Lu, D., Trimboli, M. S., Fan, G., Zhang, R., & Plett, G. L. (2021a). Implementation of for parameterization of an electrochemical Lithium-Ion battery model. Journal
a physics-based model for half-cell open-circuit potential and full-cell open-circuit of The Electrochemical Society, 165(7), A1309–A1323. http://dx.doi.org/10.1149/
voltage estimates: part i. processing half-cell data. Journal of The Electrochemical 2.0421807jes.
Society, 168(7), Article 070532.
Pereira, D. J., Fernandez, M. A., Streng, K. C., Hou, X. X., Gao, X., Weidner, J. W., et al.
Lu, D., Trimboli, M. S., Fan, G., Zhang, R., & Plett, G. L. (2021b). Implementation of (2020). Accounting for non-ideal, lithiation-based active material volume change
a physics-based model for half-cell open-circuit potential and full-cell open-circuit in mechano-electrochemical pouch cell simulation. Journal of The Electrochemical
voltage estimates: part ii. processing full-cell data. Journal of The Electrochemical Society, 167(8), http://dx.doi.org/10.1149/1945-7111/ab8a9c.
Society, 168(7), Article 070533.
Perez, H. E., Dey, S., Hu, X., & Moura, S. J. (2017). Optimal charging of Li-Ion
Lu, D., Trimboli, M. S., Fan, G., Zhang, R., & Plett, G. L. (2021c). Nondestructive pulse
batteries via a single particle model with electrolyte and thermal dynamics. Journal
testing to estimate a subset of physics-based-model parameter values for lithium-ion
of The Electrochemical Society, 164(7), A1679–A1687. http://dx.doi.org/10.1149/2.
cells. Journal of The Electrochemical Society, 168(8), Article 080533.
1301707jes.
Lueth, S., Sauter, U. S., & Bessler, W. G. (2015). An agglomerate model of lithium-
Perez, H., Shahmohammadhamedani, N., & Moura, S. (2015). Enhanced performance
ion battery cathodes. Journal of The Electrochemical Society, 163(2), A210–A222.
of Li-Ion batteries via modified reference governors and electrochemical models.
http://dx.doi.org/10.1149/2.0291602jes.
IEEE/ASME Transactions on Mechatronics, 20(4), 1511–1520. http://dx.doi.org/10.
Mai, W., Yang, M., & Soghrati, S. (2019). A particle-resolved 3D finite element model to
1109/tmech.2014.2379695.
study the effect of cathode microstructure on the behavior of lithium ion batteries.
Petit, M., Calas, E., & Bernard, J. (2020). A simplified electrochemical model for
Electrochimica Acta, 294, 192–209. http://dx.doi.org/10.1016/j.electacta.2018.10.
modelling li-ion batteries comprising blend and bidispersed electrodes for high
072.
power applications. Journal of Power Sources, 479, http://dx.doi.org/10.1016/j.
Malik, M., Li, G., & Chen, Z. (2021). An optimal charging algorithm to minimise solid
jpowsour.2020.228766.
electrolyte interface layer in lithium-ion battery. Journal of Power Sources, 482,
Pozzi, A., Torchio, M., Braatz, R. D., & Raimondo, D. M. (2020). Optimal charging of an
Article 228895.
electric vehicle battery pack: A real-time sensitivity-based model predictive control
Mao, Z., Farkhondeh, M., Pritzker, M., Fowler, M., & Chen, Z. (2015). Multi-
approach. Journal of Power Sources, 461, http://dx.doi.org/10.1016/j.jpowsour.
particle model for a commercial blended lithium-ion electrode. Journal of The
2020.228133.
Electrochemical Society, 163(3), A458.
Prada, E., Di Domenico, D., Creff, Y., Bernard, J., Sauvant-Moynot, V., & Huet, F.
Marcicki, J., Canova, M., Conlisk, A. T., & Rizzoni, G. (2013). Design and parametriza-
(2013). A simplified electrochemical and thermal aging model of LiFePO4-graphite
tion analysis of a reduced-order electrochemical model of graphite/LiFePO4 cells
Li-ion batteries: Power and capacity fade simulations. Journal of The Electrochemical
for SOC/SOH estimation. Journal of Power Sources, 237, 310–324. http://dx.doi.
Society, 160(4), A616–A628. http://dx.doi.org/10.1149/2.053304jes.
org/10.1016/j.jpowsour.2012.12.120.
Pramanik, S., & Anwar, S. (2016). Electrochemical model based charge optimization
Marquis, S. G., Timms, R., Sulzer, V., Please, C. P., & Chapman, S. J. (2020). A suite
for lithium-ion batteries. Journal of Power Sources, 313, 164–177. http://dx.doi.org/
of reduced-order models of a single-layer Lithium-Ion pouch cell. Journal of The
10.1016/j.jpowsour.2016.01.096.
Electrochemical Society, 167(14), http://dx.doi.org/10.1149/1945-7111/abbce4.
Mastali Majdabadi, M., Farhad, S., Farkhondeh, M., Fraser, R. A., & Fowler, M. (2015). Qadir, S., Li, G., & Chen, Z. (2022). Simplification of full homogenized macro-scale
Simplified electrochemical multi-particle model for LiFePO4 cathodes in lithium- model for lithium-ion batteries. Journal of Energy Storage, 46, Article 103801.
ion batteries. Journal of Power Sources, 275, 633–643. http://dx.doi.org/10.1016/ Rahman, M. A., Anwar, S., & Izadian, A. (2015). Electrochemical model based fault
j.jpowsour.2014.11.066. diagnosis of a lithium ion battery using multiple model adaptive estimation ap-
Mehta, R., & Gupta, A. (2021). An improved single-particle model with electrolyte proach. In 2015 IEEE international conference on industrial technology (pp. 210–217).
dynamics for high current applications of lithium-ion cells. Electrochimica Acta, IEEE.
http://dx.doi.org/10.1016/j.electacta.2021.138623. Reddy, S. R., Scharrer, M. K., Pichler, F., Watzenig, D., & Dulikravich, G. S. (2019).
Merla, Y., Wu, B., Yufit, V., Martinez-Botas, R. F., & Offer, G. J. (2018). An Accelerating parameter estimation in Doyle–Fuller–Newman model for lithium-ion
easy-to-parameterise physics-informed battery model and its application towards batteries. COMPEL - the International Journal for Computation and Mathematics in
lithium-ion battery cell design, diagnosis, and degradation. Journal of Power Sources, Electrical and Electronic Engineering, 38(5), 1533–1544. http://dx.doi.org/10.1108/
384, 66–79. http://dx.doi.org/10.1016/j.jpowsour.2018.02.065. compel-12-2018-0533.
Miguel, E., Plett, G. L., Trimboli, M. S., Lopetegi, I., Oca, L., Iraola, U., et al. (2021a). Ren, L., Zhu, G., Kang, J., Wang, J. V., Luo, B., Chen, C., et al. (2021). An algorithm
Electrochemical model and sigma point Kalman filter based online oriented battery for state of charge estimation based on a single-particle model. Journal of Energy
model. IEEE Access, 9, 98072–98090. http://dx.doi.org/10.1109/ACCESS.2021. Storage, 39, http://dx.doi.org/10.1016/j.est.2021.102644.
3095620. Richardson, G., Denuault, G., & Please, C. P. (2011). Multiscale modelling and analysis
Miguel, E., Plett, G. L., Trimboli, M. S., Oca, L., Iraola, U., & Bekaert, E. (2021b). of lithium-ion battery charge and discharge. Journal of Engineering Mathematics,
Review of computational parameter estimation methods for electrochemical models. 72(1), 41–72. http://dx.doi.org/10.1007/s10665-011-9461-9.
Journal of Energy Storage, 44, Article 103388. Röder, F., Sonntag, S., Schröder, D., & Krewer, U. (2016). Simulating the impact of
Moura, S. J., Argomedo, F. B., Klein, R., Mirtabatabaei, A., & Krstic, M. (2017). particle size distribution on the performance of graphite electrodes in lithium-ion
Battery state estimation for a single particle model with electrolyte dynamics. IEEE batteries. Energy Technology, 4(12), 1588–1597.
Transactions on Control Systems Technology, 25(2), 453–468. http://dx.doi.org/10. Rodríguez, A., & Plett, G. L. (2017). Controls-oriented models of lithium-ion cells having
1109/tcst.2016.2571663. blend electrodes. part 2: Physics-based reduced-order models. Journal of Energy
Moura, S. J., Chaturvedi, N. A., & Krsti?, M. (2014). Adaptive partial differential Storage, 11, 219–236. http://dx.doi.org/10.1016/j.est.2017.02.006.
equation observer for battery state-of-charge/state-of-health estimation via an Rodríguez, A., Plett, G. L., & Trimboli, M. S. (2018). Improved transfer functions mod-
electrochemical model. Journal of Dynamic Systems, Measurement, and Control, eling linearized lithium-ion battery-cell internal electrochemical variables. Journal
136(1), http://dx.doi.org/10.1115/1.4024801. of Energy Storage, 20, 560–575. http://dx.doi.org/10.1016/j.est.2018.06.015.
22
Sacci, R. L., Black, J. M., Balke, N., Dudney, N. J., More, K. L., & Unocic, R. R. (2015). Sung, W., & Shin, C. B. (2015). Electrochemical model of a lithium-ion battery
Nanoscale imaging of fundamental li battery chemistry: solid-electrolyte interphase implemented into an automotive battery management system. Computers & Chemical
formation and preferential growth of lithium metal nanoclusters. Nano Letter, Engineering, 76, 87–97. http://dx.doi.org/10.1016/j.compchemeng.2015.02.007.
15(3), 8–2011. http://dx.doi.org/10.1021/nl5048626, URL https://www.ncbi.nlm. Tang, S. B., Lai, M. O., & Lu, L. (2008). Li-ion diffusion in highly (003) oriented
nih.gov/pubmed/25706693. LiCoO2 thin film cathode prepared by pulsed laser deposition. Journal of Alloys and
Safari, M., & Delacourt, C. (2011). Modeling of a commercial graphite/LiFePO4 Compounds, 449(1–2), 300–303. http://dx.doi.org/10.1016/j.jallcom.2005.12.131.
cell. Journal of The Electrochemical Society, 158(5), http://dx.doi.org/10.1149/1. Tang, S., Wang, Z., Guo, H., Wang, J., Li, X., & Yan, G. (2019). Systematic parameter
3567007. acquisition method for electrochemical model of 4.35?V LiCoO2 batteries. Solid
Santhanagopalan, S., Guo, Q., Ramadass, P., & White, R. E. (2006). Review of models State Ionics, 343, http://dx.doi.org/10.1016/j.ssi.2019.115083.
for predicting the cycling performance of lithium ion batteries. Journal of Power Tanim, T. R., & Rahn, C. D. (2015). Aging formula for lithium ion batteries with
Sources, 156(2), 620–628. http://dx.doi.org/10.1016/j.jpowsour.2005.05.070. solid electrolyte interphase layer growth. Journal of Power Sources, 294, 239–247.
Santos-Mendoza, I. O., Henquín, E. R., Vazquez-Arenas, J., & Aguirre, P. A. (2020). http://dx.doi.org/10.1016/j.jpowsour.2015.06.014.
Simplified electrochemical model to account for different active/inactive cathode Tanim, T. R., Rahn, C. D., & Wang, C.-Y. (2015). State of charge estimation of a lithium
compositions in Li-ion batteries. Journal of Energy Storage, 31, http://dx.doi.org/ ion cell based on a temperature dependent and electrolyte enhanced single particle
10.1016/j.est.2020.101579. model. Energy, 80, 731–739. http://dx.doi.org/10.1016/j.energy.2014.12.031.
Sauerteig, D., Hanselmann, N., Arzberger, A., Reinshagen, H., Ivanov, S., & Bund, A. Taralova, V., Iliev, O., & Efendiev, Y. (2016). Derivation and numerical validation of a
(2018). Electrochemical-mechanical coupled modeling and parameterization of homogenized isothermal Li-ion battery model. Journal of Engineering Mathematics,
swelling and ionic transport in lithium-ion batteries. Journal of Power Sources, 378, 101(1), 1–27. http://dx.doi.org/10.1007/s10665-015-9842-6.
235–247. Taralova, F., & Lai, W. (2011). Derivation of Micro/Macro Lithium battery models from
Saw, L. H., Ye, Y., & Tay, A. A. O. (2013). Electrochemical–thermal analysis of 18650 homogenization. Transport in Porous Media, 88(2), 249–270. http://dx.doi.org/10.
Lithium Iron Phosphate cell. Energy Conversion and Management, 75, 162–174. 1007/s11242-011-9738-5.
http://dx.doi.org/10.1016/j.enconman.2013.05.040. Tian, J., Wang, Y., & Chen, Z. (2021a). An improved single particle model for
Schmalstieg, J., Rahe, C., Ecker, M., & Sauer, D. U. (2018). Full cell parameterization lithium-ion batteries based on main stress factor compensation. Journal of Cleaner
of a high-power lithium-ion battery for a physico-chemical model: Part I. Physical Production, 278, Article 123456.
and electrochemical parameters. Journal of The Electrochemical Society, 165(16), Tian, J., Wang, Y., & Chen, Z. (2021b). An improved single particle model for
A3799–A3810. http://dx.doi.org/10.1149/2.0321816jes.
lithium-ion batteries based on main stress factor compensation. Journal of Cleaner
Schmalstieg, J., & Sauer, D. U. (2018). Full cell parameterization of a high-power Production, 278, http://dx.doi.org/10.1016/j.jclepro.2020.123456.
Lithium-Ion battery for a physico-chemical model: Part II. Thermal parameters Tourani, A., White, P., & Ivey, P. (2014). A multi scale multi-dimensional thermo
and validation. Journal of The Electrochemical Society, 165(16), A3811–A3819.
electrochemical modelling of high capacity lithium-ion cells. Journal of Power
http://dx.doi.org/10.1149/2.0331816jes.
Sources, 255, 360–367. http://dx.doi.org/10.1016/j.jpowsour.2014.01.030.
Schmidt, A. P., Bitzer, M., Imre, r. W., & Guzzella, L. (2010). Experiment-driven
Tran, N. T., Vilathgamuwa, M., Farrell, T., Choi, S. S., Li, Y., & Teague, J. (2018). A
electrochemical modeling and systematic parameterization for a lithium-ion battery
Padé approximate model of Lithium Ion batteries. Journal of The Electrochemical
cell. Journal of Power Sources, 195(15), 5071–5080. http://dx.doi.org/10.1016/j.
Society, 165(7), A1409–A1421. http://dx.doi.org/10.1149/2.0651807jes.
jpowsour.2010.02.029.
Tranter, T. G., Timms, R., Heenan, T. M. M., Marquis, S. G., Sulzer, V., Jnawali, A., et
Shafikhani, I., et al. (2021). Energy management of hybrid electric vehicles with battery
al. (2020). Probing heterogeneity in li-ion batteries with coupled multiscale models
aging considerations: Wheel loader case study. Control Engineering Practice, 110,
of electrochemistry and thermal transport using tomographic domains. Journal of
Article 104759.
The Electrochemical Society, 167(11), http://dx.doi.org/10.1149/1945-7111/aba44b.
Sharma, A. K., Basu, S., Hariharan, K. S., Adiga, S. P., Kolake, S. M., Song, T., et al.
Traskunov, M. (2015). An extended homogenized porous electrode model for lithium-
(2019). A closed form reduced order electrochemical model for lithium-ion cells.
ion cell electrodes. Journal of Power Sources, 282, 572–580. http://dx.doi.org/10.
Journal of The Electrochemical Society, 166(6), A1197–A1210. http://dx.doi.org/10.
1016/j.jpowsour.2015.02.098.
1149/2.0411906jes.
Traskunov, I., & Latz, A. (2021). Localized fluctuations of electrochemical properties
Smiley, A., & Plett, G. L. (2018). An adaptive physics-based reduced-order model of an
in porous electrodes of lithium-ion batteries: Beyond porous electrode theory.
aged lithium-ion cell, selected using an interacting multiple-model Kalman filter.
Electrochimica Acta, 379, http://dx.doi.org/10.1016/j.electacta.2021.138144.
Journal of Energy Storage, 19, 120–134. http://dx.doi.org/10.1016/j.est.2018.07.
Valentin, O., Thivel, P.-X., Kareemulla, T., Cadiou, F., & Bultel, Y. (2017). Modeling of
004.
thermo-mechanical stresses in Li-ion battery. Journal of Energy Storage, 13, 184–192.
Smith, K. A., Rahn, C. D., & Wang, C.-Y. (2007). Control oriented 1D electrochemical
http://dx.doi.org/10.1016/j.est.2017.07.018.
model of lithium ion battery. Energy Conversion and Management, 48(9), 2565–2578.
Venugopal, G., Moore, J., Howard, J., & Pendalwar, S. (1999). Characterization of
http://dx.doi.org/10.1016/j.enconman.2007.03.015.
microporous separators for lithium-ion batteries. Journal of Power Sources, 77(1),
Smith, K., & Wang, C.-Y. (2006). Solid-state diffusion limitations on pulse operation
34–41.
of a lithium ion cell for hybrid electric vehicles. Journal of Power Sources, 161(1),
628–639. http://dx.doi.org/10.1016/j.jpowsour.2006.03.050. Verbrugge, M., Baker, D., Koch, B., Xiao, X., & Gu, W. (2017). Thermodynamic model
for substitutional materials: application to lithiated graphite, spinel manganese
Stetzel, K. D., Aldrich, L. L., Trimboli, M. S., & Plett, G. L. (2015). Electrochemical
state and internal variables estimation using a reduced-order physics-based model oxide, iron phosphate, and layered nickel-manganese-cobalt oxide. Journal of The
of a lithium-ion cell and an extended Kalman filter. Journal of Power Sources, 278, Electrochemical Society, 164(11), E3243.
490–505. Waldmann, T., Hogg, B.-I., & Wohlfahrt-Mehrens, M. (2018). Li plating as unwanted
Stiaszny, B., Ziegler, J. C., Krau?, E. E., Schmidt, J. P., & Ivers-Tiffée, E. (2014). side reaction in commercial Li-ion cells – A review. Journal of Power Sources, 384,
Electrochemical characterization and post-mortem analysis of aged LiMn2O4– 107–124. http://dx.doi.org/10.1016/j.jpowsour.2018.02.063.
Li(Ni0.5Mn0.3Co0.2)O2/graphite lithium ion batteries. Part I: Cycle aging. Journal Waldmann, T., Wilka, M., Kasper, M., Fleischhammer, M., & Wohlfahrt-Mehrens, M.
of Power Sources, 251, 439–450. http://dx.doi.org/10.1016/j.jpowsour.2013.11.080. (2014). Temperature dependent ageing mechanisms in Lithium-ion batteries – A
Streeter, I., Wildgoose, G. G., Shao, L., & Compton, R. G. (2008). Cyclic voltammetry Post-Mortem study. Journal of Power Sources, 262, 129–135. http://dx.doi.org/10.
on electrode surfaces covered with porous layers: An analysis of electron transfer 1016/j.jpowsour.2014.03.112.
kinetics at single-walled carbon nanotube modified electrodes. Sensors and Actuators Wang, Z., Ma, J., & Zhang, L. (2017). Finite element thermal model and simulation
B (Chemical), 133(2), 462–466. http://dx.doi.org/10.1016/j.snb.2008.03.015. for a cylindrical [li-ion] battery. IEEE Access, 5, 15372–15379. http://dx.doi.org/
Sturm, J., Ennifar, H., Erhard, S. V., Rheinfeld, A., Kosch, S., & Jossen, A. (2018). 10.1109/access.2017.2723436.
State estimation of lithium-ion cells using a physicochemical model based extended Wang, A., O’Kane, S., Planella, F. B., Houx, J. L., O’Regan, K., Zyskin, M., et al. (2021).
Kalman filter. Applied Energy, 223, 103–123. http://dx.doi.org/10.1016/j.apenergy. Parameterising continuum level Li-ion battery models & the LiionDB database. arXiv
2018.04.011. preprint arXiv:2110.09879.
Sturm, J., Rheinfeld, A., Zilberman, I., Spingler, F. B., Kosch, S., Frie, F., et al. (2019). Wang, Y., Tian, J., Sun, Z., Wang, L., Xu, R., Li, M., et al. (2020). A comprehensive
Modeling and simulation of inhomogeneities in a 18650 nickel-rich, silicon-graphite review of battery modeling and state estimation approaches for advanced battery
lithium-ion cell during fast charging. Journal of Power Sources, 412, 204–223. management systems. Renewable and Sustainable Energy Reviews, 131, http://dx.doi.
http://dx.doi.org/10.1016/j.jpowsour.2018.11.043. org/10.1016/j.rser.2020.110015.
Subramaniam, A., Kolluri, S., Santhanagopalan, S., & Subramanian, V. R. (2020). An Wei, Z., Zhao, J., He, H., Ding, G., Cui, H., & Liu, L. (2021). Future smart battery
efficient electrochemical-thermal tanks-in-series model for Lithium-Ion batteries. and management: Advanced sensing from external to embedded multi-dimensional
Journal of The Electrochemical Society, 167(11), Article 113506. measurement. Journal of Power Sources, 489, http://dx.doi.org/10.1016/j.jpowsour.
Subramanian, V. R., Diwakar, V. D., & Tapriyal, D. (2005). Efficient macro-micro 2021.229462.
scale coupled modeling of batteries. Journal of The Electrochemical Society, 152(10), Xavier, M. A., de Souza, A. K., Karami, K., Plett, G. L., & Trimboli, M. S. (2021). A
A2002. computational framework for Lithium Ion cell-level model predictive control using
Subramanian, V. R., Tapriyal, D., & White, R. E. (2004). A boundary condition for a physics-based reduced-order model. IEEE Control Systems Letters, 5(4), 1387–1392.
porous electrodes. Electrochemical and Solid State Letters, 7(9), A259. http://dx.doi.org/10.1109/lcsys.2020.3038131.
23
Xia, H., Lu, L., & Ceder, G. (2006). Li diffusion in LiCoO2 thin films prepared Yuan, S., Jiang, L., Yin, C., Wu, H., & Zhang, X. (2017b). A transfer function type of
by pulsed laser deposition. Journal of Power Sources, 159(2), 1422–1427. http: simplified electrochemical model with modified boundary conditions and Padé ap-
//dx.doi.org/10.1016/j.jpowsour.2005.12.012. proximation for Li-ion battery: Part 2. Modeling and parameter estimation. Journal
Xia, B., Zhao, X., de Callafon, R., Garnier, H., Nguyen, T., & Mi, C. (2016). Accurate of Power Sources, 352, 258–271. http://dx.doi.org/10.1016/j.jpowsour.2017.03.061.
Lithium-ion battery parameter estimation with continuous-time system identifica- Zelič, K., & Katrašnik, T. (2019). Thermodynamically consistent and computationally
tion methods. Applied Energy, 179, 426–436. http://dx.doi.org/10.1016/j.apenergy. efficient 0D lithium intercalation model of a phase separating cathode particle.
2016.07.005. Journal of The Electrochemical Society, 166(14), A3242.
Xiong, R., Li, L., Li, Z., Yu, Q., & Mu, H. (2018). An electrochemical model based Zhang, D., Dey, S., Couto, L. D., & Moura, S. J. (2019). Battery adaptive observer for a
degradation state identification method of lithium-ion battery for all-climate electric single-particle model with intercalation-induced stress. IEEE Transactions on Control
vehicles application. Applied Energy, 219, 264–275. http://dx.doi.org/10.1016/j. Systems Technology, 1–15. http://dx.doi.org/10.1109/tcst.2019.2910797.
apenergy.2018.03.053. Zhang, X., Lu, J., Yuan, S., Yang, J., & Zhou, X. (2017). A novel method for iden-
Xu, J., Wang, T., Pei, L., Mao, S., & Zhu, C. (2020). Parameter identification of tification of lithium-ion battery equivalent circuit model parameters considering
electrolyte decomposition state in lithium-ion batteries based on a reduced pseudo electrochemical properties. Journal of Power Sources, 345, 21–29. http://dx.doi.org/
two-dimensional model with Padé approximation. Journal of Power Sources, 460, 10.1016/j.jpowsour.2017.01.126.
http://dx.doi.org/10.1016/j.jpowsour.2020.228093. Zhang, L., Lyu, C., Hinds, G., Wang, L., Luo, W., Zheng, J., et al. (2014). Parameter
Yang, S.-c., Hua, Y., Qiao, D., Lian, Y.-b., Pan, Y.-w., & He, Y.-l. (2019). A coupled sensitivity analysis of cylindrical LiFePO4Battery performance using multi-physics
electrochemical-thermal-mechanical degradation modelling approach for lifetime modeling. Journal of The Electrochemical Society, 161(5), A762–A776. http://dx.doi.
assessment of lithium-ion batteries. Electrochimica Acta, 326, http://dx.doi.org/10. org/10.1149/2.048405jes.
1016/j.electacta.2019.134928. Zhang, L., Lyu, C., Wang, L., Zheng, J., Luo, W., & Ma, K. (2015). Parallelized
Yin, Y., & Choe, S.-Y. (2020). Actively temperature controlled health-aware fast genetic identification of the thermal-electrochemical model for Lithium-Ion battery.
charging method for lithium-ion battery using nonlinear model predictive control. Advances in Mechanical Engineering, 5, http://dx.doi.org/10.1155/2013/754653.
Applied Energy, 271, http://dx.doi.org/10.1016/j.apenergy.2020.115232. Zhao, L., Liu, Z., & Ji, G. (2018). Lithium-ion battery state of charge estimation with
Yin, Y., Hu, Y., Choe, S.-Y., Cho, H., & Joe, W. T. (2019). New fast charging method of model parameters adaptation using h extended Kalman filter. Control Engineering
lithium-ion batteries based on a reduced order electrochemical model considering Practice, 81, 114–128.
side reaction. Journal of Power Sources, 423, 367–379. http://dx.doi.org/10.1016/ Zou, C., Hu, X., Wei, Z., Wik, T., & Egardt, B. (2017). Electrochemical estimation
j.jpowsour.2019.03.007. and control for lithium-ion battery health-aware fast charging. IEEE Transactions
Yuan, S., Jiang, L., Yin, C., Wu, H., & Zhang, X. (2017a). A transfer function on Industrial Electronics, 65(8), 6635–6645.
type of simplified electrochemical model with modified boundary conditions and Zou, C., Hu, X., Wei, Z., Wik, T., & Egardt, B. (2018). Electrochemical estimation
Padé approximation for li-ion battery: Part 1. lithium concentration estimation. and control for Lithium-Ion battery health-aware fast charging. IEEE Transactions
Journal of Power Sources, 352, 245–257. http://dx.doi.org/10.1016/j.jpowsour. on Industrial Electronics, 65(8), 6635–6645. http://dx.doi.org/10.1109/tie.2017.
2017.03.060. 2772154.
24

Electrochemical Modeling and Parameterization Towards Co - 2022 - Control Engine

Uploaded by

Copyright:

Available Formats

Electrochemical Modeling and Parameterization Towards Co - 2022 - Control Engine

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Electrochemical Modeling and Parameterization Towards Co - 2022 - Control Engine

Uploaded by

Copyright:

Available Formats

Control Engineering Practice 124 (2022) 105176

Contents lists available at ScienceDirect

Control Engineering Practice

Electrochemical modeling and parameterization towards control-oriented

ARTICLE INFO ABSTRACT

𝜂𝑘 = 𝜙𝑠,𝑘 − 𝜙𝑒,𝑘 − 𝑈𝑜𝑐𝑝,𝑘 , 𝑘 ∈ {𝑛, 𝑝} (9)

𝑉 (𝑡) = 𝜙𝑠 (𝐿, 𝑡) − 𝜙𝑠 (0, 𝑡) − 𝑅𝑓 𝐼 (10)

where 𝑅𝑓 is the film resistance.

The Padé approximation of the transcendental transfer functions for

where 𝑅𝑠 is the particle radius, 𝐹 is the Faraday constant, 𝐷𝑠 is the

Fig. 7. Li-ion battery structures at multiple length scales.

one classical solution is to extend P2D or simplified 1D electrochemical

Fig. 10. The coupled 1D electrochemical model and thermal model.

Fig. 12. Main coupling phenomena.

Fig. 13. Future trends of control-oriented modeling.

You might also like