testing, analizing and modelling of lithium-ion … · analyzing and modelling of lithium-ion...
TRANSCRIPT
ESCUELA TÉCNICA SUPERIOR DE INGENIERÍA (ICAI)
TESTING, ANALIZING AND MODELLING OF
LITHIUM-ION BATTERIES
Autor: Alejandro Morras Lorenzo
Director: Michael Pecht
Madrid
Junio de 2016
Alejandro Morrás
2
AUTORIZACIÓN PARA LA DIGITALIZACIÓN, DEPÓSITO Y DIVULGACIÓN EN RED DE
PROYECTOS FIN DE GRADO, FIN DE MÁSTER, TESINAS O MEMORIAS DE BACHILLERATO
1º. Declaración de la autoría y acreditación de la misma.
El autor D. Alejandro Morras Lorenzo____________________________________
DECLARA ser el titular de los derechos de propiedad intelectual de la obra: ____________ Testing,
analyzing and modelling of Lithium-ion batteries ____________________, que ésta es una obra
original, y que ostenta la condición de autor en el sentido que otorga la Ley de Propiedad Intelectual.
2º. Objeto y fines de la cesión.
Con el fin de dar la máxima difusión a la obra citada a través del Repositorio institucional de la Universidad,
el autor CEDE a la Universidad Pontificia Comillas, de forma gratuita y no exclusiva, por el máximo plazo
legal y con ámbito universal, los derechos de digitalización, de archivo, de reproducción, de distribución y de
comunicación pública, incluido el derecho de puesta a disposición electrónica, tal y como se describen en la
Ley de Propiedad Intelectual. El derecho de transformación se cede a los únicos efectos de lo dispuesto en la
letra a) del apartado siguiente.
3º. Condiciones de la cesión y acceso
Sin perjuicio de la titularidad de la obra, que sigue correspondiendo a su autor, la cesión de derechos
contemplada en esta licencia habilita para:
a) Transformarla con el fin de adaptarla a cualquier tecnología que permita incorporarla a internet y hacerla
accesible; incorporar metadatos para realizar el registro de la obra e incorporar “marcas de agua” o
cualquier otro sistema de seguridad o de protección.
b) Reproducirla en un soporte digital para su incorporación a una base de datos electrónica, incluyendo el
derecho de reproducir y almacenar la obra en servidores, a los efectos de garantizar su seguridad,
conservación y preservar el formato.
c) Comunicarla, por defecto, a través de un archivo institucional abierto, accesible de modo libre y gratuito
a través de internet.
d) Cualquier otra forma de acceso (restringido, embargado, cerrado) deberá solicitarse expresamente y
obedecer a causas justificadas.
e) Asignar por defecto a estos trabajos una licencia Creative Commons.
f) Asignar por defecto a estos trabajos un HANDLE (URL persistente).
4º. Derechos del autor.
El autor, en tanto que titular de una obra tiene derecho a:
a) Que la Universidad identifique claramente su nombre como autor de la misma
b) Comunicar y dar publicidad a la obra en la versión que ceda y en otras posteriores a través de cualquier
medio.
c) Solicitar la retirada de la obra del repositorio por causa justificada.
d) Recibir notificación fehaciente de cualquier reclamación que puedan formular terceras personas en
relación con la obra y, en particular, de reclamaciones relativas a los derechos de propiedad intelectual
sobre ella.
5º. Deberes del autor.
El autor se compromete a:
a) Garantizar que el compromiso que adquiere mediante el presente escrito no infringe ningún derecho de
terceros, ya sean de propiedad industrial, intelectual o cualquier otro.
b) Garantizar que el contenido de las obras no atenta contra los derechos al honor, a la intimidad y
a la imagen de terceros.
c) Asumir toda reclamación o responsabilidad, incluyendo las indemnizaciones por daños, que pudieran
ejercitarse contra la Universidad por terceros que vieran infringidos sus derechos e intereses a causa de
Alejandro Morrás
3
la cesión.
d) Asumir la responsabilidad en el caso de que las instituciones fueran condenadas por infracción de
derechos derivada de las obras objeto de la cesión.
6º. Fines y funcionamiento del Repositorio Institucional.
La obra se pondrá a disposición de los usuarios para que hagan de ella un uso justo y respetuoso con los
derechos del autor, según lo permitido por la legislación aplicable, y con fines de estudio, investigación, o
cualquier otro fin lícito. Con dicha finalidad, la Universidad asume los siguientes deberes y se reserva las
siguientes facultades:
La Universidad informará a los usuarios del archivo sobre los usos permitidos, y no garantiza ni
asume responsabilidad alguna por otras formas en que los usuarios hagan un uso posterior de las obras
no conforme con la legislación vigente. El uso posterior, más allá de la copia privada, requerirá que se
cite la fuente y se reconozca la autoría, que no se obtenga beneficio comercial, y que no se realicen
obras derivadas.
La Universidad no revisará el contenido de las obras, que en todo caso permanecerá bajo la
responsabilidad exclusive del autor y no estará obligada a ejercitar acciones legales en nombre del autor
en el supuesto de infracciones a derechos de propiedad intelectual derivados del depósito y archivo de las
obras. El autor renuncia a cualquier reclamación frente a la Universidad por las formas no ajustadas a la
legislación vigente en que los usuarios hagan uso de las obras.
La Universidad adoptará las medidas necesarias para la preservación de la obra en un futuro.
La Universidad se reserva la facultad de retirar la obra, previa notificación al autor, en supuestos
suficientemente justificados, o en caso de reclamaciones de terceros.
Madrid, a ……15….. de …………Junio………………... de ……2016….
ACEPTA
Fdo………………………Alejandro Morras Lorenzo………………………
Motivos para solicitar el acceso restringido, cerrado o embargado del trabajo en el Repositorio Institucional:
La institución (CALCE) con la que el trabajo ha sido realizado va a publicar artículos sobre la
investigación y me han pedido persistentemente que el proyecto no sea de acceso público para
asegurar que su artículo es el primero en publicarse, al menos el primer año.
Alejandro Morrás
4
Declaro, bajo mi responsabilidad, que el Proyecto presentado con el título
……………… Testing, analyzing and modelling of Lithium ion batteries …… en la ETS
de Ingeniería - ICAI de la Universidad Pontificia Comillas en el curso
académico…2015/2016……………. es de mi autoría, original e inédito y no ha sido presentado
con anterioridad a otros efectos. El Proyecto no es plagio de otro, ni total ni parcialmente /y la
información que ha sido tomada de otros documentos está debidamente referenciada.
Fdo.: Alejandro Morras Lorenzo Fecha: ……/ ……/ ……
Autorizada la entrega del proyecto
EL DIRECTOR DEL PROYECTO
Fdo.: Michael Pecht Fecha: ……/ ……/ ……
Vº Bº del Coordinador de Proyectos
Fdo.: Fernando de Cuadra Fecha: ……/ ……/ ……
Alejandro Morrás
5
PRUEBA, ANALISIS Y MODELADO DE BATERIAS DE ION-LITIO
Autor: Morrás Lorenzo, Alejandro
Director: Pecht, Michael.
Supervisores: Das, Diganta
Xing, Laura
Zheng, Daisy
Entidad Colaboradora: CALCE (at University of Maryland) – Center for Advanced
Life Cycle Engineering
RESUMEN
Como resultado de la gran cantidad de contaminación producida por las centrales de
combustibles fósiles y por diversos medios de transporte contaminantes, la atmósfera está repleta
de gases que, unidos con la absorción por parte de éstos de la radiación solar, han derivado en el
conocido efecto invernadero. Ante el peligro de cambiar irreversiblemente el planeta, los
científicos y los gobiernos de todo el mundo han impulsado la búsqueda de alternativas a las
centrales eléctricas tradicionales y a los vehículos propulsados con derivados del petróleo. Se han
producido grandes avances en diferentes tecnologías sostenibles como las alternativas de
generación eléctrica a las tradicionales, las llamadas renovables, y vehículos que contaminan y
consumen cada vez menos, llegando a la contaminación cero del coche eléctrico. El rápido avance
de estas tecnologías ha ido ligado a una búsqueda continua de un método de almacenamiento de
energía eléctrica. Este método es absolutamente necesario en el caso de las energías renovables
para poder competir en eficiencia y coste con las energías tradicionales, y así pasar de ser una
alternativa a una fuente principal.
Análogamente, en el sector del transporte, impulsado por cambios en la legislación,
impuestos más bajos para el comprador y subvenciones para las empresas que investigan, hacen
más atractiva la compra y el uso de vehículos “limpios” para el cliente final.
Esta nueva generación de coches son los híbridos (funcionan con combustible y con un
motor eléctrico) y los coches eléctricos. Un factor crítico para ambos es el dispositivo de
almacenamiento de energía para el vehículo. El almacenamiento de energía tiene un impacto
directo en la autonomía, calidad y seguridad. Los dos principales dispositivos de almacenamiento
de energía son los condensadores y las baterías. Este proyecto se centra en las baterías,
específicamente en las baterías de iones de litio, muy usadas en el sector de la automoción. Este
Alejandro Morrás
6
tipo de baterías también tiene una gran importancia en el sector de los móviles, en el que debido
al gran avance de éstos y a su uso cada vez más intensificado, requieren mejores fuentes de energía,
más eficientes, seguras y duraderas.
El estudio se ha realizado para CALCE (Center for Advanced Life Cycle Engineering), una
institución en estrecha colaboración con la Universidad de Maryland (UMD) en College Park
(EE.UU.). El CALCE es una institución que principalmente investiga la fiabilidad de componentes
electrónicos.
Los objetivos a largo plazo del grupo encargado del trabajo con baterías son construir
modelos para la estimación del estado de la batería en cada instante. Para esto hay que tener en
cuenta los efectos de la temperatura, comparar diferentes modelos y evaluar el rendimiento de los
mismos en diferentes tipos de baterías de iones de litio (células de alta energía y de alta tasa de
corriente). El modelado de las baterías se engloba en una parte del objetivo del CALCE de mejorar
continuamente el Sistema de Gestión de la Batería (BMS en inglés). Este sistema tiene muchas
aplicaciones tales como la optimización y protección de las baterías en los vehículos eléctricos. El
BMS es un conjunto de programas y componentes electrónicos que gestionan y protegen las
baterías frente a condiciones de riesgo. Adicionalmente, el BMS analiza el comportamiento de la
batería, mejora su eficiencia optimizando los valores de corriente y tensión, además de mejorar la
vida de la batería.
En el estudio se seleccionaron cuatro baterías con diferentes composiciones químicas (hay
que notar que la química de cada batería puede diferir de las especificaciones que el fabricante
indica al cliente), también de diferentes tipos, específicamente dos tipos de baterías de alta tasa de
carga (permite tanto la carga rápida y con alta corriente como la descarga con alta corriente) y dos
tipos de baterías de alta energía (alta densidad de energía).
El objetivo del proyecto es diseñar, analizar, comparar y evaluar las pruebas realizadas a
diferentes baterías. También se hace un estudio de la posibilidad de aplicar diferentes modelos de
baterías de iones de litio a una misma batería que se encuentra a diferentes temperaturas y cómo
la temperatura afecta a la exactitud del modelo. Además se busca ver el comportamiento de dos
de los tipos de baterías en un test de degradación. Otra parte del objetivo es comparar las baterías
de alta energía y las de alta corriente.
Con el fin de tener unos datos fiables, lo primero que se ha hecho es realizar una serie de
mediciones iniciales que implican:
• Medición de la capacidad inicial: Con una carga completa y una descarga completa hasta las
respectivas tensiones de corte, podemos obtener la capacidad de la batería como corriente (A) *
tiempo (h) también llamado el método de recuento de Coulomb (Coulomb counting method).
• Medición de la impedancia inicial: Este valor es necesario para hacer un análisis posterior y para
comparar el valor inicial con los valores obtenidos después de determinadas pruebas. Este test
mide tanto la resistencia óhmica como la resistencia de transferencia de carga, incluyendo las
partes real e imaginaria.
Estas pruebas se comparan con las especificaciones de la célula (en la mayoría de los casos
son similares con pequeñas diferencias), pero para los cálculos se utilizan los datos medidos, los
llamados "reales".
Alejandro Morrás
7
Hay que remarcar que hay 3 células por cada tipo de batería (por ejemplo para el tipo SP20
hay SP20-1, SP20-2 y SP20-3), esto se debe a que alguna de las células puede fallar por lo que es
más seguro si se tienen varias y porque algunas de las pruebas degradan a la batería demasiado e
impiden la realización de otros test. Por esta última razón, diferentes pruebas se hicieron a las
células 1 y 3 (test incremental de OCV-SOC) que a la celda 2 (perfil dinámico de corriente) de
cada tipo.
Después de comprobar las especificaciones, la atención se centra en la selección de las
prueba que se van a realizar a las baterías para caracterizarlas en un buen modelo. Al final se
decidió que para identificar los parámetros del modelo se usaría un test de corriente dinámica y,
para estimar la relación entre la tensión de circuito abierto (OCV en inglés) y el estado de carga
(SOC en inglés) se realiza una prueba llamada test incremental del SOC que para cada batería
estima el valor de la tensión de circuito abierto para cada porcentaje del estado de carga. Así con
esta última relación se puede calcular la tensión estimada y calcular el error con respecto a la
medida.
El procedimiento de los test es el siguiente: En primer lugar las células 1 y 3 se sometieron
a un ensayo que carga la batería completamente y luego la descarga en periodos del 10% del SOC
(como está completamente carga el punto inicial es el 100% SOC) , después de cada periodo hay
un tiempo de descanso de 2 horas para tener una buena estimación del OCV en ese específico SOC
y, cuando se haya descargado completamente el procedimiento se invierte para que la batería se
cargue en periodos del 10% del SOC hasta llegar al 100% del SOC . En segundo lugar, la célula
2 se somete a una combinación de 4 perfiles de corriente dinámica (DST, US06, FUDS, BJDST y
una combinación de todos). El DST (Dynamic Stress Test) es el que está siendo utilizado para la
estimación del modelo, mientras que los otros son para probar su robustez.
Todas las pruebas se llevan a cabo a diferentes temperaturas (0C, 25C y 45C) para
comparar el comportamiento de la célula y el modelo en diversas condiciones.
El modelo seleccionado para el estudio es ampliamente utilizado por la industria, este es el
R+RC (primer orden). Otros modelos como el de la resistencia simple o el de segundo orden han
sido estudiados, pero han sido descartados por su gran error y su gran complejidad
respectivamente. Será utilizado Matlab como la herramienta para construir y verificar el modelo.
Este modelo ha sido con anterioridad por otros investigadores para modelar las células de litio-
fosfato, por lo que se quiere probar a ver si funciona con baterías con diferente composición
química.
Este proyecto no ha incluido el parámetro temperatura en el modelado, aunque si en las
pruebas, siendo un aspecto en el que se podría profundizar más para mejorar el modelo.
Por otra parte, se realiza un estudio de degradación de baterías similares, que consiste en
la carga y descarga de las células durante 500 y 1000 ciclos (tomando mediciones al principio y
cuando algo inesperado sucede). En el CALCE, después de terminar los ciclos, se hará una
inspección destructiva con un microscopio de escaneado de electrones (SEM, scanning electron
microscope), para la posterior comparación de la estructura interna de las baterías con diferente
número de ciclos de carga y descarga.
Resultados
Alejandro Morrás
8
Las baterías de alta corriente se diferencian de las de alta energía en que, aplicando la
misma corriente a ambas, las primeras se descargarán más lentamente que las de alta energía. La
temperatura es un factor importante, ya que produce cambios en la capacidad de la batería, las
baterías se descargan más rápidamente a bajas (0C) y altas temperaturas (45C) que a temperatura
ambiente (25C aprox.). Además, El modelo funciona de forma aceptable con todos los tipos de
baterías aunque funciona mejor con las baterías de alta corriente que con las de alta energía y a
altas temperaturas que a bajas o a 25C.
Los resultados de las pruebas de ciclo de vida ha resultado en el fallo inesperado de ambas
células PEA, que muestra un patrón interesante cuando se alcanza por primera vez la capacidad de
0Ah. Las pruebas con las SE28 están todavía en curso, pero los datos disponibles muestran una
tendencia de reducción de la capacidad, como era de esperar.
Conclusiones
Para este informe, se ha realizado un trabajo de investigación profunda de las posibles
pruebas a realizar para caracterizar el comportamiento baterías. Se ha realizado la comparación
entre los diferentes tipos de baterías han sido sometidos a las pruebas y también entre los datos de
la misma prueba que se hace a diferentes temperaturas. El modelo ha sido calculado y probado
para todos y cada uno de los tipos. Este modelo ha sido elegido por sus excelentes resultados y
simplicidad.
Queda como línea para futuros trabajos de investigación, que este modelo pueda predecir
el comportamiento de las baterías a diferentes temperaturas para cualquier tipo de batería
(personalizando sus parámetros y el modelo según su composición química), así el BMS podrá
proteger y gestionar mejor la batería.
Alejandro Morrás
9
ABSTRACT
As a result of the pollution of the fossil fuel plants and means of transportation, a
phenomena called greenhouse effect has forced scientists and governments all world to find
alternatives to the traditional fossil fuel power plants and vehicles powered with derivatives of the
petroleum. The advances of energy technologies specifically with renewable energies has been an
incentive for the research of new and more efficient methods for energy storage. This has its
explanation in the problems that renewable energies have to compete with the traditional energies,
with an efficient method to store the energy the renewable energies will not only be a
complementary energy but also could be a main power source for the society.
Analogically, cars are being forced to be more and more efficient when talking about fuel
consumption and to achieve the lower and lower levels of pollution every year. This research´s
incentive is mostly in the form of lower taxes for the user, grants for the companies that investigate
it and other measures to make the new generation of cars more attractive to the user. These car are
the hybrids (function with fuel and electricity) and the electric cars. For both of them a critical
factor is the energy storage device for the vehicle. The energy storage has a direct impact on
autonomy, quality and safety. The main two energy storage devices are capacitors and batteries.
This report focuses on batteries, specifically in Lithium ion batteries. This kind of batteries are
very important in the mobile phone´s industry, because due to the fast development of these
technologies and the intensive use that society does of the mobile phones, the mobiles require a
better power source, more efficient, durable and safe.
In the study four different batteries with different chemistries were selected (it should be
remarked that the chemistry of each battery may differ from the specifications that the
manufacturer writes on the battery´s specification table), also the type is different there are two
types of high rate batteries (allows fast and high current charge and high current discharge) and
two types of high energy (high energy density) batteries.
The study has been done at CALCE (Center for Advanced Life Cycle Engineering), an
institution in close collaboration with the University of Maryland (UMD) at College Park (USA).
The long term objectives of the battery group are to build models for battery online state
estimation taking into account temperature effects and to compare different models and evaluate
model performance for different types of lithium-ion batteries (high energy cells and high rate
cells). This is a part of the CALCE’s objective to continuously improve the Battery Management
System (BMS), which has many applications such as vehicles. The BMS is set of programs and
electronics that manage and protect the battery form risky conditions and behavior, it improves the
working efficiency of the battery as it controls the current and the voltage, also improving the
battery’s life.
Alejandro Morrás
10
The goal of my project is to build, analyze, compare and evaluate the test done to different
batteries together with the viability and study of the possibility of applying different models of
Lithium Ion batteries to batteries with different temperatures and how this temperature affects the
accuracy of the model. Also to compare the differences between high rate and high energy cells.
Moreover, there an interest to compare the behavior of the two types of cells in a degradation test.
In order to have good data collection, the first thing that has been done is to do the initial
measurements tests that involve:
Initial capacity measurement: With a full charge and a full discharge to the respective cut
off voltages, we can obtain the capacity by current*time also called the Coulomb counting
method.
Initial impedance measurement: This value was necessary to do later analysis and to
compare the initial value with the values after certain tests. This measures the ohmic and
the charge transfer resistance, with real and imaginary part.
These test are compared to the cell’s specifications (in most cases were similar with small
differences), but for all the calculations the data used is the one that has been measured, the one
we called “real”.
It should be remarked that there are 3 cells for each battery type (for SP20 there is SP20-1,
SP20-2 and SP20-3), this is because random failure can occur with one of the cells and because
some of the tests might degrade the battery and can make it useless for different characterizations.
For this last reason, different tests were done to cells 1 and 3 (10% OCV-SOC curve) than to cell
2 of each type.
After the specifications are checked, the focus is on selecting which tests should be done to
the batteries so that a good model can be built. The final decision taken is that the battery will be
tested using dynamic stress tests (using current as input) to identify the model parameters and a 10
%SOC test (described below) that will estimate the relation between the SOC and OCV for the
different types of battery, so that the model can be checked and the error is calculated.
The procedure that is selected is as follows: Firstly cells 1 and 3 are subjected to a test that
fully charges the battery and then discharges the battery in steps of 10% of the SOC (as it is fully
charge, starts at 100% SOC), after each step there is a resting time of 2 hours to have a good
estimation of the OCV at that specific SOC, when it is fully discharged the procedure is reversed
so that the battery is charged in steps of 10% SOC till 100%SOC. Secondly, cell 2 is subjected to
a combination of 4 current stress tests (DST, US06, FUDS, BJDST and a combination of all). The
DST is the one being used for the model estimation while the other are to verify the model
robustness.
All test are done at different temperatures (0C, 25C and 45C) to compare the behavior of the
cell and the model under various conditions.
The model selected for the study is widely used by the industry, this is the R+RC (first order),
using Matlab as the tool to build and verify the model. Other models have been considered, as the
simple resistance and the second order model but have been discarded for their big error and
complexity respectively. This model has been used before to model Lithium-phosphate cells, so it
is tested to see if it works with other chemistries.
Alejandro Morrás
11
The scope of this project just achieved to model and prove the batteries without taking into
account the temperature in the model’s parameters.
The procedure that is selected is as follows: Firstly cells 1 and 3 are subjected to a test that
fully charges the battery and then discharges the battery in steps of 10% of the SOC (as it is fully
charge, starts at 100% SOC), after each step there is a resting time of 2 hours to have a good
estimation of the OCV at that specific SOC, when it is fully discharged the procedure is reversed
so that the battery is charged in steps of 10% SOC till 100%SOC. Secondly, cell 2 is subjected to
a combination of 4 current stress tests (DST, US06, FUDS, BJDST and a combination of all). The
DST is the one being used for the model estimation while the other are to verify the model
robustness.
All test are done at different temperatures (0C, 25C and 45C) to compare the behavior of the
cell and the model under various conditions.
The model selected for the study is widely used by the industry, this is the R+RC (first order),
using Matlab as the tool to build and verify the model. This model has been used before to model
Lithium-phosphate cells, so it is tested to see if it works with other chemistries.
The scope of this project just achieved to model and prove the batteries without taking into
account the temperature in the model’s parameters.
Moreover, test are run on similar batteries to check its life cycle and behavior, doing a constant
cycling until 500 and 1000 cycles (taking measurements at the beginning and when anything
unexpected happens).
Results
High rate cells differ from the high energy cells in the impact that the same current has on
each of them, it will discharge more a high energy than a high rate cell. The temperature is an
important factor because the battery’s capacity changes with temperature so the batteries are
discharged faster with low temperatures and high temperatures than at room temperature (25C
approx.). Furthermore the model works better with high rate than with high energy batteries, also
at high temperature than at room or low temperatures.
The life cycling test results in the unexpected failure of both PEA cells, showing an
interesting pattern when it first reaches a cycle where the capacity is near 0Ah. The SE28 test are
still ongoing but the data available show a trend of capacity reduction. At CALCE, after finishing
the cycling test, there will be a destructive inspection with the SEM (scanning electron
microscope), to see the differences between the batteries with different number of cycles done.
Conclusions
For this report, it has been done a labor of deep research of the possible tests to characterize
batteries’ demeanor. There is a comparison between the different types of batteries have been
subjected to the tests and also between the data of the same test done at different temperatures.
The model is made for all the different data collected and has been chosen for its relation between
simplicity and results (very good results with a relatively simple model). More research should be
Alejandro Morrás
12
done to make this model to predict the behavior of batteries at different temperatures for the all
the different chemistries, like this the BMS can better protect and manage the battery.
I want to express my gratitude to
CALCE for this amazing opportunity.
I sincerely thank Laura and Daisy for
their guidance during my internship at
CALCE and Dr. Das for providing the
opportunity.
I especially thank my family for their
continuous support, especially to
Cristina and my mother for their
patience and caring, also to my father for
all his advice about the report.
Alejandro Morrás
13
Table of Contents
1 Introduction to Li ion batteries ..................................................................... 18
1.1 Types of Li-ion batteries ................................................................................................. 20
1.2 Environment where the research was done: CALCE at the University of Maryland
(College Park) ............................................................................................................................ 25
1.3 Research Objective of the Battery Group (long term) .................................................... 25
1.3.1 Batteries where the tests were performed ............................................................................26
2 Procedure for experiments and measurement of Li/ion batteries ............. 27
2.1 Procedure for Arbin test (storage or korean cells) .......................................................... 28
2.2 Procedure for resistance measuring ................................................................................ 30
2.3 Procedure for tests at Cadex ........................................................................................... 32
3 Battery Tests .................................................................................................... 34
3.1 Initial Measurements for All Samples ............................................................................ 34
3.2 Open Circuit Voltage Tests ............................................................................................ 37
3.2.1 Deeper analysis of the incremental SOC experiment ..........................................................41
3.3 Dynamic Charge/Discharge Testing ............................................................................... 43
4 Application of typical R+RC battery model for all the cells tested ........... 52
5 Comparison of the three tests done to the same battery, degradation
study ........................................................................................................................ 57
6 Cycling Life Test with SEM Observation ..................................................... 58
6.1 Impedance measurement ................................................................................................ 60
7 Conclusion and future research to be done .................................................. 65
8 References and Bibliography: ........................................................................ 66
Alejandro Morrás
14
TABLE OF FIGURES
FIGURE 1: CYLINDRICAL CELL AS THE ONE USED, DESCRIPTION PICTURE (2) ........................... 19
FIGURE 2: SNAPSHOT OF A TYPICAL LICOO2 BATTERY (2**) .................................................... 20
FIGURE 3: SNAPSHOT OF PURE LITHIUM MANGANESE BATTERY (2*) ....................................... 21
FIGURE 4: SNAPSHOT OF NCA (2*) ............................................................................................. 22
FIGURE 5: SNAPSHOT OF NMC (2*) ............................................................................................. 22
FIGURE 6: SNAPSHOT OF LIFEPO4 BATTERY (2*) ....................................................................... 23
FIGURE 7: SNAPSHOT OF LI-TITANATE BATTERY (2*) ................................................................ 24
FIGURE 8: SAMPLES PHOTOGRAPH .............................................................................................. 26
FIGURE 9 TEST BENCH .................................................................................................................. 27
FIGURE 10: SCREENSHOT OF THE PROGRAM USED AT ARBIN .................................................... 29
FIGURE 11: PICTURE OF THE BATTERY TESTING SYSTEM USED ................................................ 30
FIGURE 12: C-CODE FOR CADEX PROGRAM ............................................................................... 32
FIGURE 13: CADEX TESTING SYSTEM USED .............................................................................. 33
FIGURE 14: INITIAL IMPEDANCE OF SP20-1 ................................................................................ 35
FIGURE 15: CC,CV TEST PROFILE OF SAMPLE SP20-1 ............................................................... 36
FIGURE 16: TEST PROCEDURE OF INCREMENTAL OCV TEST ..................................................... 38
FIGURE 17: OCV-SOC CURVE OF LOW-CURRENT OCV TEST ................................................... 39
FIGURE 18: OCV-SOC CURVE OF INCREMENTAL OCV TEST .................................................... 40
FIGURE 19: OCV-SOC CURVES OF TWO OCV TESTS AT ROOM TEMPERATURE (25OC) AND
HIGH TEMPERATURE (45OC) ................................................................................................. 40
FIGURE 20: OCV-SOC AVERAGE CHARGE/DISCHARGE CURVES OF OCV TESTS AT LOW
TEMPERATURE (0ºC), ROOM TEMPERATURE (25OC) AND HIGH TEMPERATURE (45OC)...... 41
Alejandro Morrás
15
FIGURE 21: INCREMENTAL OCV-SOC TEST OF SE28-4 AT 45OC ............................................. 42
FIGURE 22: BATTERY DYNAMIC TEST LOAD PROFILES FOR SP20: 1) DYNAMIC STRESS TEST, 2)
FEDERAL URBAN DRIVING SCHEDULE, 3) HIGHWAY DRIVING SCHEDULE, AND 4) BEIJING
DYNAMIC STRESS TEST .......................................................................................................... 45
FIGURE 23: DYNAMIC STRESS TEST (DST) CYCLING FROM 80%SOC ....................................... 46
FIGURE 24:FEDERAL URBAN DRIVING SCHEDULE (FUDS) CYCLING FROM 80%SOC .............. 46
FIGURE 25: US06 HIGHWAY DRIVING SCHEDULE CYCLING FROM 80%SOC............................. 47
FIGURE 26: BEIJING DYNAMIC STRESS TEST CYCLING FROM 80%SOC .................................... 47
FIGURE 27: PEA-5 DST EXAMPLE .............................................................................................. 49
FIGURE 28: EXAMPLE OF THE VOLTAGE MEASUREMENT AS A RESULT OF THE CURRENT STRESS
TEST ....................................................................................................................................... 49
FIGURE 29: PEA-5 DIFFERENT TEMPERATURES CURRENT STRESS TESTS ................................ 50
FIGURE 30: CURRENT STRESS TEST COMPARISON BETWEEN HIGH ENERGY AND HIGH RATE
CELLS ..................................................................................................................................... 51
FIGURE 31: PROPOSED R+RC BATTERY MODEL (6) ................................................................... 52
FIGURE 32: PEA-5 AT 25OC ......................................................................................................... 54
FIGURE 33: ABSOLUTE ERROR OF THE ESTIMATED VOLTAGE AT 25OC ..................................... 55
FIGURE 34: PPPF-2 AT 25OC ....................................................................................................... 56
FIGURE 35: ABSOLUTE ERROR OF THE ESTIMATED VOLTAGE AT 25OC ..................................... 56
FIGURE 36: CAPACITY CALCULUS EXPLANATION USING A CURRENT DISCHARGE EXAMPLE .... 59
FIGURE 37: IMPEDANCE MEASUREMENT EXAMPLE..................................................................... 60
FIGURE 38: PEA-2 IMPEDANCE COMPARISON ............................................................................ 61
FIGURE 39 PEA-2 CYCLING RESULTS ......................................................................................... 62
FIGURE 40: PEA-3 CYCLING RESULTS ........................................................................................ 62
FIGURE 41: SE28-2 CYCLING RESULTS ....................................................................................... 63
FIGURE 42: SE28-3 CYCLING RESULTS ....................................................................................... 63
Alejandro Morrás
16
TABLE OF TABLES
TABLE 1: SAMPLE SPECIFICATIONS ............................................................................................. 26
TABLE 2: INITIAL CAPACITY OF EACH BATTERY SAMPLE COMPARED TO THE ONE WRITTEN IN
THE SPECIFICATIONS ............................................................................................................. 36
TABLE 3: INCREMENTAL OCV-SOC TEST DATA OF SE28-4 AT 45OC ....................................... 42
TABLE 4: MODEL RESULTS SUMMARY FOR DATA AT 25OC ......................................................... 53
TABLE 5: PEA-5 MODEL RESULTS AT DIFFERENT TEMPERATURES ............................................ 54
TABLE 6: DATA FOR DEGRADATION STUDY ................................................................................. 57
Alejandro Morrás
17
Acronyms and essential concepts definition
For this report it is important to understand the following concepts:
SOC (State Of Charge): It is the percentage of power that is available to be used at that
moment. It is measured in percentage. But when representing it in figures, it is often shown
in a scale from 0 to 1 (1-0.9-0.8-0.7-0.6-… 0).It is often calculated a: 𝑆𝑂𝐶 =𝐼×𝑡
𝑅𝑒𝑎𝑙 𝑐𝑎𝑝𝑎𝑐𝑖𝑡𝑦
OCV (Open Circuit Voltage): is the difference of electrical potential between two terminals
of the battery when disconnected from any circuit.
DOD (Depth of Discharge): It is the inverse to the SOC, percentage of power that has
already been used.
CALCE: The Center for Advanced Life Cycle Engineering
C-rate: Is a measure of the rate at which a battery is discharged relative to its maximum
capacity. 1C rate means that the discharge current will discharge the entire battery in
approximately 1 hour.
Cut-off Voltage: It can be upper or lower, is the voltage limit for the battery (it is given by
the specifications of the manufacturer).
CVCharge: Charge performed applying the higher cut off voltage until a specific low
current (usually 0.01C) is reached.
CCCharge/CCDischarge: Charging/discharging with a constant current (CC).
RCT: Charge Transfer Resistance
R0: Ohmic resistance in series of the battery model
Rd: Polarization resistance
Cp: Polarization capacitance
Ud: Polarization voltage
Alejandro Morrás
18
1 Introduction to Li ion batteries
One of the main problems that our society has nowadays is to find new, sustainable and
cost efficient alternatives to fossil fuels when producing energy, either for transport purposes
or to obtain electricity. There is a high dependence on this kind of resources (fossil fuels) that
when they are used, produce a big amount of undesirable contaminant particles. As the result
of many years of pollution because of the use of traditional sources, the world has come up with
the greenhouse effect. This is that the temperature of the earth is increasing as radiation from
the planet's atmosphere warms the planet's surface to a temperature above what it would be with
the atmosphere that the Earth originally had. The technology is trying to solve this problem by
developing new methods of energy production, these are the renewable energies. With efficient
renewable energies the amount of pollution will significantly decrease and the greenhouse
effect will be stopped before there are any bigger consequences than the raise of a couple of
degrees in the earth temperature.
There are many different renewable energies that use natural resources as wind, solar
power or water (waterfalls, tides…), they are becoming more and more efficient but they have
found a big obstacle to increase significantly their efficiency and it is that in most of the
renewable energies there is no efficient way to store the electricity that they produce. Scientists
and engineers are working on this issue and for now, the best approach has been the batteries.
Not only we pollute when producing energy but also there is a big pollution when using
the traditional means of transport like cars, which is the most used. Research has been done in
how to make a sustainable car and for now there are mainly two options, the hybrid vehicle and
the electric vehicle (involving direct electricity use). For this type of cars their main fuel is the
electricity and a big challenge has been how to store this electricity so that the cars have an
efficiency and an autonomy similar to traditional cars powered by fossil fuels. The answer is,
in the case of electric vehicles, electric batteries.
An electric battery is an energy storage element that with electrochemical reactions, works
as a power supply for electric devices. This kind of batteries are very important in the mobile
phone´s industry, because due to the fast development of these technologies and the intensive
Alejandro Morrás
19
use that society does of the mobile phones, they require a better power source, more efficient,
durable and safe. There are many different types but the one that is mostly used and where this
report is centered is in the study of Lithium ion (intercalated lithium compound) batteries.
As it is explained in the US patent ‘US7563541 B2’: The Lithium Ion battery is composed
by two electrodes, “a positive electrode including a current collector and a first active material
and a negative electrode including a current collector, a second active material, and a third
active material. The first active material, second active material, and third active material are
configured to allow doping and undoping of lithium ions” (1). The battery will degrade if the
voltage is smaller than the “corrosion potential of the current collector of the negative electrode”
(1) and if it is bigger than the “decomposition potential of the first active material” (1).
As it is clearly explained in the Figure 1, there is a separator between the positive plate
and the negative plate. Despite of the variety of chemistries that is used to build batteries
Figure 1: Cylindrical cell as the one used, description
picture (2)
Alejandro Morrás
20
modern Li ion batteries use a gelled electrolyte to build the well-known and broadly used
Lithium –polymer batteries (in the 1970’s the electrolyte was solid).
1.1 Types of Li-ion batteries
There are also different types of Li ion batteries based on its chemistry, the main ones are:
LiCoO2, LiMn2Co4, LiNiCoAlO2, LiNiMnCoO2, and LiFePO4 (2).
LiCoO2 (2):
o Cathode: CoO2-
o Anode: Graphite carbon
Advantages: Very high specific energy
Disadvantages: As we can observe in the snapshot, it has low life span,
specific power and is subject to safety issues regarding temperature.
Examples where it is used: cell phones, personal computers.
Figure 2: Snapshot of a typical LiCoO2 battery (2**)
LiMn2O4 or LMO (2):
o Cathode: Manganese oxide
Alejandro Morrás
21
o Anode: Graphite carbon
Advantages: Low internal resistance that results in high rate cells and
thermally stable.
Disadvantages: very low life span
Examples where it is used: is only used for specific projects and has
evolved into the Lithium Nickel Manganese Cobalt battery for its better
specifications.
Figure 3: Snapshot of pure Lithium Manganese battery (2*)
LiNiCoAlO2 (NCA) (2):
o Cathode: Combination of nickel-cobalt-aluminum (NMC).
o Anode: Graphite carbon
Advantages: High specific energy, remarkable specific power and good
life span.
Disadvantages: Low safety and high costs
Examples where it is used: A possible application are electric vehicles
powertrains.
Alejandro Morrás
22
Figure 4: Snapshot of NCA (2*)
LiNiMnCoO2 (NMC) (2):
o Cathode: Combination of nickel-manganese-cobalt (NMC).
o Anode: Graphite carbon
Advantages: Very high specific energy, improved life span compared to
the LMO, it is the one that has more equilibrium within the
characteristics.
Disadvantages: If silicon is added to the battery it becomes unstable.
Examples where it is used: It is the preferred battery for electric vehicles,
e-bikes and other devices.
Figure 5: Snapshot of NMC (2*)
Alejandro Morrás
23
LiFePO4(2)
o Cathode: Nano-scale phosphate
o Anode: Graphite carbon
Advantages: Very high specific energy, low internal resistance (high rate
current), good life span.
Disadvantages: As we can observe in the snapshot, it has very low
specific energy.
Examples where it is used:” to replace the lead acid battery starter” (2).
Figure 6: Snapshot of LiFePO4 battery (2*)
Lithium Titanate (Li4Ti5O12) (2):
o Cathode: Graphite carbon
o Anode: Li-Titanate
Advantages: High rate cell, safe and very good life cycle capacity.
Disadvantages: As we can observe in the snapshot, it has very low
specific energy and is expensive.
Examples where it is used:” electric powertrains, UPS and solar-powered
street lighting” (2).
Alejandro Morrás
24
Figure 7: Snapshot of Li-Titanate battery (2*)
For the batteries it is extremely important to know its State of Charge (SOC) to improve
the battery´s performance. Problems like overcharging when there is a change in capacity at
different temperatures can be avoided with an accurate estimate of the SOC. In addition, this
estimation is basic to model de behavior of the batteries.
Electric battery modelling, in particular with lithium ion batteries (is the type of
rechargeable batteries that is used the most), is very difficult because there is not a clear
understanding of what makes a battery fail and how is its performance when it is going to fail.
There has been a lot of research on this topic (battery modelling), but there are many different
chemistries and each one, as it has been shown in the introduction, works differently so a model
might be good for just one type of batteries.
The study that is done in this report will analyze the behavior of several batteries with
different chemistries when specific tests are done. Moreover, a model built for a specific
chemistry will be tried in different chemistries, an analysis of it and a comparison between them
will be made.
Alejandro Morrás
25
1.2 Environment where the research was done: CALCE at the University of
Maryland (College Park)
The Center for Advanced Life Cycle Engineering (CALCE), as it explained in its
webpage: is “the largest electronic products and systems research center focused on electronics
reliability, is dedicated to providing a knowledge and resource base to support the development
of competitive electronic components, products and systems”(3).
1.3 Research Objective of the Battery Group (long term)
1. Build the models for battery online state estimation taking into account temperature
effects.
2. Compare different models and evaluate model performance for different types of
lithium-ion batteries (high energy cells and high rate cells.)
In this report the focus is on which tests were done or will be done to the batteries so that
the objective can be achieved, what is the purpose of each test, which equipment was necessary
to do it and analysis of the results of the test with a comparison between the results with high
rate cells and high energy cells. Furthermore, a typical R+RC battery model for LiFePO4
batteries has been tried with batteries with a different chemistry, then the results for high rate
and high energy cells were also compared.
Alejandro Morrás
26
1.3.1 Batteries where the tests were performed
Table 1: Sample Specifications
Brand Model Type Name Material No Nominal
voltage
Rated
capacity
Voltage
Range
Samsung
SDI
INR
18650-20R High rate SP20
LiNiMn
CoO2 3 3.6V 2.0Ah 2.5 ~ 4.2V
Samsung
SDI
ICR
18650-28A
High
energy SE28 LiCoO2 3 3.75V 2.8 Ah 2.75 ~ 4.3 V
Panasonic NCR
18650-PF High rate PPPF
LiNiCoA
lO2 3 3.6V 2.9 Ah 2.5 ~ 4.2 V
Panasonic NCR
18650-A
High
energy PEA
LiNiCoA
lO2 3 3.6V 3.1 Ah 2.5 ~ 4.2 V
Panasonic NCR
18650-B
High
energy PEB
LiNiCoA
lO2 3 3.6V 3.4 Ah 2.5 ~ 4.2 V
LG Chem. ICR
18650-E1
High
energy LEE1 LiCoO2 3 3.75 V 3.2 Ah 2.75 ~4.35 V
All the batteries testes are commercially distributed so its specifications are public to everyone
(see bibliography).
Figure 8: Samples photograph
Alejandro Morrás
27
2 Procedure for experiments and measurement of Li/ion batteries
The tools used for the experiments were an Arbin instrument to charge/discharge the
batteries, a Cadex product with the same functions that the Arbin but with a different procedure
that will be explained. Finally the last part of the equipment used is an instrument that, with the
help of a program, measured the internal resistance of the different batteries. The experiment
platform is already set up, as shown in Fig. 2. It consists of 1) test samples, 2) Arbin BT2000
battery test system, 3) Yamato drying oven DVS 402C, 3*) Freeze concepts freezer with
Temperature control and 4) PC with Arbin software to give test system orders (i.e. charging,
discharging) and 5) monitor data information. Test samples are placed inside the oven or in the
freezer so that the samples’ ambient temperature can be controlled. To note, all the test data is
measured and recorded in 1-60 second interval, depending on the test that’s is being done.
Figure 9 Test bench
1 2
3
4
5 3*
Alejandro Morrás
28
2.1 Procedure for Arbin test (storage or korean cells)
Arbin works with a program called MITS Pro, with an intuitive interface, we designed the
different programs for the battery test. Here is an example of the procedure for a SP20 battery
capacity test.
Steps
1. Open the Arbin program and select the channels you want to use
2. Open/create a new program, in the case of storage cells we will use
PL_ChargeDischargeCharge
3. Check the values of the different steps:
a. Rest: 1 min with DV_Time = 10 s.
b. CCCharge: The current must be C/2 (in this case 0.75 A), till
Voltage>=4.2(according to the upper cut-off voltage of the cell to be tested).
DV_Time = 10 s.
c. CVcharge: The voltage must be 4.2 V, till current <=0.01*C (in this case
0.015A). DV_Time = 10 s.
d. DCIR (Internal resistance phase): Here the values must be Amp=0.06, ms=10
and offset=0.12.
e. Rest: 5 min with DV_Time = 10 s.
f. CCDischarge: The current must be -C/2 (in this case -0.75 A), until
Voltage<=2.5(according to the lower cut-off voltage of the cell to be tested).
DV_Time = 10 s.
g. DCIR (Internal resistance phase): Here the values must be Amp=0.06, ms=10
and offset=0.12.
h. Rest: 5 min with DV_Time = 10 s.
Alejandro Morrás
29
After setting the program we should look carefully on how the test is going to be saved.
In this case: Month_Day_Year_PLN_temperatureC_3W. In the comments we should specify
the SOC of each channel, saying also the name of the battery.
Figure 10: Screenshot of the Program used at Arbin
Alejandro Morrás
30
2.2 Procedure for resistance measuring
First we have to switch on the equipment and connect the cell.
Open the program in the desktop with the name Kane.
Check the values of the table (the ones that are showed in the photo below)
Click two time in the auto filter tick and make sure it is as shown
Click on the ON button and press play.
Check that the picture shown when it is finished is half a circle and positive slope ramp
(drawn with a continuous line).
After finishing change the excel file from the IMB folder, to the Korean cells folder (if
they are different batteries there will be a different folder)and save it as
month_day_year_battery name
Figure 11: Picture of the Battery Testing System used
Alejandro Morrás
31
To finish press OFF and close the window and press anything on the keyboard for the
black window to disappear.
Turn off the equipment and disconnect everything.
Alejandro Morrás
32
2.3 Procedure for tests at Cadex
Open the program and select a C-Code that has already been done or complete a new one with
the corresponding values, based on test cell’s specification, as shown in the image below.
Figure 12: C-Code for Cadex Program
i. For Korean cells the procedure will be the following:
a. Rest: 5 min.
True: Next step
False: No action.
b. CCCharge: The current must be C/2, till Voltage>=4.2 (cell upper cut-off
voltage). DV_Time = 1 s.
True: Next step
False: No action.
Alejandro Morrás
33
c. Rest: 30 min.
True: Next step
False: No action.
d. CCDischarge: The current must be -C/2, till Voltage<=2.5 (cell lower cut-off
voltage). DV_Time = 1 s.
True: Next step
False: No action.
e. Rest: 30 min.
True: No action.
False: No action.
The CADEX Testing System used was used mainly for the life cycling test
Figure 13: CADEX Testing System used
Alejandro Morrás
34
3 Battery Tests
3.1 Initial Measurements for All Samples
When the samples were received, we conducted initial tests for each cell at room
temperature (25oC±5oC) as follows. First, sample’s initial impedance was measured using
electrochemical impedance spectroscopy (EIS) technique. Secondly, the sample was fully
charged using constant current/constant voltage (CCCV) charging protocol. Then the sample
was fully discharged by a constant current. During the discharging process, sample’s capacity
was calculated by the integral of current. Lastly, the sample was charged to full-charged state
at which the sample’s impedance was measured.
EIS technique can be used to assess the health of a battery. When an AC voltage or current
signal is injected into a battery, the corresponding current or voltage response of the battery can
be used to calculate the complex impedance of the battery. For a given frequency AC signal,
the battery will exhibit both a real and imaginary impedance response. This can be visualized
using a Nyquist plot seen in the figure below. The real impedance is graphed on the horizontal
axis and the imaginary impedance is graphed on the vertical axis. The offset of the curve from
the vertical axis is labeled Rohm, or the ohmic resistance. The ohmic resistance accounts for
the resistance of electron flow due to the cell components including the current collectors,
electrodes, and electrolyte. The width of the semi-circle represents the charge transfer
resistance, Rct. The charge transfer resistance accounts for the resistance to charge transfer
processes at the surface of the electrode particles. This is due to the solid electrolyte interphase
(SEI) layer inhibiting charge transfer. Fig.14 shows the initial impedance of SP20-1 which is
one of the three SP20 samples.
Alejandro Morrás
35
Figure 14: Initial impedance of SP20-1
As mentioned before, sample’s capacity is characterized by charging and discharging
processes using coulomb counting method. Battery should be charged and discharged at an
appropriate rate based on their data sheet using CC, CV charging and CC discharging. For our
test samples, they were all charged at a rate of 0.5C until they reached the upper cut-off voltage.
Then they were charged at the upper cut-off voltage until their charging current is less than
0.01C. Next, the samples were rested for 0.5 hour. At last, the samples were discharged at a
rate of 0.5C until the lower cut-off voltages were reached. The fully discharged capacity is
recorded as the initial capacity of the sample. Fig.14 shows the current and voltage profile of
SP20-1 for initial capacity measurement.
Increasing Frequency
Rohm Rct
0.0125 Hz
1640 Hz
Alejandro Morrás
36
Figure 15: CC,CV test profile of sample SP20-1
Table 2: initial capacity of each battery sample compared to the one written in the
specifications
Name Initial discharge capacity (Ah) Nominal capacity (Ah)
SP20-1 2.009 2.000
2.000
2.000
SP20-2 2.028
SP20-3 1.978
SE28-1 2.739 2.800
SE28-2 2.704 2.800
SE28-3 2.698 2.800
PPPF-1 2.786 2.900
PPPF-2 2.767 2.900
PPPF-3 2.777 2.900
PEA-1 2.927 3.100
PEA-2 2.970 3.100
PEA-3 2.958 3.100
PEB-1 3.195 3.400
PEB-2 3.189 3.400
PEB-3 3.204 3.400
LEE1-1 2.929 3.200
LEE1-2 3.109 3.200
LEE1-3 3.110 3.200
Alejandro Morrás
37
3.2 Open Circuit Voltage Tests
The open circuit voltage (OCV) of a battery is the cell potential when no load is applied.
OCV curve is an electrochemical property of the electrode materials and represents the potential
difference between the anode and cathode when varying amounts of lithium are intercalated
into their structures. Battery’s OCV is a function of its state of charge (SOC). Therefore, OCV
can be used for online SOC estimation if we have established an accurate OCV-SOC
relationship. In industry and academia, there are 2 OCV tests commonly used for OCV-SOC
relationship determination. One OCV test is called low-current OCV test in which the battery
sample is charged and discharged by a small current (i.e.0.05C). The terminal voltage under
small current is considered as equilibrium potential since the small current minimizes the
polarization effects inside the cell. In addition, the average value of voltage during charging
and discharging is calculated. This value is recorded as battery’s true OCV because the effects
of hysteresis and ohmic resistance are reduced by the averaging.
The other OCV test is called incremental OCV test in which the battery sample is charged
and discharged using pulse current-relaxation duration at every 10% state of charge (SOC).
Similarly, the voltages of charging and discharging are averaged to get the true OCV. In our
work, 2 hour relaxation duration is adopted after pulse charging/discharging to eliminate the
polarization effects. Fig.16 shows the test procedure of incremental OCV test.
Alejandro Morrás
38
Figure 16: Test procedure of incremental OCV test
Two OCV tests were conducted on sample SP20-1 at room temperature (25oC±5oC). In
low-current OCV test, sample was charged and discharged at a rate of 1/20C. A 2h resting time
was applied between the charging and discharging portion. In incremental OCV test, sample
was charged/ discharged by a pulse current at a rate of C/2. The duration of the current pulse is
12 minutes so that the sample’s SOC is increased/ decreased by 10%. Besides, a 2h resting time
was applied between every two pulses.
C/2 discharging
C/2 charging
Alejandro Morrás
39
Fig.17 and Fig. 18 show the OCV-SOC curves obtained by 2 OCV tests at room
temperature (25oC±5oC). To investigate the temperature dependency of OCV-SOC
relationship, OCV tests are going to be conducted at other temperatures. We have finished 2
OCV tests at high temperature (45oC) and obtained the OCV-SOC curves. The OCV-SOC
relationships at two temperatures are shown in Fig.19.
Figure 17: OCV-SOC curve of low-current OCV test
Alejandro Morrás
40
Figure 18: OCV-SOC curve of incremental OCV test
Comparing figures 18 and 19, it can be concluded that the low current OCV-SOC test, achieves
more values when reaching 0% SOC. The other values are fairly similar.
Figure 19: OCV-SOC curves of two OCV tests at room temperature (25oC) and high
temperature (45oC)
Alejandro Morrás
41
We can observe in Fig.20, the incremental OCV-SOC test done at different
temperatures, it should be remarked that the biggest difference between the data is the OCV at
0%SOC. Moreover, in Fig.19 the OCV value at 0%SOC of the battery at 45 ºC is bigger than
at 25oC while Fig.20 shows that the smaller the temperature, the bigger is the OCV value at
0%SOC. The reason for this difference might be because of their different chemistry.
3.2.1 Deeper analysis of the incremental SOC experiment
The purpose of this test is to try to reach, with a different strategy than the ones used in
other studies done with SE28-4, a percentage of SOC greater than 90%. When using strategies
of low current charging and discharging the greatest point that was measured was 88%.
The strategy performed in this study is to charge a battery with a C/2 current and a
resting time of 2 hours every 10%SOC. This resting time is very important to have a precise
2,9
3,1
3,3
3,5
3,7
3,9
4,1
4,3
0 0,2 0,4 0,6 0,8 1
Vo
ltag
e
SOC
PEA-4
45C
25C
0C
Figure 20: OCV-SOC average charge/discharge curves of OCV tests at
low temperature (0ºC), room temperature (25oC) and high temperature
(45oC)
Alejandro Morrás
42
measurement of the voltage of the battery. The frequency of the resting times was chosen
looking at previous analysis (4) that were carried on by other teams.
Prior to this test, the cell was charged up to its maximum capacity (100%SOC).
Table 3: Incremental OCV-SOC test data of SE28-4 at 45oC
Here we can find the results of doing the test of charging and discharging the battery, the
data were collected just when the resting time was about to end.
SE28-4 OCV 45oC
SOC Vdisch 45C Vch 45C
1 4.25178051 4.26424551
0.9 4.14606905
0.8 4.05023241 4.040681362
0.7 3.968966007 3.996972084
0.6 3.90162158 3.926551819
0.5 3.825211525 3.845770836
0.4 3.78992033 3.805461407
0.3 3.766123056 3.779397726
0.2 3.721928358 3.753657818
0.1 3.668668032 3.679837942
0 3.363837004 3.363837004
3,3
3,4
3,5
3,6
3,7
3,8
3,9
4
4,1
4,2
4,3
4,4
0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1
Vo
ltag
e (V
)
SOC
SE28-4
Vdisch 45C
Vch 45C
Figure 21: Incremental OCV-SOC test of SE28-4 at 45oC
Alejandro Morrás
43
When did it reach each percentage of the SOC was determined by the calculation of the time
that it took to gain 10%SOC. For this battery each charging/discharging period was 727 second
long.
Regarding at the figure that shows current vs time, we can clearly see that every time that
the current goes to zero is the resting time. During the 8th period of charging (80%-90%SOC),
we can observe a strange behavior in the current value, it goes to zero earlier than expected. For
that reason we can infer that the battery has already reached his maximum charge.
The hysteresis phenomena can be identified in the difference between the OCV for the
charging process and the OCV for the discharging process. The difference between the charging
and discharging process, is around 1.5 mV (1). This difference is small compared to the range
of voltage covered by the whole process. The gradient stays almost the same for the whole
process with the exception of the first two points. The practical capacity of the battery used in
the experiment is 2.8 Ah and the change in capacity per 1 mV change of OCV is 0.029 Ah/mV
in the vicinity of the 50% DOD. A 1.5 mV uncertainty in the OCV means a 0.04 Ah capacity
estimate uncertainty, with a corresponding SOC uncertainty of around 1.53%. This uncertainty
is small enough to conclude that the test is accurate
3.3 Dynamic Charge/Discharge Testing
Battery dynamic charge/discharge test uses different load profiles to verify that the
battery can deliver its specified power and thus to evaluate the battery performance. Dynamic
stress test (DST), federal urban driving schedule (FUDS), and US06 highway driving schedule
are three representative battery dynamic performance tests based on a time–velocity profile
from an automobile industry standard vehicle. Beijing dynamic stress test (BJDST) is a dynamic
test based on a simulated load profile for Beijing No.90 electric bus. For example, to conduct a
DST test in the laboratory, a dynamic current sequence was transferred from the corresponding
time–velocity profile, programmed to charge or discharge the battery and applied to battery
performance test. Likewise, FUDS, US06, and BJDST tests can be run on the test sample based
on the corresponding current profiles. Notably, the current sequence should be scaled to fit the
specification of the test battery and the limitation of the testing system (Arbin). The following
Alejandro Morrás
44
two variable power discharge regimes have been defined for USABC (data obtained from the
USABC Manual Revision 2) testing:
1. FUDS – “A second-by-second dynamic regime calculated using the FUDS
vehicle time- velocity profile with a hypothetical electric van having a peak
power demand of 111 W/kg and average power of about 10 W/kg. The actual
profile used for testing, sometimes referred to as FUDS79, is derived by
artificially limiting the peak power demand to 79 W/kg. This "clipped" power
profile can then be scaled to any desired maximum power demand”, (5).
2. “DST (Dynamic Stress Test) - a simplified variable power discharge cycle with
the same average characteristics as the FUDS regime. The DST uses a 360
second sequence of power steps with only 7 discrete power levels”. (5).
The sequence durations of DST, FUDS, US06 and BJDST are 360 seconds, 1372
seconds, 600 seconds, and 916 seconds, respectively. Four current load profiles of SP20 cell is
shown in Fig. 22. To note, negative current denotes discharging process and vice versa.
Alejandro Morrás
45
Figure 22: Battery dynamic test load profiles for SP20: 1) dynamic stress test, 2) federal
urban driving schedule, 3) highway driving schedule, and 4) Beijing dynamic stress test
We have conducted the dynamic charging/discharging tests of battery sample SP20-2 at
room temperature (25oC ±5oC) and high temperature (45oC). In these tests, the sample was first
fully charged, and then discharged to a certain SOC point which is recorded as initial SOC. This
initial SOC can be considered as a true value since it is precisely calculated by the integral of
discharging current which starts from a fully charged state (i.e. 100%SOC). Initial SOCs for
SP20-2 were set at 80% and 50% so that the true initial value covered its major working range
from 20% to 85% SOC. Then the sample was run by DST test until the sample reached its lower
cut-off voltage. The measured current and voltage of SP20-2 cell with initial 80%SOC are
shown in Fig.23. Likewise, FUDS, US06, and BJDST tests were also run following the same
procedure and the corresponding test results are shown in Fig.24, Fig.25 and Fig.26.
0 100 200 300 400 500 600
-4-202
Cu
rren
t (A
)
DST
0 200 400 600 800 1000 1200 1400
-4-202
Cu
rren
t (A
)
FUDS
0 100 200 300 400 500 600
-4-202
Cu
rren
t (A
)
US06
0 100 200 300 400 500 600 700 800 900 1000-2
-1
0
1
Time (s)
Cu
rren
t (A
)
BJDST
Alejandro Morrás
46
Figure 23: Dynamic stress test (DST) cycling from 80%SOC
Figure 24:Federal urban driving schedule (FUDS) cycling from 80%SOC
Alejandro Morrás
47
Figure 25: US06 highway driving schedule cycling from 80%SOC
Figure 26: Beijing dynamic stress test cycling from 80%SOC
Based on OCV tests data and battery dynamic charge/discharge tests data. Based on the
OCV-SOC mapping results, four battery models have been proposed. Then the battery
modeling and model parameters identification have been done using DST test data. FUDS test
data were used for model selection. Data of US06 and BJDST tests with different initial SOCs
were used to verify the robustness of selected model.
Alejandro Morrás
48
In order to investigate the effect of temperature on battery OCV-SOC mapping and the
accuracy of SOC estimation models, OCV tests and dynamic tests have to be done at room
temperature (25oC ±5oC), high temperature (45oC), and low temperature (0oC).
In addition, dynamic charge/discharge test at different temperatures is also conducted on
Panasonic NCR 18650-PF (PPPF) cells, Samsung SDI ICR 18650-28A (SE28) and Panasonic
NCR 18650-A (PEA). The test procedure is similar to the one that is described above:
1. The battery is charged to its full capacity, first 1C (CC) and when the upper cut off
voltage is reached the process of CV charge begins until a current of 0.01C is achieved.
2. The battery is discharged to 80%SOC (discharge of 20% SOC) because the current
profile for the dynamic charge/discharge test might overcharge the battery if it started
at 100%SOC. This discharge process is achieved using a discharge current of 0.5C for
the calculated time for the battery to reach 80%SOC.
3. When the battery is at 80%SOC the dynamic charge/discharge current profile is applied.
The current profile consists in the four current profiles plus a combination of the four
of them.
4. The dynamic charge/discharge current profile is applied until the battery reaches its
lower cut off voltage.
Alejandro Morrás
49
Figure 27: PEA-5 DST Example
With Figure 27 I want to show a full charge/discharge process. It begins with CC charge,
when the voltage gets to 4.2 V the CV charges continues charging the battery until the charging
current is <=0.01C. After that, the battery is discharged 20% of its SOC, so that we begin the
-6
-5
-4
-3
-2
-1
0
1
2
3
4
5
0 2000 4000 6000 8000 10000 12000
time (s)
PEA -5 DST Dynamic charge/discharge
Voltage(V)
Current(A)
2,5
2,7
2,9
3,1
3,3
3,5
3,7
3,9
4,1
4,3
4,5
0 5 10 15 20
Vo
ltag
e
Time (h)
Current Stress Test PEA-5
25ºC
Figure 28: Example of the voltage measurement as a result of the Current
Stress Test
DST FUDS BJDST DFUB US06
Alejandro Morrás
50
test at 80%SOC.Then, the battery is discharged using the charge/discharge DST current profile
until the lower cut off voltage is reached (in this case 2.5V). As it can be observed at Figure
28, the test continues with the different charge/discharge profiles.
Figure 28 shows a complete Current Stress voltage results for the PEA-5 battery at room
temperature.
Figure 29: PEA-5 Different temperatures Current Stress Tests
In Figure 29 a comparison between the same test done to the same battery but at different
temperatures is shown. As it is observed, there is not much of a difference between 25ºC and
0ºC but when we compare this last two with the test run at 45ºC the difference is bigger. The
test at 45ºC seems ‘shifted’ to the right; this is because applying the same current to the battery
at 45 ºC the voltage change is smaller. As a result, it takes more time for the battery to fully
discharge and reach the lower cut off voltage. The difference between the test at 0ºC and 25ºC
is smaller but there is a difference in the broadness of the voltage curve and also there is an
appreciable time difference.
From the analysis of Figure 29 it can be concluded that at high temperatures it takes more
time for the battery to fully discharge than at room temperature with the same current profile
applied. Moreover, at low temperatures it also takes more time for the battery to discharge but
the difference is very small.
2,5
2,7
2,9
3,1
3,3
3,5
3,7
3,9
4,1
4,3
4,5
0 5 10 15 20
Vo
ltag
e
Time (h)
Current Stress Tests PEA-5
25ºC
0ºC
45ºC
Alejandro Morrás
51
Current Stress test Comparison between High Energy and High Rate cells
Figure 30: Current Stress test Comparison between High Energy and High Rate cells
In Figure 30 it is shown the same Current Stress test applied to two different cells PEA-
5 and PPPF-2 that are high energy and high rate cells respectively. The main difference between
them is clear when looking at Figure 30, having both a proportional current going through each
of them the voltage change for PPPF-2 was much smaller than the voltage change of PEA-5, as
a result the full discharge process takes more time for PPPF-2 than for PEA-5.
2,5
2,7
2,9
3,1
3,3
3,5
3,7
3,9
4,1
4,3
4,5
0 20000 40000 60000 80000
Vo
ltag
e
Time (seconds)
Current Stresss test
PEA-5
PPPF-2
Alejandro Morrás
52
4 Application of typical R+RC battery model for all the cells
tested
The model selected for the study is widely used by the industry, this is the R+RC (first
order), that is displayed in Figure 31, using Matlab as the tool to build and verify the model.
This model has been used before to model Lithium-phosphate cells (6), so it is tested to see if
it works with other chemistries.
R0 is the ohmic resistance, Rd and Cp are the resistance and capacitance associated with the
polarization voltage (Ud). I is the current through the battery, when it is discharging is positive
and when charging is negative.
Following the model that is used in (6) and a similar is used in (7), the set of equations that
are used to model the battery are:
𝑑 =
1
𝐶𝑝−
𝑈𝑑𝐶𝑝𝑅𝑑
𝑈𝑡 = 𝑂𝐶𝑉(𝑆𝑂𝐶) − 𝐼 × 𝑅0 − 𝑈𝑑
(1)
Moreover, an Uscented Kalman Filter (UKF) has been used so that a better estimation of
the SOC can be obtained. The state function is:
Figure 31: Proposed R+RC battery model (6)
Cp
Alejandro Morrás
53
𝑆𝑂𝐶𝑡−1 = 𝑆𝑂𝐶𝑡 + 𝐼𝑡 (−
1
𝐶𝑁) + 𝑘𝑆𝑂𝐶,𝑡
𝑈𝑑𝑡−1 = 𝑈𝑑𝑡𝑒−
1𝑅𝑃𝐶𝑝 + 𝐼𝑡𝑅𝑝 (1 − 𝑒
−1
𝑅𝑃𝐶𝑝) + 𝑘𝑈𝑑,𝑡 (2)
𝑅0𝑡−1 = 𝑅0𝑡 + 𝑘𝑅0,𝑡
CN is the nominal capacity (not the one measured). Measurement function as in (6) is:
𝑈𝑡,𝑡 = 𝑂𝐶𝑉(𝑆𝑂𝐶𝑡) − 𝐼𝑡 × 𝑅0,𝑡 − 𝑈𝑑,𝑡 + 𝞷𝒕
The battery will be tested using dynamic stress tests (using current as input) to identify the
model parameters and an incremental SOC test (described before) that will estimate the relation
between the SOC and OCV for the different types of battery, so that the model can be checked
and the error calculated. The DST is the one being used for the model estimation while the
others are to verify the model robustness. However the model is just tested using the DST,
because these are preliminary results, it will be tested with the other profiles in the future.
All test are done at different temperatures (0C, 25C and 45C) to compare the behavior of
the cell and the model under various conditions.
The scope of this project just achieved to model and prove the batteries without taking into
account the temperature in the model’s parameters.
Table 4: Model Results summary for data at 25oC
In table 4 the results for the modelling for all the batteries at 25C are shown. There is not a clear
pattern in the error. However, there are many differences between the high rate cells and the
Model Results
Battery Type R0 Rp Cp MAE RMSE
PEA-5 0.16597 0.1976 1977.80 0.0012 0.0053
PPPF-2 0.0412 0.1801 1024.8 7.5789e-04 0.0028
SE28-5 0.1818 0.2409 1347.4 7.4423e-04 0.0030
SP20-2 0.1045 0.1776 797.1160 0.0010 0.0037
Alejandro Morrás
54
high energy cells. The first ones (SP20 and PPPF) have significantly less resistance and
capacitor values because, at it has been shown in the dynamic stress test, to generate the same
voltage difference in a high rate cell than in a high energy cell you need a bigger current. As a
result, as the voltage difference is similar and the current needs to be bigger, the resistance has
a smaller value for the high rate cells.
Table 5: PEA-5 model results at different temperatures
Model Results
Temperature R0 Rp Cp MAE RMSE
0oC 0.1658 0.1526 499.07 0.0022 0.0074
25oC 0.16597 0.1976 1977.80 0.0012 0.0053
45oC 0.0987 0.2442 1641.60 0.0016 0.000525
Table 5 shows the results for the characterization of a PEA type cell, with the test done at
different temperatures. The results show than the average error is smaller at 25C while the
RMSE is smaller at 45C, this means that there are sporadic bigger errors at 25C. RMSE focuses
more on penalizing the big errors than on showing the error trend. Overall the model at 25C is
slightly better than at the other temperatures.
2,5
2,7
2,9
3,1
3,3
3,5
3,7
3,9
4,1
4,3
12
85
56
98
53
11
37
14
21
17
05
19
89
22
73
25
57
28
41
31
25
34
09
36
93
39
77
42
61
45
45
48
29
51
13
53
97
56
81
59
65
62
49
65
33
68
17
71
01
73
85
76
69
79
53
PEA -5 at 25oC
Model Voltage estimation Voltage measured
Figure 32: PEA-5 at 25oC
Alejandro Morrás
55
Figures 32 shows the plot of the data estimated by the model versus the data measured.
Apparently it looks quite good but analyzing the absolute error, plotted in figure 33, it can be
observed that there are significant differences between them.
Table: PPPF-2 model results at different temperatures
Model Results
Temperature R0 Rp Cp MAE RMSE
25oC 0.0412 0.1801 1024.8 7.5789e-04 0.0028
45oC 0.0313 0.1832 1378.7 6.0349e-04 0.0020
Analyzing the data in table 2, it can be concluded that the model at 45oC is better than
at 25oC, both MAE and RMSE are smaller for the model at 45oC. The RMSE error is still bigger
than the value for SE28 at 45C.
-0,08
-0,06
-0,04
-0,02
0
0,02
0,04
0,06
0,08
0,1
0 1000 2000 3000 4000 5000 6000 7000 8000 9000
Absolute Voltage Error
Figure 33: Absolute Error of the estimated voltage at 25oC
Alejandro Morrás
56
Looking at figure 34 the model, seems to work perfectly. On the other hand, taking a closer
look with the help of the absolute error, figure 35 shows that when the battery is near the point
of full discharge the errors suddenly increases dramatically. This is an observation that suggests
further research on this behavior in the future.
2,95
3,15
3,35
3,55
3,75
3,95
4,15
1
40
9
81
7
12
25
16
33
20
41
24
49
28
57
32
65
36
73
40
81
44
89
48
97
53
05
57
13
61
21
65
29
69
37
73
45
77
53
81
61
85
69
89
77
93
85
97
93
10
20
1
10
60
9
11
01
7
PPPF-2 at 25oC
Model Voltage Estimated Voltage Measured
-0,10
-0,05
0,00
0,05
0,10
0,15
0 2000 4000 6000 8000 10000 12000
Absolute Voltage Error
Figure 34: PPPF-2 at 25oC
Figure 35: Absolute Error of the estimated voltage at 25oC
Alejandro Morrás
57
5 Comparison of the three tests done to the same battery,
degradation study
To complete the full study of the Lithium-ion batteries, a degradation study is done, here
with SP20 type.
In order to see the degradation of a battery we have to look at the capacity and how does
it vary form one test to another. In the next table we can find the calculated capacity of the
battery for three consecutive tests:
Table 6: Data for degradation study
%SOC
Charge Discharge
Test 1 Test 2 Test 3 Test 1 Test 2 Test 3
100,0000 100,0000 100,0000 100,00000 100,00000 100,00000
87,82000 88,15073 84,97360 90,00347 90,00422 89,99987
79,89000 79,89171 74,97556 80,01007 80,00858 80,00036
69,97026 69,93695 64,97758 70,01839 70,01181 69,99945
59,97290 59,94381 54,97955 60,04942 60,01672 59,99848
49,97786 49,94862 44,98150 50,05568 50,02093 49,99756
39,97961 39,95517 34,98208 40,06280 40,02584 39,99762
29,98467 29,95652 24,98312 30,06654 30,03126 29,99657
19,99156 19,95942 14,98391 20,07105 20,03837 19,99533
9,99803 10,00528 4,98505 10,07470 10,04494 9,99544
3,72452 1,03810 0,04985 3,72452 1,13815 1,93481
The percentage of SOC is calculated from the values of the current during the charge and
discharge. Test 3 was performed with different equipment than test 1 and 2. There are minor
differences caused by this change. Usually a battery doesn’t degrade until it has completed more
than 100 test. Here this theory is true, with just three test we cannot observe big differences in
the capacity from one test to next one.
Alejandro Morrás
58
6 Cycling Life Test with SEM Observation
Battery’s degradation is contributed by the degradation of its internal components, such
as the two electrodes. Changes in the electrodes (i.e. cathode, and anode) during battery cycling
should be studied to investigate the degradation mechanism. The electrode can be inspected via
scanning electron microscope (SEM) once it is disassembled from the battery cell. Then the
changes in the electrodes can be studied by the comparison of cell electrodes at different
degradation levels.
However, cell disassembly is a destructive test since further electrical or structural tests
are difficult to perform once the cell is opened. Therefore, more than one sample is required for
cycling life test with SEM observation.
In this work, three Samsung ICR18650-28A cells and three Panasonic NCR18650-A
cells are used to do disassembly test and SEM observation. Taking SE28 cells as an example,
the experiments are conducted by the following steps:
1) Impedance tests for SE28-1, SE28-2and SE28-3.
2) Cell disassembly test for SE28-1.
3) SEM observation and potential measurement for each electrode of SE28-1.
4) Cycling life test for SE28-2and SE28-3.
5) Impedance test for SE28-2 once it finishes the 500th cycle.
6) Cell disassembly test for SE28-2.
7) SEM observation and potential measurement for each electrode of SE28-2.
8) Impedance test for SE28-3 once it finishes the 1000th cycle.
9) Cell disassembly test for SE28-3.
10) SEM observation and potential measurement for each electrode of SE28-3.
Cycling process consists CCCV charging using 1C current with 0.05C termination
point, 10 minutes rest, and CC discharging using 1C current.
It has to be remarked that the cycling was done, due to the equipment limitations, 24
cycles each time and then the battery rested for some time until the test was done again, this
might have an effect in the capacity recording. Moreover, the SEM observation will be done in
the future when the test are finished.
Alejandro Morrás
59
In addition, in order to calculate the capacity of the battery after each cycle a member
of the team (Fangdang Zheng) suggested that there was an error when calculating the capacity
just as ‘current x time(in hours)’. The Figure 36 illustrates the problem with the calculations:
It has been taken one discharge cycle to show the correction:
The problems comes from the data acquisition, there were so many data to be stored that
the decision of just taking a measurement every 60 seconds was taken. Taking the
measurements every second would have result in a problem when analyzing the data as the .xlsx
would have so many measurements that the computer will have problems to open the file and
to calculate from them the capacity. Here is the formula that was used at first:
Capacity (Ah) = (I1– I0)*Time(s)* 3600(s/h)
The following formulas were used in order to calculate a better estimation of the
capacity at that point:
For the beginning and the end of the Discharge:
Capacity at point 1 (Ah) =I previous+ (I1– I0)/2*Time(s)* 3600(s/h)
For all the other points (not the beginning nor the end):
-3,5
-3
-2,5
-2
-1,5
-1
-0,5
0
0 500 1000 1500 2000 2500 3000 3500
Cu
rren
t (A
)
Time (s)
Current discharge example
Current (A)
Figure 36: Capacity calculus explanation using a current discharge
example
I 0
I 1 I 2 I 3
Alejandro Morrás
60
+ Capacity at point 3 (Ah) = I previous+ (I3– I2)/2*Time(s)* 3600(s/h) +
+I3*Time(s)* 3600(s/h)
The correction of adding the green triangle to calculate the middle points was done
because there are some points in the discharge process where the current varied from the value
measured in the prior point, so by just calculating the ‘square’ a higher capacity than the real
could have been added.
The results are presented as follows:
6.1 Impedance measurement
In Figure 37 it can be found a typical impedance measurement done to the battery
PEA-2. The data collected clearly describe the diffusion in active electrode material and
electrolyte function while showing the typical charge transfer reaction circumference
(described in the Initial Measurements section).
0
0,001
0,002
0,003
0,004
0,005
0,006
0,007
0,008
0,009
0,01
0,07 0,08 0,09 0,1 0,11 0,12
Imag
inar
y Im
ped
ance
(Ω)
Real Impedance (Ω)
PEA-2 Impedance Measurement
0 Cycles
Figure 37: Impedance measurement example
Alejandro Morrás
61
Moreover, as at the beginning of the test the impedance was measured another
measurement was taken when a big change in the capacity was noticed getting to the plot in
Figure 38.
As we can appreciate in Figure 38, the impedance has changed radically from 0cycles
to 100 cycles, especially in the charge transfer resistance (remember that is the diameter of the
imaginary circle). Also there is an uncommon feature in this Figure 38, in its first part, we
cannot clearly identify the part that shows the diffusion in active electrode material and
electrolyte, this might be because the battery reaches the battery cycling limit (when it degrades
up to failure). The measurement was taken at 307 cycles when it was considered that the battery
was not working anymore.
Here are presented the results for the continued battery cycling:
0
0,02
0,04
0,06
0,08
0,1
0,12
0,14
0,16
0 0,1 0,2 0,3 0,4
Imag
inar
y Im
ped
ance
(Ω)
Real Impedance (Ω)
PEA-2 Impedance Comparison
0 Cycles
100 cycles
307 cycles
Figure 38: PEA-2 Impedance comparison
Increasing Frequency
Alejandro Morrás
62
Figure 39 PEA-2 Cycling Results
From Figure 39 it should be mentioned that between cycles 69 and 70 there is a big drop
in the capacity of the battery, this might be a consequence of the fast degradation of the battery
with the specific test procedure.
Figure 40: PEA-3 Cycling Results
0,00
0,50
1,00
1,50
2,00
2,50
3,00
0 50 100 150 200 250 300 350 400
Cap
acit
y (A
h)
Number of cycles
PEA-3
0,00
0,50
1,00
1,50
2,00
2,50
3,00
0 50 100 150 200 250 300
Cap
acit
y (A
h)
Number of cycles
PEA-2
Alejandro Morrás
63
Figure 40 shows the degradation of PEA-3 that is a battery with the same specifications
and chemistry than PEA-2 (Figure 39), the battery continues to work for many more cycle.
There is no big drop in the capacity of this battery but the final result is similar but a 150 cycles
later.
0,00
0,50
1,00
1,50
2,00
2,50
3,00
0 50 100 150 200 250 300 350 400
Cap
acit
y (A
h)
Number of cycles
SE28-2
0,00
0,50
1,00
1,50
2,00
2,50
3,00
0 50 100 150 200 250 300 350 400
Cap
acit
y (A
h)
Number of cycles
SE28-3
Figure 41: SE28-2 Cycling Results
Figure 42: SE28-3 Cycling Results
Alejandro Morrás
64
Figures 41 and 42 show the plot of batteries SE28-2 and SE28-3, both seem to have a
similar degradation pattern, maybe SE28-3 has less capacity at the final cycle. The performance
of these batteries in the test is clearly better than the performance of PEA batteries.
Alejandro Morrás
65
7 Conclusion and future research to be done
After all the test that have been performed on the batteries some conclusions can be
obtained. The main difference between high rate cells and high energy cells is the impact that
the same current has on each of them, it will discharge more a high energy than a high rate cell.
In addition, the values of the resistors and capacitors for high rate cells is smaller, as the current
needs to be bigger than in a high energy cell to produce the same voltage difference.
The temperature is an important factor because the battery’s capacity changes with
temperature so the batteries are discharged faster with low temperatures and high temperatures
than at room temperature (25C approx.). Furthermore, the model works better with high rate
than with high energy batteries, also at high temperature than at room or low temperatures.
The Life cycling test results in the unexpected failure of both PEA cells, showing an
interesting pattern when it first reaches a cycle where the capacity is near 0Ah. The SE28 test
are still ongoing but the data available show a trend of capacity reduction. At CALCE, after
finishing the cycling test, there will be a destructive inspection with the SEM (scanning electron
microscope), to see the differences between the batteries with different number of cycles done.
For this report, it has been done a labor of deep research of the possible tests to
characterize batteries’ demeanor. There is a comparison between the different types of batteries
have been subjected to the tests and also between the data of the same test done at different
temperatures. The model is made for all the different data collected and has been chosen for its
relation between simplicity and results (very good results with a relatively simple model).
The models should be tested in the future with the other current profile tests. More
research should be done to make this model better predict the behavior of batteries at different
temperatures for all the different chemistries. The model used works fine for the different
chemistries, but to improve the model, the test and project should be more focused in just one
type of cells. By specializing the model, the BMS can better protect and manage the battery.
There is also more work to be done on the degradation test, the test should be continued,
finished and a comparison between the SEM observations of the different samples should be
done.
Alejandro Morrás
66
8 References and Bibliography:
(1)William G.Howard, Craig L. Schmidt, Erik R. Scott. Lithium-ion battery. Available at
http://www.google.com/patents/US7563541
(2)Isidor Buchmann. Basics about Batteries (CEO and Founder). Available at
http://batteryuniversity.com/learn/article/bu_1502_basics_about_batteries
(2*) In source (2) mentioned to be courtesy of The Boston Consulting Group
(2**) In source (2) mentioned to be courtesy of Cadex
(2) Caiping Zhang, Jiuchun Jiang, Weige Zhang and Suleiman M. Sharkh. Estimation of State
of Charge of Lithium-Ion Batteries Used in HEV Using Robust Extended Kalman Filtering. In
energies ISSN 1996-1073 published: 19 April 2012
(3)Available at http://www.calce.umd.edu/
(4)Tarun Huria, Massimo Ceraolo, Javier Gazzarri, Robyn Jackey. High Fidelity Electrical
Model with Thermal Dependence for Characterization and Simulation of High Power Lithium
Battery Cells.
(5)USABC Manual Revision 2. Available at
https://avt.inl.gov/sites/default/files/pdf/battery/usabc_manual_rev2.pdf
(6) Yinjiao Xing and Kwok-Leung Tsui. State of Charge Estimation for Lithium-Ion Batteries
Using a Temperature-Based Equivalent Circuit Model. Available at
http://www.sciencedirect.com/
(7)Caiping Zhang, Jiazhong Liu, S.M.Sharkh, Chengning Zhang. Identification of Dynamic
Model Parameters for Lithium-Ion Batteries used in Hybrid Electric Vehicles. National
Engineering Laboratory for Electric vehicle, School of Mechanical and Vehicular Engineering,
Beijing Institute ofTechnology, Beijing 1 00081 , China;
W.Y. Low J.A. Aziz , N.R.N. Idris , R. Saidur Electrical model to predict current voltage
behaviours of lithium ferro phosphate batteries using a transient response correction method.
In Journal of Power Sources. Available online 23 August 2012
Sijia Liu, Jiuchun Jiang, Senior Member, IEEE, Wei Shi, Zeyu Ma, Le Yi Wang, Fellow, IEEE,
and Hongyu Guo. Butler-Volmer Equation Based Electrical Model for High-Power Lithium
Titanate Batteries Used in Electric Vehicles. Available on IEEE TRANSACTIONS ON
INDUSTRIAL ELECTRONICS
W.Y. Low a, J.A. Aziz a,*, N.R.N. Idris a, R. Saidur Electrical model to predict current voltage
behaviours of lithium ferro phosphate batteries using a transient response correction method.
Available on www.elsevier.com/locate/jpowsour
Alejandro Morrás
67
I.J.Fernández, C.F.Calvillo, A. Sánchez-Miralles, J. Boal. Capacity fade and ageing models for
electric batteries and optimal charging strategy for electric vehicles. Available on
www.elservier.com/locate/energy
Xiaosong Hua, Shengbo Li, Huei Penga. A comparative study of equivalent circuit models for
Li-ion batteries. Available at: www.elsevier.com/locate/jpowsour
Fei Feng , Rengui Lu, Guo Wei and Chunbo Zhu. Online Estimation of Model Parameters and
State of Charge of LiFePO4 Batteries Using a Novel Open-Circuit Voltage at Various Ambient
Temperatures. Available at Energies 2015, 8, 2950-2976; doi: 10.3390/en8042950
Seongjun Lee, Jonghoon Kim, Jaemoon Lee, B.H. Cho. State-of-charge and capacity estimation
of lithium-ion battery using a new open-circuit voltage versus state-of-charge. Available at:
www.elsevier.com/locate/jpowsour
Eric A. Wan and Rudolph van der Merwe. The Unscented Kalman Filter for Nonlinear
Estimation. Available at https://www.seas.harvard.edu/courses/cs281/papers/unscented.pdf
Suleiman Abu-Sharkh, Dennis Doerffel. Rapid test and non-linear model characterisation of
solid-state lithium-ion batteries. Available at file:///C:/Users/Alex/Downloads/rapid-test.pdf
T. Stockley, K. Thanapalan, M. Bowkett & J. Williams. Design and implementation of an open
circuit voltage prediction mechanism for lithium-ion battery systems. Available at
http://www.tandfonline.com/loi/tssc20
Sandra Patricia Castaño Solís. Modelado y caracterización funcional en régimen dinámico de
sistemas electroquímicos de almacenamiento de energía. Aplicación a supercondensadores y
baterías de iones de litio. Available at http://e-
archivo.uc3m.es/bitstream/handle/10016/19281/tesis_sandra-
patricia_castano_solis_2014.pdf?sequence=1
Qishui Zhonga,b , Bo Huangb, Jianhua Mac , Hui Lia. Experimental study on relationship
between SOC and OCV of lithium-ion batteries. Available at
http://www.ijsgce.com/uploadfile/2015/0824/20150824072600339.pdf
http://batteryuniversity.com/
Battery specification are available at:
INR 18650-20R: http://www.avacom.cz/Datasheety/Samsung/INR18650-20R.pdf
ICR 18650-28: http://www.meircell.co.il/files/Samsung%20ICR18650-28A.pdf
NCR 18650-PF: http://industrial.panasonic.com/cdbs/www-
data/pdf2/ACA4000/ACA4000CE240.pdf
NCR 18650-A: http://industrial.panasonic.com/cdbs/www-
data/pdf2/ACI4000/ACI4000CE25.pdf