ANALYSIS AND DEVELOPMENT OF ENERGY STORAGE SYSTEM MODELS FOR SYSTEM OPTIMIZATION

Transcription

1 UNIVERSIDAD PONTIFICIA COMILLAS ESCUELA TÉCNICA SUPERIOR DE INGENIERÍA (ICAI) INGENIERO INDUSTRIAL FINAL DEGREE PROJECT ANALYSIS AND DEVELOPMENT OF ENERGY STORAGE SYSTEM MODELS FOR SYSTEM OPTIMIZATION AUTHOR: DIRECTOR: Isaac Prada y Nogueira MADRID, May 2012

2 AUTORIZACIÓN PARA LA DIGITALIZACIÓN, DEPÓSITO Y DIVULGACIÓN EN ACCESO ABIERTO DE DOCUMENTACIÓN 1º. Declaración de la autoría y acreditación de la misma. El autor D., como alumno de la UNIVERSIDAD PONTIFICIA COMILLAS (COMILLAS), DECLARA que es el titular de los derechos de propiedad intelectual, objeto de la presente cesión, en relación con el proyecto de fin de carrera Analysis and Development of Energy Storage System Models for System Optimization, que ésta es una obra original, y que ostenta la condición de autor en el sentido que otorga la Ley de Propiedad Intelectual como titular único o cotitular de la obra. En caso de ser cotitular, el autor (firmante) declara asimismo que cuenta con el consentimiento de los restantes titulares para hacer la presente cesión. En caso de previa cesión a terceros de derechos de explotación de la obra, el autor declara que tiene la oportuna autorización de dichos titulares de derechos a los fines de esta cesión o bien que retiene la facultad de ceder estos derechos en la forma prevista en la presente cesión y así lo acredita. 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 y hacer posible su utilización de forma libre y gratuita ( con las limitaciones que más adelante se detallan) por todos los usuarios del repositorio y del portal e-ciencia, 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, 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. Sin perjuicio de la titularidad de la obra, que sigue correspondiendo a su autor, la cesión de derechos contemplada en esta licencia, el repositorio institucional podrá: (a) Transformarla para adaptarla a cualquier tecnología susceptible de incorporarla a internet; realizar adaptaciones para hacer posible la utilización de la obra en formatos electrónicos, así como incorporar metadatos para realizar el registro de la obra e incorporar marcas de agua o cualquier otro sistema de seguridad o de protección. 1

3 (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 y ponerla a disposición del público a través de un archivo abierto institucional, accesible de modo libre y gratuito a través de internet. (d) Distribuir copias electrónicas de la obra a los usuarios en un soporte digital. 4º. Derechos del autor. El autor, en tanto que titular de una obra que cede con carácter no exclusivo a la Universidad por medio de su registro en el Repositorio Institucional tiene derecho a: a) A que la Universidad identifique claramente su nombre como el autor o propietario de los derechos del documento. 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. A tal fin deberá ponerse en contacto con el vicerrector/a de investigación (curiarte@rec.upcomillas.es). d) Autorizar expresamente a COMILLAS para, en su caso, realizar los trámites necesarios para la obtención del ISBN. 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 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. 2

4 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: a) Deberes del repositorio Institucional: - 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 exclusiva 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. b) Derechos que se reserva el Repositorio institucional respecto de las obras en él registradas: - retirar la obra, previa notificación al autor, en supuestos suficientemente justificados, o en caso de reclamaciones de terceros. Madrid, a 29 de Mayo de 2012 ACEPTA 3

5 Proyecto realizado por el alumno: Fdo.: Fecha: / / Autorizada la entrega del proyeto cuya información no es de carácter confidencial EL DIRECTOR DEL PROYECTO Isaac Prada y Nogueira Fdo.: Fecha: / / V. B. DEL COORDINADOR DE PROYECTOS José Ignacio Linares Hurtado Fdo.: Fecha: / /

6

7 UNIVERSIDAD PONTIFICIA COMILLAS ESCUELA TÉCNICA SUPERIOR DE INGENIERÍA (ICAI) INGENIERO INDUSTRIAL FINAL DEGREE PROJECT ANALYSIS AND DEVELOPMENT OF ENERGY STORAGE SYSTEM MODELS FOR SYSTEM OPTIMIZATION AUTHOR: DIRECTOR: Isaac Prada y Nogueira MADRID, May 2012

8 IV

9 Resumen El aumento tanto de la demanda de energía como de la penetración de las fuentes de energía impredecibles e incontrolables, como la eólica, hace necesarios nuevos métodos más efectivos para almacenar la energía. Muchos sistemas de almacenamiento de energía (SAE) están actualmente en desarrollo, pero una forma eficaz de almacenar la energía debe ir más allá ser capaz de contener la energía para que se pueda emplear posteriormente. También es necesario ser capaz de gestionar adecuadamente dicha energía, tanto para minimizar los costes de instalación y operación como para maximizar el valor de la energía entregada, y así maximizar los beneficios. Las distintas tecnologías de almacenamiento de energía, como condensadores, baterías y volantes de inercia, ofrecen ventajas e inconvenientes diferentes, por lo que para optimizar su rendimiento se deben combinar diversas tecnologías que permitan al sistema cumplir con tantos requisitos como sea necesario. Las típicas herramientas de gestión de SAE son los BMS y DSS. Los BMS (Battery Management Systems) son dispositivos que controlan los procesos de carga y descarga de las baterías para mantener su voltaje, corriente, temperatura y carga dentro de los límites convenientes. Funcionan controlando que esas variables se mantengan dentro de sus valores máximos y mínimos, pero no son capaces de optimizar el funcionamiento de la batería dentro de dichos rangos. Los DSS (Decision Support Systems) son sistemas informáticos destinados a ayudar a los gestores en la toma de decisiones. Los DSS se utilizan en varios campos relacionados con el comercio, simulando diferentes escenarios y mostrando las consecuencias de las distintas opciones que se puedan tomar, para ayudar a tomar las decisiones correctas. Uno de los DSS más populares para la industria de la energía es Hatch Vista DSS, utilizado para la gestión de centrales hidroeléctricas. La limitación de los actuales DSS destinados a aplicaciones energéticas es que están diseñados para operar con una única tecnología de SAE y orientados a aplicaciones de gran escala, tales como centrales hidroeléctricas de bombeo o plantas de energía de aire comprimido (CAES). Este proyecto pretende establecer la base para el desarrollo de una herramienta de optimización destinada V

10 RESUMEN a operar tanto en grandes instalaciones como instalaciones de menor tamaño y con capacidad para gestionar SAE compuestos por varias tecnologías. Las principales tecnologías de SAE que serán estudiadas en este proyecto son las baterías de plomo ácido, baterías de iones de litio y SAE hidráulicos. Las baterías de plomo han sido seleccionadas porque constituyen una forma tradicional de almacenar la energía que a lo largo de los años se ha ido perfeccionando. A pesar de que son muy pesadas, por lo que ofrecen baja potencia y energía específica, resultan mucho más baratas que otras tecnologías de almacenamiento de energía, por lo que aún son muy utilizadas, especialmente en aplicaciones donde no se requiere mover el SAE. Otra de las ventajas de las baterías de plomo-ácido es que son capaces de soportar condiciones de funcionamiento irregulares, por lo que no requieren complejos controladores o BMS. Las baterías de plomo se pueden clasificar en dos tipos: VRLA y ventiladas. El electrolito en baterías ventiladas está abierto a la atmósfera y puede liberar hidrógeno y oxígeno cuando la batería se está cargando. Estas baterías requieren un mantenimiento para rellenar el electrolito con agua y compensar esas pérdidas por gasificación. Las baterías VRLA, por el contrario, son baterías selladas que no requieren mantenimiento. Bajo condiciones de operación normales no liberan gases, y sólo debido a una temperatura excesiva u otras condiciones anormales la presión interna podría ser demasiado alta, haciendo que la válvula de seguridad abra. El estado de carga de baterías de plomo ácido se puede medir a través de la tensión en los bornes, la densidad del electrolito (si es accesible, no en VRLA) o la resistencia interna. Los principales métodos para medir el estado de salud son mediante una carga completa seguida por una descarga completa y por espectroscopia de impedancia. Su elevado peso hace que sean aptas sólo para aplicaciones fijas en las que no es necesario mover el SAE, como las fuentes de alimentación ininterrumpida, luces de emergencia y apoyo a sistemas solares; o las aplicaciones móviles en el que la autonomía no es especialmente importante (carretillas elevadoras, carros de golf...) Las baterías de litio se estudiarán en este proyecto porque, a pesar de que ya están en uso, se considera que tienen un largo desarrollo aún pendiente. Dicho desarrollo se espera que se centre en la mejora de la fiabilidad de las baterías, la reducción del peso y volumen (aumento de potencia y energía específica) y la reducción de sus costes, lo que llevará a la expansión de su uso. Hay dos tipos de baterías de litio: las baterías de litio convencionales y las baterías de litio polímero. La única diferencia entre ellas es que en las primeras el material del electrolito es una solución salina, mientras que en las segundas es un compuesto polimérico sólido. Las baterías de litio polímero tienen menor peso y mayor potencia y energía específica que las baterías convencionales, pero su funcionamiento y propiedades son muy similares. El proceso VI

11 RESUMEN fundamental que tiene lugar en estas baterías es la intercalación de átomos de litio del electrodo positivo (hecho de LiCoO 2 o similar) en el electrodo negativo (grafito). Durante la carga los electrones son forzados a moverse desde el electrodo positivo al electrodo negativo a través del cargador, mientas que los iones de litio se mueven a través del electrolito. Muchos autores consideran que este proceso es más próximo a la forma en que funcionan los condensadores que a una reacción REDOX típica de baterías, ya que consiste en el almacenamiento de la energía en un material no activo (grafito). Los métodos de medición del estado de carga y estado de salud en baterías de litio son similares al de los de baterías de plomo: la integración de la corriente de descarga, el voltaje de las celdas y la impedancia en el caso de la carga y la medición de la capacidad mediante carga y descarga y la espectroscopia de la impedancia (EIS) en el caso de la salud. Las baterías de litio son muy sensibles a la sobrecarga, sobre descarga, temperatura y las altas tasas de carga/descarga; por lo que requieren complejos sistemas para controlar sus procesos. Los BMS son elementos esenciales para estas baterías y frecuentemente están incluidos en los propios paquetes de celdas. Los principales modelos de baterías existentes son modelos electroquímicos y modelos de circuitos eléctricos equivalentes. Los modelos electroquímicos se basan en las reacciones electroquímicas que tienen lugar en la batería. Estos modelos proporcionan resultados muy precisos pero su configuración es complicada, ya que requieren parámetros que no están disponibles para usuarios en general, y el poder de computación necesario para ejecutar las simulaciones es muy alto. Los circuitos eléctricos equivalentes proporcionan una manera mucho más fácil de predecir el comportamiento de la batería que los modelos electroquímicos, pero su exactitud es menor y el número de efectos que pueden ser estudiados a través de ellos es limitado. La mayoría de los circuitos eléctricos equivalentes se basan en un conjunto de condensadores y resistencias. Los SAE hidráulicos funcionan almacenando la energía por medio de un fluido a presión. Hay otros sistemas (CAES y plantas hidroeléctricas) que también funcionan aplicando presión a un fluido, pero su forma de trabajar es completamente diferente y aunque son muy interesantes para compensar las diferencias entre demanda y producción de energía a gran escala, estarán fuera del alcance del proyecto actual, que se centrará en aplicaciones de menor escala. El dispositivo principal del SAE hidráulico que estamos proponiendo es el acumulador hidráulico. El acumulador hidráulico consiste en dos cámaras limitadas por barreras móviles. En una de las cámaras se almacena el fluido incompresible (aceite hidráulico) que se utiliza para intercambiar la energía con la aplicación; en la otra cámara hay un gas compresible utilizado para aportar la presión al fluido. El estado de la carga en los acumuladores hidráulicos se mide mediante la presión del sistema. La temperatura VII

12 RESUMEN del gas aumentará al ser comprimido y disminuirá al expandirse, por lo que si la presión es la única variable conocida habrá que realizar suposiciones adicionales sobre el intercambio de calor con el entorno. Dichas suposiciones suelen consistir en considerar la carga y descarga del acumulador como procesos adiabáticos o isotérmicos. Procesos adiabáticos son aquellos en los que el calor intercambiado es despreciable. Procesos isotérmicos son aquellos en los que la temperatura del gas es constante, es decir, se considera que son suficientemente lentos como para dar tiempo para que la temperatura del gas se iguale con la del entorno. El tiempo para tomar uno u otro caso depende de las características particulares del sistema considerado. La principal ventaja de los sistemas hidráulicos es su capacidad para resistir un funcionamiento irregular, y la sencillez de sus sistemas de gestión en comparación con la que requieren otros sistemas, como el BMS de las baterías de litio. También vale la pena mencionar los altos regímenes a los que pueden trabajar, lo que lleva a altas potencias específicas. Gracias a estas características, los SAE hidráulicos se pueden añadir a muchos sistemas hidráulicos ya existentes, que no estaban destinados inicialmente al almacenamiento de energía, sino a otros fines de transmisión de potencia. Su principal desventaja es el riesgo que implica el operar a alta presión. Los sistemas hidráulicos de almacenamiento de energía también ofrecen eficiencias energéticas mucho más bajas que otras tecnologías de almacenamiento de energía. Para validar los modelos de SAE estudiados se han construido dos bancos de pruebas. El primero tiene como objeto estudiar el proceso de descarga de las baterías de plomo ácido. Dicho banco consiste en una resistencia de descarga, el circuito de carga y los dispositivos de medición necesarios. Se ha visto que el uso de condensadores para modelar el proceso de descarga da lugar a errores significativos. Para resolver este problema, se propone una nueva función matemática que representa la reducción de iones en la zona de electrolito próxima a los electrodos, donde tiene lugar la reacción química. El segundo banco es un SAE hidráulico que consiste en un acumulador hidráulico de diafragma, una bomba de engranajes accionado por un motor de inducción de CA y el motor de engranajes que acciona un alternador eléctrico. Este banco se destina a estudiar la eficiencia energética de los diferentes componentes del sistema. La baja eficiencia que muestran el alternador y la transmisión de polea y correa desde el motor al alternador afecta significativamente a la medición de las otras variables del sistema. Al estudiar la autodescarga del acumulador, se ve cómo la compresión de los gases puede ser modelada como una composición de los procesos adiabáticos y isotérmica considerados anteriormente. VIII

13 Abstract Increase in energy demand and penetration of unpredictable and uncontrollable sources of energy such as wind power makes necessary effective methods to store energy. Many energy storage systems (ESS) are currently under development, but efficient ways to store energy go further than just being able to contain the energy to be delivered at a given time. It is also necessary to be able to manage properly said energy so system installation and operation costs are reduced and output energy value is maximized, in order to maximize total revenues. Different energy storage technologies, such as capacitors, batteries and flywheels, have different advantages and disadvantages. In order to optimize their performance ESSs should combine a number of technologies that permit to meet as many requirements as possible. Typical ESS managing tool are BMS and DSS. Battery Management Systems (BMS) are devices that control the charging and discharging processes of batteries to maintain their voltage, current, temperature and charge within the desirable ranges. They work by controlling that those variables don t excess their maximum or minimum values, but within the battery operating range they are not capable to optimize the battery performance. Decision Support Systems (DSS) are computer systems intended to assist managers on decision making. DSSs are used in several fields related to business, simulating different scenarios and showing the consequences of the different options that can be taken, giving assistance to make the right decisions. The most popular DSS for the energy industry is Hatch Vista DSS, used for hydroelectric power plant managing. Limitation of current DSSs is that they are implemented to operate with a single ESS type and especially designed for large scale applications, such as pumped hydroelectric power plants or compressed air (CAES) power plants. This project is setting the background for the development of an optimization tool intended for operating both large scale and small scale installations and with capability to manage ESSs made up of several energy storage technologies. ESS technologies studied in this project are lead acid batteries, lithium ion batteries and hydraulic ESSs. IX

14 ABSTRACT Lead acid batteries have been selected because they represent a traditional way to store energy that over the years has been improved. Although they are very heavy, and therefore they show low power to mass and energy to mass ratios, they result much cheaper than other energy storage technologies, so they are still very used in applications where no movable ESS is required. Another advantage of lead acid batteries is that they are capable of withstanding rough operation, so they don t require complex battery chargers or BMS as lithium ion batteries do. Lead acid batteries can be classified in two types: vented batteries and VRLA. Vented batteries electrolyte is open to atmosphere and hydrogen and oxygen gasses can be released while the battery is being charged. These batteries require maintenance to refill the electrolyte with water. VRLA by contrast are sealed batteries that don t require maintenance. Under normal operation they don t release gasses; only if due to excessive temperature or other abnormal conditions inner pressure is too high, relief valve will open. The State of Charge of lead acid batteries can be measured through the terminal voltage, electrolyte density (when it is accessible, not in VRLA) or the internal resistance. The State of Health can be measured by a complete charge followed by a complete discharge or by impedance spectroscopy. Their high weight makes them suitable only for stationary applications in which the displacement of the ESS is not necessary, such as uninterruptible power sources, emergency lighting and solar system backup, or movable applications in which range is not an important concern (forklift, golf carts... ) Lithium ion batteries are covered by this project because, although they are already in use, they are considered to have a long development still pending. Said development is considered to be focused on improve the battery reliability, reduce weight and volume (increase specific power and specific energy) and reduce their costs. There are two types of lithium batteries: the conventional lithium ion batteries and the polymer lithium ion batteries. The only difference between them is that for the first type the electrolyte material is a salt solution while for the second type it is a solid polymer composite. Polymer lithium ion batteries show lower weight, so higher specific power and specific energy, than conventional lithium ion batteries, but their performance and properties are very similar, so we will use the term lithium ion to refer to both types of batteries. The fundamental process that takes place in lithium ion battery is the intercalation of lithium atoms from the positive electrode (made of LiCoO 2 or similar) into the negative electrode (graphite). While electrons are forced to move from positive electrode to negative electrode when charging (and vice versa when discharging) through the electrical circuit, lithium ions move through the electrolyte. Many authors consider this process to be closer to the way capacitors work than a typical REDOX battery reaction, as it consists on storing the energy in a non-active material (graphite). Processes for State of Charge and State X

15 ABSTRACT of Health measurement for lithium ion batteries are similar than those for lead acid batteries: current integration, cell voltage and cell impedance for SoC; and capacity measurement and electrochemical impedance spectroscopy (EIS) for SoH. Lithium ion batteries are very sensitive to overcharge, over discharge, high temperature and high charge/discharge rates. They require complex systems to control their processes. BMSs are essential elements to work with these batteries and are usually included in the battery packs. Main existing battery models are electrochemical models and equivalent electrical circuits models. Electrochemical models are based on the electrochemical reactions that take place in the battery; these models provide accurate results but shaping them is complicated, as they require parameters that are no available for general users, and high computing power is required to run the simulations. Equivalent electrical circuits provide a much easier way to predict battery behavior than electrochemical models, but their accuracy is lower and the number of effects that can be studied through them is limited. Most equivalent electrical circuits are based on a set of capacitors and resistors. Hydraulic ESSs work storing energy by means of pressurized fluids. There are some other systems (CAES and hydroelectric) that also work by applying pressure to a fluid, but the way they work is completely different and; although both systems are very interesting for high scale energy demand and production compensation, they will be out of the scope of current project as we will be focused on smaller scale applications. Main device of the hydraulic ESS we are proposing is the hydraulic accumulator. Hydraulic accumulators consist of two chambers limited by movable barriers. In one of the chambers there is the incompressible fluid (hydraulic oil) used to exchange the energy with the application; in the other chamber there is a compressible gas used to pressurize the fluid. The State of Charge in hydraulic accumulators is measured though the system pressure. As the gas will increase its temperature when compressing and reduce it when expanding, additional assumptions about the heat exchange must be made if pressure is the only known variable. According to this statement, accumulator charge and discharge can be modeled as adiabatic or isothermal process. Adiabatic processes are those in which heat exchange is negligible, i.e. processes are fast enough so no significant heat rejection or absorption happens. Isothermal processes are those in which the gas temperature is constant, i.e. there is time enough during charge or discharge to reject heat or absorb heat until temperatures are homogenized. Time required to consider one or the other case depends on particular characteristics of the system under study. Main advantage of hydraulic systems is their ability to withstand rough operation, and the simplicity of their management systems compared to the one that other systems require, as the BMS of lithium ion batteries. It is also worth to mention XI

16 ABSTRACT the high charge and discharge rates at which they can work, leading to high power capacities. Thanks to these characteristics, hydraulic ESSs can be added to many already existing hydraulic systems, even though they were not initially intended for energy storage but for other power transmission purposes. Its main disadvantage is the hazard implicit in operating high pressure recipients. Hydraulic energy storage systems also show much lower charge/discharge cycle efficiency than other energy storage technologies. To validate the studied ESS models, two test benches have been built. The first one is intended to study the discharge process of lead acid batteries, consisting on a discharge load, the charging circuit and the required measurement devices. It has been seen that using capacitors to model del discharge process lead to significant errors. To solve this issue, it is proposed a new function that models the ion reduction in the electrolyte zone close to the electrodes, where the reaction takes place. The second one is a hydraulic ESS consisting of a hydraulic diaphragm accumulator, a gear pump driven by an AC induction motor and gear motor that drives an electric alternator. This bench is intended to study performance of the different components of the system. The low energy efficiency that shows the alternator and the pulley and belt transmission from the motor to the alternator affected significantly to the measurement of the other variables of the system. It is seen that selfdischarge process of the accumulator results to be a composition of the adiabatic and isothermal processes. XII

17 A mi familia T. S. R. Throughout space there is energy... It is a mere question of time when men will succeed in attaching their machinery to the very wheelwork of nature NIKOLA TESLA XIII

18 XIV

19 Acknowledgements The author wishes to express his gratitude to KeelWit Technology & Beyond S.L. for the opportunity to work and learn from them during this year. Special thanks to Isaac Prada y Nogueira and José María Cancer Aboitiz for their guidance during the course of this project and for the enthusiasm they have transmitted. XV

20 Contents I. Memory 1 1. Introduction Aim of the project Technologies covered by the project The necessity of energy storage Existing optimization tools Description of ESS General parameters Specific energy Specific power Energy efficiency Temperature State of Charge State of Health C Rate Lifetime Lead acid batteries Operating principle Battery structure and types Parameter measurements and estimations Battery models Maintenance and failure prevention Charge/discharge processes Advantages, disadvantages and applications Lithium batteries Operating principle Battery structure and types Parameter measurements and estimations Battery models Maintenance and failure prevention Charge/discharge processes XVI

21 CONTENTS Advantages, disadvantages and applications Hydraulic energy storage systems Operating principle Structure and types Parameter measurement and estimations System models Maintenance and failure prevention Charge/discharge processes Advantages, disadvantages and applications Optimization function Description of the function Model restrictions Additional constrains Size limitations ESS odometer Lead acid battery test Aim of the test Test bench design and components Measurements and data collection Results Hydraulic ESS test Aim of the test Test bench design and component Measurement and data collection Results Bibliography 89 XVII

22 List of Figures Flexibility sources on the electrical grid, [GOGG09] Ragone plot, [RAGO68] Lead acid electrolyte density variation Voltage variations on a 12V lead acid battery, [PERE93] KiBaM model representation Equivalent circuits for SoH estimation KiBaM model circuit representation Lead acid capacity fade Total cycles and equivalent cycles variation with DOD kWh deep cycle lead acid batteries life test Basic battery charger scheme Lithium batteries price evolution forecast Li + intercallation in LiCoO 2 and graphite electrodes, [ELEC11] Non linear circuit model for lithium ion batteries Battery performance according to Shepherd model Battery performance according to Unnewehr model Lithium ion battery charge Gas loaded accumulators [PARK03] Energy vs pressure during accumulator charge Example power scheme Inverse proportionality between num of cycles and DoD for lead acid battery Battery test bench Battery voltage variation during discharge Current variation during discharge Battery voltage variation during charge Current variation during charge Proposed circuit Voltage output from proposed circuit Voltage output from proposed circuit 1 with discharge end Voltage output from proposed circuit 2 with discharge end Proposed circuit XVIII

23 CONTENTS Hydraulic ESS scheme Electric motor torque-speed characteristic Alternator output vs equivalent engine speed Results of the electric motor and pump characterization test Results of the electrovalves characterization test Results of the hydraulic motor and alternator mechanical characterization test Results of the on-load test Results of the on-load test Results of the self-discharge test XIX

24 CONTENTS XX

25 L MEMORY

26

27 Chapter 1 Introduction The continuous increase in energy demand and its unpredictability both in consumption (due to user habits) and generation (due to the proliferation of renewable energies) makes necessary adequate managing systems, which require the storage of energy. Efficient energy storage will lead to more flexible energy systems, as production and generation don t have to be simultaneous. Also economical benefits will be achieved by optimizing the production, sale, purchase and consumption times. During the last years, many technologies that accentuate the necessity of new ways to store energy has been developed. Electric vehicles, renewable energy sources and electric smart grids are clear examples of this tendency. Also the increase of fossil fuel prices will be a key point for new technologies to be success, as the cost of the systems must be covered by the profits from trading and energy saving. A rise in the capacity of energy storage, keeping the equipment within reasonable cost, size and weight limits, will enable the expansion of existing technologies and the development of some new ones, as well as the implantation of new applications. In this project it will be developed a mathematical tool for optimizing the design and management of any energy storage system, selecting the most appropriate energy storing technologies that it should include, and determining optimum energy exchanges in terms of power and time, as well as adequate maintenance schedules and disposal times. Also different energy storage technologies will be studied to define their characteristics and evaluate their feasibility for energy storage applications. 3

28 1. INTRODUCTION 1.1. Aim of the project The aim of this project is to develop a tool for optimizing energy storage systems (ESS) that permits to determinate easily the best configuration of different energy storage technologies that should make up the system. It will also optimize the management of the energy flow, schedule maintenance operations and determine disposal time to achieve the best performance during the system life time, both in energy and economical terms. The parameters that have in common any energy storage technology will be defined, so different devices from different technologies can be arranged together to make up a single complex ESS 1.2. Technologies covered by the project This project will study in detail the parameters of lithium ion batteries, lead acid batteries and hydraulic systems as energy storage technologies. Further research will include other technologies such as flywheels, supercapacitors and different battery chemistries. However, common ESS parameters will be defined as generic as possible in order to be able to apply them to any kind of technology. We first want to focus on said technologies because we consider lithium ion batteries an incipient technology with much development still pending and a considerable future usage, lead acid batteries to be based on a mature technology that has successfully been used for a long time on several applications and it still is due to its relatively low costs, and hydraulic systems to be an already existing technology widely used for different purposes that we want to suggest for energy storage as innovative application. Further reasons to select these technologies will be given in the descriptions of each technology on chapter The necessity of energy storage Several authors discuss about the necessity of improved energy storage system for large scale applications. One of the areas in which they are normally focused is the proliferation of renewable energies, because they come from a variable and unpredictable source that cannot be controlled to adapt the generation to the demand as we can do with energy generated from fossil fuel combustion. 4

29 1. INTRODUCTION Figure Flexibility sources on the electrical grid, [GOGG09] On it Wind Power and Energy Storage report, the American Wind Energy Association affirm that wind power generation still can increase without energy storage if variability is compensated by more flexible power sources, such as natural gas, and adapting market demand by offering different electric rates depending on the demand time; and shows Denmark, Spain, Ireland, and Germany as examples of wind power penetration through market and flexible generation regulation. Smart grids are a good example of how demand flexibility will lead to a significant improvement in renewable generation and energy trading. AWEA also proposes on this article the wind power curtailment as a way to regulate electrical grid more economical than new systems of energy storage (figure 1.3.1). The main reason they provide against new generation energy storage is the high cost they represent, but they also acknowledge that flexibility from traditional sources will be tapped in the future. This evidences the necessity of new energy storage technologies as well as the improvement of the efficiency of existing technologies to reduce operating costs. [GOGG09] But further than enabling higher penetration of renewable energy sources into the electrical system, energy storage offers additional benefits such as peak load shifting, frequency regulation, and protection against unexpected outages or blackouts, both for large and small scale installations. The research that is going on to make the plug-in electric cars (EV and PHEV) able to provide energy to the grid when not in use, what is known as vehicle to grid services (V2G), in a good example of the increasing interest in distributed energy storage Furthermore, when dealing with large scale renewable energy production, benefits from energy storage go further than just energy saving or additional energy production. In this case, other benefits such as capacity firming, voltage stability and grid reliability must be considered, as well as environmental concerns. [QUAN08] The benefits from increase in renewable energy integration are easy to calculate in monetary terms as can be expressed just as the product of the additional energy 5

30 1. INTRODUCTION integrated multiplied by the market price of said energy. The other aspects of the benefits, such as the environmental benefit, are more complicate to expresses in monetary terms, as they lead to improvement in welfare that are not directly reflected on costs savings or economical profits. However, most authors assign an arbitrary economical value to all those benefits that cannot be measured as money profits, in terms of $ per year or similar. Therefore, the target of operating an ESS will be maximizing the total benefits, accounting for the energy integration increase, energy related additional profits, non economical benefits and, of course, system installation and operating costs Existing optimization tools There are two main products currently available for ESS management and optimization: BMSs and DSSs Battery Management Systems (BMS) are physical devices that control main parameters of electric batteries to ensure that they operate under safe conditions. Such parameters are commonly the input and output current, voltage and temperature. BMS stops the charging process if battery charge excess the maximum charge or the discharge process if charge falls bellow minimum charge. It also controls the rate at which energy is delivered or absorbed while the battery is operating. In addition, multi-cell batteries BMSs control every single cell variables to compensate the charge differences through internal energy exchanges among cells to avoid located damage that could extend to healthy cells causing the complete battery failure. Any way these systems are not intended to optimize the battery performance, but just to maintain the operating conditions within the design range to ensure safe operation. [STUA02] [PLET06] Decision Support System (DSS) are computer programs that assist on decision making by evaluating different scenarios and proposing the best one. They consist of a data base that provides the required information about the different options, a model to predict the consequences of each decision and a interface to let the user introduce the input information and read the results. DSSs are widely used on business environment to assist managers in their products and trade decisions. In the energy industry, DSSs are used to plan and schedule the energy production based on energy predicted consumption as well as energy trading estimating energy prices. An example of DSS applied to the energy sector is Hatch Vista DSS, intended to maximize revenues by optimizing the running of pumped hydroelectric plants. [YANG06] [RAMA05] 6

31 1. INTRODUCTION The system we pretend to develop goes further than those existing systems as it will include optimization tool both for the design and operation of the ESS 7

32 1. INTRODUCTION 8

33 Chapter 2 Description of ESS 2.1. General parameters Energy storage systems can be described through a set of general parameters that are common to any energy storage technology. This section will describe those parameters that this project will work with. Parameters cited on this section are only those considered the most relevant for ESSs study. Particular parameters of each ESS technology covered by this project will be described on their corresponding sections Specific energy The task of an ESS is to absorb certain amount of energy in a given time, either from the excess energy of the process (like regenerative braking) or by purchasing it at lower price (energy trade), so it can be delivered later. Therefore energy that can be stored is critical for describing the system and it is usually referred to as the system capacity. But capacity alone provides little information about the performance of a given energy storage technology, as it can be increased just by increasing the size of the ESS, by adding cells. To solve this issue, capacity if often expressed in terms for energy per mass unit, which is the specific energy. Similar expressions are defined for special applications, such as the autonomy or range (in distance units) for hybrid or electric vehicles, or the discharge time for a given load. 9

34 2. DESCRIPTION OF ESS Specific power Just as important as the total amount of energy that an ESS can store is the rate at which said energy can be delivered during discharge or absorbed during charge processes, it is, the maximum power of the system. But again, the absolute power of the ESS can be increased just by increasing the size of the system, so the specific power is expressed as the power per mass unit. Specific power and specific energy can be represented together on a chart named Ragone plot, after David V. Ragone first used it to compare different energy storage systems. ESS on Ragone plots are not represented by a single point, as the total energy available varies with the required power, but they are represented by a line or, more generally, a thick area. Ragone plots can also represent the desirable limits of the ESS we are looking for or expected discharge profiles. On Figure there is an example of Ragone plot comparing different battery technologies together with the time for constant discharge. These charts have been widely used on later papers to compare the characteristics of different ESSs. [RAGO68] Figure Ragone plot, [RAGO68] 10

35 2. DESCRIPTION OF ESS Energy efficiency The energy efficiency of an ESS is the ratio between output and input energy. The difference between input and output energy is due two types of losses: exchange losses and self-discharge. Exchange losses are those involved in energy transformation process, in this case, the charge and discharge processes. It is mainly due to the friction on moving part or fluids, or the electromagnetic properties of the materials. Knowing the losses is important not just to know the total energy available or the energy costs, but also to size auxiliary systems such as refrigeration of the ESS in those cases where they are necessary. Self-discharge is the phenomenon in which ESSs lose the energy the have stored even when there is no intentional energy exchange. Both terms account for the overall efficiency of the systems, but depending on the time between charge and discharge processes any of them could be negligible. Equation expresses overall efficiency in terms of exchange losses and selfdischarge. How this terms can be calculated will be discussed later in this project. η = W out W in = 1 W lost W in = 1 W ex losses + W self discharge W in (2.1.1) Temperature Temperature of the system affect to the internal processes of the ESS, especially in those depending on chemical reactions, as batteries, both because normal reactions rates vary (so the maximum power varies) and because new reactions may appear, causing cell aging or damage. The variation of the reaction rate with the temperature is modelled by the Arrhenius equation. Many battery management systems (BMS) control the performance of ESS to keep the temperature in an operating range, but optimize the performance of an ESS requires to monitor and control the temperature even within that operating range. The main challenge of an adequate temperature management system for the batteries is not just to control the overall temperature, but to control the temperature of each cell that make up the battery. Control of cell temperatures can be carried out just by controlling the energy exchanges or through complex heating and cooling systems, depending of the requirements of the application and the sophistication of the system. [PESA01] 11

36 2. DESCRIPTION OF ESS State of Charge The State of Charge (SoC) is a measurement of the actual energy stored in a system compared to the maximum energy it can store. Thus, it is not a characteristic of the system as specific energy or power are, but a measurement of the actual state. SoC can be calculated by integrating the power exchanged over the time, or by indirect measurement such as pressure, voltage, density... Particular methods will be studied for each technology, as each one has its own limitations (energy losses, measurement variations...) State of Health Similarly to the SoC, State of Health (SoH) is a measurement of the actual condition of the system. It measures the actual maximum energy that the system can store compared to that when it was new. An easy way to measure it is just by recharging the system until completely charged and then discharge it until fully discharged, but this method cannot be performed with the system in service, so many applications require to estimate it through any other parameters that change with age, as we will see later on C Rate Usually the charge or discharge rate at which an ESS (specially batteries) is operating is expressed on units relative to the system capacity. When dealing with batteries, capacity is given in ampere-hours (Ah). If a battery capacity is 80Ah, it means that said battery can provide the amount of charge equivalent to a current of 80A during 1 hour. It doesn t mean that the battery can actually provide 80A during 1h, as maximum power and capacity fade constrains must be considered, but total amount of charge (current integrated over time) under optimum conditions will be 80Ah. The C-rate is then defined as the quotient of the actual discharge current (expressed in amperes), divided by the battery capacity (in ampere-hour). Therefore, 1C is the current that fully discharges the total capacity of the battery in one our, 2C is twice the current of 1C, i. e. the current that fully discharges the total battery capacity in 30 minutes, as so on. I would like to point out again that running the battery at 1C doesn t mean it will be able to run during 1h, as current related capacity fade must be considered. 12

37 2. DESCRIPTION OF ESS When dealing with other systems than batteries, equivalent measurements can be done through the E-Rate. The E-Rate is analog to the C-Rate, but it is expressed in energy units instead of amperes Lifetime In scientific literature, depending on the author the term lifetime is used to refer either the time that a given ESS may run form fully charged to fully discharged, or the time that the system may work been charged and discharged (measured for example in number of cycles) until ESS can no longer be operated due to system aging. This term won t be used on this text to avoid confusion. Equivalent terms such as discharge time for the first meaning and number of cycles that the system lasts will be used instead. 13

38 2. DESCRIPTION OF ESS 2.2. Lead acid batteries Lead acid batteries, invented back in 1859 by Gaston Planté, are the oldest type of rechargeable battery. Despite the fact that it is an old technology and some new types of batteries have been developed since then, they are still one of the most commercially used batteries due to its lower costs with respect to the other types Operating principle Lead acid batteries are based on the chemical REDOX reaction represented by equation When fully charged the positive electrode of a single cell is composed of lead oxide (P bo 2 ) and the negative is metal lead (P b). During discharge positive electrode is reduced and negative electrode is oxidized both to lead sulfate (P bso 4 ) The electrolyte is a sulphuric acid aqueous solution, that provides both the bisulfate ions (HSO4 ) and the protons (H + ). P bo 2 + HSO 4 + 3H + disch. charge P bso 4 + 2H 2 O (2.2.1) P b + HSO 4 disch. charge P bso 4 + H + The reduction of P bo 2 (first line of equation 2.2.1) have a standard electrode potential of V. The oxidation of P b to P b 2+ have an electrode potential of V (second line) [DELL01]. Combining both values, the total open cell obtained for that battery is: V cell = ( 0.358) = 2.048V (2.2.2) Actual voltage of the cell may vary due to state of charge, state of health, discharge rate, temperature and electrode alloying elements, among other factors. Cell voltage on a battery can raise up 2.12 or 2.13 volts, while same battery may have a cell voltage of 1.85 Volts when fully discharged. This variations will be used to estimate the State of Charge of the battery as we will see on section However, the nominal voltage of lead acid batteries is considered to be 2V/cell. 14

39 2. DESCRIPTION OF ESS Battery structure and types Lead acid battery electrodes are made up of several porous plates of lead (negative terminal) and lead oxide (positive terminal). The porosity of these plates increases the contact surface, and therefore the reaction rate, so more power is available. As lead is too weak to support mechanical stress, antimony is added to strengthen the material, enabling deeper discharge cycles. Antimony content is commonly between 2% and 6%, higher values will increase the amount of hydrogen and oxygen released during the charge, so the consumption of water in the charge/discharge process will increase. Cadmium and strontium are sometimes used instead of antimony to strengthen the electrodes because they also reduce the self-discharge rate. Calcium can be used to reduce self-discharge only for shallow discharge batteries, as it reduces the depth of discharge the battery can support. [ODON05] [WILL07] Notice that according to equation 2.2.1, during charge process a net amount of protons is produced. Ideally all this protons should remain in the water to recombine with the bisulfate ions (HSO4 ) to restore the sulfuric acid consumed during the discharge, but during normal operation some of these protons are combined together producing hydrogen gas (H 2 ) and molecular oxygen (O 2 ) from the dissociated water ions. Lead acid batteries can be classified into two categories: vented batteries and VRLA batteries. Vented batteries (also called flooded) are those whose electrolyte is liquid and open to atmosphere, so the produced hydrogen is just released to atmosphere. Adequate ventilation is required in the recharge room and water must be added periodically to maintain the electrolyte level. By contrast, the VRLA are sealed batteries where no gasses can be released to atmosphere. Their name stands for Valve Regulated Lead-Acid as they are equipped with a security valve to maintain pressure under a safety limit that in normal use should remain closed. VRLA batteries can be further classified AGM or gel, depending on their electrolyte type. AGM (Absorbed Glass Mat) batteries have fiber glass mat in which the sulphuric acid solution is impregnated, while gel batteries have the electrolyte gelified by mixing it with silica dust. [MACC11] Parameter measurements and estimations State of Charge On applications in which discharge load is constant and known, the State of Charge of the battery can be calculated by measuring the discharge time. On applications where the discharge load is not known or constant, a similar method 15

40 2. DESCRIPTION OF ESS can be used by integrating current over time. The problem of these methods is that they are based on the difference between the initial energy of the battery and the energy released, but initial energy is not always known as partial charge and energy losses must be considered. Also capacity fade associated to the discharge rate must be accounted to obtain realistic estimations (see Peukert equation on section 2.2.4). For vented batteries, the SoC can be estimated through the electrolyte density. As the acid concentration varies with the charge (HSO4 ions react to form P bso 4 ) and the acid and water have different densities (1.8g/cm 3 for sulfuric acid), the electrolyte density will vary linearly with the charge that the battery contains (Figure 2.2.1). The density values might be altered by water evaporation and gas release during charge, but as acid doesn t evaporate and water is added periodically, the actual and theoretical density must be close enough to be a valid estimation if adequate maintenance is carried out. When electrolyte is not accessible for density measurement (VRLA batteries for example) the battery SoC can be estimated through the battery voltage. As explained on section 2.2.3, although theoretical open circuit cell voltage is volts, the actual voltage varies with the SoC, temperature and current. Temperature affects to the speed of reaction (higher temperature means higher rate, and therefore, higher power), so for adequate estimation of SoC, voltage data must be referred to a specific temperature. Voltage variations from current are caused by the electrical resistance of the battery components. Batteries can me modeled through a set of capacitors and resistances that must account both the effect of the metallic conductors losses and the electrolyte charge transfer resistance. To reduce the effect of said resistance, the voltage should be measured on open circuit. If the measurement of the open circuit voltage is not possible (because batteries cannot be disconnected), SoC can be estimated through the resistance of the equivalent Thévenin circuit. Notice that at low SoC, battery will show low resistance for charging and high resistance for discharging, while at high SoC levels, battery will have high resistance for charging and low resistance for discharging; so the battery, at a given SoC level, will have different values of resistance for charging and discharging processes. This difference is shown on figure 2.2.2, for a 6 cells 12V battery, where we can see how voltage varies for a given charge/discharge rate and SoC. [PERE93] State of Health Easiest way to measure the battery SoH is through direct measurement of its capacity. This test consists in recharging the battery until fully charged and then 16

41 2. DESCRIPTION OF ESS Figure Lead acid electrolyte density variation Figure Voltage variations on a 12V lead acid battery, [PERE93] 17

42 2. DESCRIPTION OF ESS discharging it until fully discharged, integrating the discharge current over time so total capacity is calculated. Then actual capacity is compared to theoretical battery capacity to obtain SoH as relative measurement. Disadvantages of this method are that it requires the battery not to be in operation while the test is carried out, it takes long time to fully charge and fully discharge the battery, compared with the time that other methods take (charge and discharge rates must be in accordance with the intended application), and it is an aggressive process for the battery as very low SoC levels must be reached. Main advantages are its simplicity and relatively low costs. [ALBE99] A more accurate way to estimate the battery SoH is through impedance spectroscopy. This process consist in measuring the cell impedance by applying a AC currents or voltages of different frequencies and monitoring the battery response (voltage or current respectively). Then the cell impedance is determined as the complex quotient of voltage and current. Notice that in order to obtain the real and imaginary part of the impedance, voltages and currents must be measured in amplitude and phase. This method is used also to study electrode interface reactions, and characterize materials in corrosion science. In the battery industry, it is valid to measure both SoC and SoH, as well as model the battery behavior. AC current injectors permit the in operation test, but the high cost of the equipment makes the impedance spectroscopy method unsuitable for on-board monitoring, being applied only in laboratory and workshop environments. [ALBE99] [KARD99] Battery models Batteries in general can be modeled through equivalent electrical circuit or electrochemical models. Electrochemical models Electrochemical models describe the behavior of the battery by simulating the chemical reactions that take place inside the battery during the charge and discharge processes. This models provide results more similar to real batteries than the ones that electrical circuit models can provide, but shaping the models is complicate as they require many parameters, some of which are not available for general public, and high computing power to be able to run the simulations and obtain results. Normally they are not used to obtain first results, but to collate data previously obtained [JONG08]. 18

43 2. DESCRIPTION OF ESS Electrochemical models can be simplified through a set of equations that characterize particular phenomena. Peukert equation states the reduction in available charge in the battery caused by an increase in discharge current when the battery is discharged as constant current. Peukert equation is formulated as follows: I n T I = C (2.2.3) Where: I is the discharge current n is a battery type constant (n = 1.35 for typical lead acid batteries) T I is the time that the battery discharge lasts at a current I C is a battery constant (discharge time at 1A) The Kinetic Battery Model (KiBaM), developed by J. Manwell and J. McGowan, is based on chemical kinetics. In this model, the battery is distributed over two wells: the bound-charge well supplies charges (electrons) to the available-charge well, which supplies charge directly to the load. Flow rate between the two wells will defined by fluid level height difference and a throttle constant. This model is represented on figure [MANW93] Figure KiBaM model representation Differential equations represent the dynamics of the fluid model: dv 1 dt = Q 12 = ρg k (h 1 h 2 ) dv 2 dt = Q 12 Qout = ρg k (h 1 h 2 ) Qout (2.2.4) As volume is proportional to height (V = A h): A 1 dh 1 dt = ρg k (h 1 h 2 ) A 2 dh 2 dt = ρg k (h 1 h 2 ) Qout (2.2.5) 19

44 2. DESCRIPTION OF ESS The fluid dynamics model can be converted to an electrical model if voltage is identified with height and current with flow rate. I = V R ; Q = ρg k (h 1 h 2 ) R = k ρg I = C d dr ; Adh dt = Q C = A (2.2.6) C 1 dv 1 dt = 1 R (V 1 V 2 ) C 2 dv 2 dt = 1 R (V 1 V 2 ) I out (2.2.7) Where V 2 is the battery terminal voltage, V 1 is the open circuit stable voltage, C 1 and C 2 are constants expressed in capacitance units and R is a constant expressed in resistance units. Electrical circuits Equivalent electrical circuits provide a much easier way to predict battery behavior than electrochemical models, but their accuracy and the number of effects that can be studied through them is also lower. A single model can be applied to predict only one or a few variables and for multi-variable simulations (e.g. voltage drop, capacity fade, self-discharge, efficiency... ), several independent models may be required. For shallow short time high power discharges (e.g. automotive battery powering the starter motor of the engine) the Thevenin dipole, made up of a resistor in series with a constant voltage power source, could be accurate enough as open circuit voltage difference between maximum and minimum SoC will be negligible compared to the difference between the open circuit voltage and the in charge voltage. For deeper discharges voltage reduction during discharge should be considered. In order to take into account that, the constant voltage power source can be replaced by a capacitor, whose voltage decreases with the energy it contains. For intermittent discharges, the battery experiences a recovery effect during the resting periods that increases battery capacity. This recovery effect will be modeled through a resistor and capacitor parallel branch in series with the main capacitor. On [ALBE99] an equivalent circuit made up of a set of resistor and a capacitor is proposed. The purpose os this circuit is to estimate the SoH of the battery from 20

45 2. DESCRIPTION OF ESS (a) Proposed in [ALBE99] (b) Simplified Figure Equivalent circuits for SoH estimation the measurement of this parameters. Although each resistor models a different phenomenon, obtain every resistor value won t be easy and additional information is not necessary in most of cases. This task can be simplified by considering only two resistors, all those ones in parallel with the capacitor as a single one and all those in series with the capacitor as another single resistor. This simplified circuit will be valid to determine whether the battery is in appropriate conditions to operate or not, regardless of the causes of failure. Figure shows these two possible circuits. Equation of the Kinetic Battery Model can also be represented through a resistance and capacitors circuit. Figure represent said circuit. 21

46 2. DESCRIPTION OF ESS Figure KiBaM model circuit representation Maintenance and failure prevention If low charge level is maintained for a long time, P bso 4 that initially had a porous and spongy structure will crystallize, forming a solid coat that will hinder the formation of βp bo 2 (tetragonal structure), so αp bo 2 (orthorhombic structure) or P bo will be formed instead, causing capacity fade and early battery failure [KELL81]. This phenomenon is illustrated on figure where capacity fade is compared at different depth of discharge. Therefore, battery must be stored fully charged when not in use. Sulfated batteries can be partially recovered by recharging it with a pulse charging current. If a maximum capacity fade is defined to determine the battery failure, higher depth of discharge (DoD) will mean less life cycles, as the threshold capacity fade will be reached sooner. But taking just this measurement is not valid for battery performance optimization as cycles with higher DoD provide also more energy per cycle. To homogenize these values it is proposed the use of an equivalent number of cycles. This proposed equivalent number of cycles is calculated as the product of the actual number of cycles that the battery lasts and the DoD percentage, so it is the hypothetical number of cycles at 100% DoD that provide the same energy than the battery has supplied during its life. Figure represents the total number of cycles and equivalent number of cycles versus DoD for same batteries represented on figure considering the battery failed when it capacity reaches 60% of initial capacity. Figure data is taken from [BIND05] and represent the number of cycles that the batteries last, total energy throughput and equivalent number of cycles for a deep cycle solar 2.1 kwh. 22

47 2. DESCRIPTION OF ESS Figure Lead acid capacity fade Figure Total cycles and equivalent cycles variation with DOD 23

48 2. DESCRIPTION OF ESS (a) (b) (c) Figure kWh deep cycle lead acid batteries life test Notice on figure that there is certain capacity increase during the first cycles. This is because some of the lead alloys that make up the terminal grids are converted into active material (P bo 2 ). Excessive transformation of these alloys into active material will cause electrode decay and, in the end, battery failure. There is no possible recovery of decayed electrodes, so battery must be replaced. Alloying elements as tin are added to the grid to prolong battery life. Another reason for battery failure is the electrolyte stratification. If acid concentration at the bottom of the electrolyte container is significantly greater than it is on the top, most of the electrical power will be provided by the bottom of the electrode plates, resulting on a higher degradation of this zone. As sulphuric acid is heavier than water, it will have a natural tendency to remain the in the deeper zone of the electrolyte. The gas generated during recharge will help to homogenize the electrolyte. 24

49 2. DESCRIPTION OF ESS Figure Basic battery charger scheme Charge/discharge processes Lead acid batteries are characterized for the capacity to withstand rough charging and discharging processes. Thanks to this characteristic, battery charger are simple and inexpensive. Basic chargers are made up of a transformer to reduce voltage from grid voltage to battery voltage and a diode bridge to rectify AC current, as seen in figure On automotive applications, charging voltage is regulated by controlling the alternator excitation current. Automatic battery charger, for example for backup emergency systems, include a voltage regulator to disconnect the charger when the battery is fully charged, preventing the heat generation caused by overcharging that may lead to electrolyte evaporation. When charge is completed, battery voltage suffers a slight decrease, due to the battery stabilization, and selfdischarge is compensated with a floating current Advantages, disadvantages and applications Lead acid batteries are widely use for applications in which a long autonomy is not required or battery weight is not critical. They are an inexpensive systems due to the low battery cost and the simplicity of the battery management systems they operate with. Main applications are uninterruptible power sources, emergency lighting, solar system backup, and marine electronics power source. Despite of their weight, lead acid batteries are also used in mobile applications such as golf carts and forklift, as the range they will move in is not very wide. 25

50 2. DESCRIPTION OF ESS 2.3. Lithium batteries Thanks to its low weight (compared with other battery technologies) lithium batteries have become very popular for portable applications. Such applications include from small electronic devices to electric vehicles with high power and energy demand. Main disadvantage of lithium batteries is its still high cost, although there is a downward tendency caused by the increase in mass production and development of both batteries and manufacturing process. The increase in battery production and, therefore, in lithium consumption will lead to a rise in lithium price, but overall battery price is still expected to decrease, because lithium share in battery price is small (less than 1%) and the other components will be cheaper. [GROS11] [GARC11] Figure Lithium batteries price evolution forecast Operating principle Similar to lead acid batteries, lithium batteries consist of the positive electrode, negative electrodes and electrolyte. The negative electrode is made of graphite (C 6 ) while the positive electrode is a metal oxide (generally LiCoO 2, but some other materials are also possible). Lithium cobalt oxide is arranged in cobalt and oxygen octahedra slabs intercalated with lithium layers. During charge, electron are forced to move from the Lithium cobalt oxide electrode to the graphite electrode, what causes the migration of Li + ions from the positive to the negative electrode. Li + ions are then reduced to Li atoms and stored in between the carbon layers of the graphite. Lithium atoms will tend to release the absorbed electron, moving back to the positive electrode during discharge. The process by which electrons are inserted in the electrode structures is know as intercalation. This reaction is represented by equation 26

51 2. DESCRIPTION OF ESS and illustrated in figure Other materials used for the construction of the negative electrode are LiF ep O 3 and LiMn 2 O 4, with analogous behavior. charge LiCoO 2 disch. Li 1 x CoO 2 + xli + (2.3.1) xli + charge disch. xli Figure Li + intercallation in LiCoO 2 and graphite electrodes, [ELEC11] Battery structure and types Cell voltage depends on the the electrode material. While the negative electrode is most of times graphite, for the positive electrode material there are various options. Most used material is LiCoO 2, considered on equation 2.3.1, but recent years research has lead to the development of some alternative materials. The desired characteristics for the electrode material are that active material structure remains unaltered by the intercalation process, so high reversibility and long battery life is obtained; electronic conductivity to allow the electron movement to the battery terminal (conductive additives may be added to the electrode material to increase conductivity); and high capability to accommodate and release Li + ions (it is considered that at least one ion per transition metal atom must be movable to obtain a high enough specific power). LiNiO 2, with similar slab structure than LiCoO 2, may be used to produce electrodes at lower cost, but it offers lower power capability as the diffusion on Li + ions is not as quick as it is in LiCoO 2 ; moreover, this 27

52 2. DESCRIPTION OF ESS compound is unstable and therefore dangerous in contact with organic liquids, so it is incompatible with some electrolyte dissolvents. There are two families of manganese electrodes: the spinel LiMn 2 O 4 and the layered LiMnO 2. The LiMnO 2 electrode is an environmental friendly and cheaper alternative to the LiCoO 2 electrode; to stabilize its structure, some of the Mn atoms are substituted for Co or Ni atoms. LiF ep O 4 is a low cost, environmental friendly material that admit the intercalation of up to 1 Li every F e in a reversible process. All these positive electrode materials offer a equilibrium potential around 3.6V that raises up to 4.2 when fully charged. [REYN04] [RUFF08] Conventional lithium ion batteries electrolytes consist of a organic solution of an ionic lithium salt, typically LiP F 6. The solvent is a blend of various organic compounds such as ethylene carbonate or diethyl carbonate. Between the two terminal there is a separator made of Polypropylene (PP) and Polyethylene (PE) permits the movement of the lithium ions but not the electrons. Lithium ion polymer batteries (LiP) use instead of the the organic solution a solid polymer such as polyethylene oxide or polyacrylonitrile. The advantages of these batteries over the conventional lithium-ion ones include lower manufacturing costs, adaptability to many packaging shapes, and reliability. Apart from the electrolyte, Lithium polymer batteries work the same as conventional lithium ion batteries, so same electrochemical considerations can be applied Parameter measurements and estimations State of Charge Since lithium ion batteries are used in applications in which high DoD is reached, an appropriate estimation of the SoC is necessary to know how long the battery will last and when it must be recharged. Except for the pressure relief safety valve, lithium-ion cells are hermetically closed, so electrolyte is not accessible for density measurement as it is on lead-acid cells. An easy way to obtain a rough estimation of the actual SoC is through the cell voltage. Similar to lead acid batteries, cell voltage decreases as energy stored decreases, but while on lead acid the voltage variation is observed throughout the discharge, on lithium cells it is accentuated at upper and lower SoC limits, while at partial SoC voltage remains almost steady. Conventional electronic devices use the voltage measurement as the accuracy that they can obtain is good enough for their applications and installing alternative systems will mean higher costs. 28

53 2. DESCRIPTION OF ESS Current integration can also be used to estimate SoC with same considerations than lead-acid batteries. The fact that the measurement of each state is based on the measurement of the previous state makes that errors from noise, resolution and rounding are accumulated causing large deviations from actual state. A more accurate method for SoC estimation is through the cell impedance. The cell impedance is measured by applying an AC current or voltage to the cell. Output current or voltage amplitude and phase is used to determine the real and imaginary part of the impedance. This measurement is performed at various frequencies so data in frequency domain is obtained. Limitations of this methods are that the cell impedance, apart from frequency and SoC, also depends on temperature and DC current, so these values also have to measured and accounted in order to get a good estimation of SoC from cell impedance. [KARD99] [KELL10] State of Health SoH on lithium ion batteries can be measured through direct measurement of the capacity with similar considerations than in lead acid batteries: it requires time and energy waste to fully charge and fully discharge the battery (Section 2.2.3). But there are additional considerations to take into account when dealing with lithium ion batteries. First, lithium batteries support deep discharges much better than lead acid batteries, therefore this test is not as aggressive as it is for lead acid batteries. Second, lithium batteries are normally used in mobile applications and they are easier to remove and, if necessary they can be replaced by another battery while testing. And third, many times the own application fully discharges the batteries, this deep cycles can be used to update SoH if it is follow by a complete charge. These considerations makes the direct measurement of the capacity a very used technique and some BMS update the battery SoH by monitoring continuously the SoC through the terminal voltage. The electrochemical impedance spectroscopy (EIS) (section 2.2.3) can also be applied to lithium ion batteries. Through this technique, ohmic and activation polarization, double layer capacitance and solid electrolyte interphase can be studied. [MOSS08] Other ways used not to obtain a measurement of the SoH, but to be able to determinate whether the battery can still be operated or not are through those manifestations of the own battery aging. Capacity loss has already been considered as a direct way of SoH measurement, but on the same line battery temperature rise at a given charge/discharge rate or self-discharge rate can also be used to check the state of the battery. 29

54 2. DESCRIPTION OF ESS Battery models Same than lead acid batteries, lithium batteries can be modeled through equivalent electrical circuits or electrochemical models. Electrochemical models One of the most relevant electrochemical models for describing lithium ion cells is the model developed by M. Doyle, T. F. Fuller and J. Newman. Said model consists of six differential equations that give voltage and current as function of time. Also some other variables such as salt concentration in the electrolyte, current density and electrode potential can be calculated as function of time and position in the cell through this model. [DOYL93] Dualfoil is program written in Fortran language for lithium-ion simulation through electrochemical models. By setting an user defined load profile, the program computes the evolution of the battery properties over the time. Main limitation of this program is that it requires entering over 50 battery parameters, some of them are not available for an average user, as they require detailed knowledge of the battery to be simulated.[jong08] Peukert equation considered of section and equation for lead acid batteries are also valid, but D. Rakhmatov and S. Vrudhula extended it to give a model that describes the battery performance through the diffusion processes of active materials that take place in the battery. This model considers the diffusion process to be one dimensional, so it can be described through the Fick s law: δc(x, t) J(x, t) = D δx δc(x, t) δt = D δ2 C(x, t) δx 2 (2.3.2) Where x is the position t is the time J(x, t) is the flux of active material C(x, t) is the concentration of active material 30

55 2. DESCRIPTION OF ESS D is the diffusion constant Boundary conditions of the Fick s law are given by Faraday s law: [YUAN01] J(x = 0, t) = i(t) νf A J(x = L, t) = 0 (2.3.3) Where i(t) is the current ν is the number of electrons involved in the electrochemical reaction at the electrode surface F is Faraday s constant ( C/mol) A is the area of the electrode surface Electrical circuits Non linear circuit developed by Ziyad M. Salameh is based on a set of resistors and capacitors, represented in figure 2.3.3, whose components are described here: C b is a capacitor that represents battery capacity R p is a resistor that represents battery selfdischarge The parallel branch C 1, R 1, R 2 represent the overvoltage. Resistance value depends whether the battery is under charge or discharge, so diodes are used to select the appropriate resistance value. R s represents electrolyte and plates resistance. The model is called non linear because the values of C b, R s, R p, R 1 and R 2 vary with voltage and temperature. Only C 1 is truly constant. However, some assumptions can be made for given states to get an approximate solution from constant values. Shepherd Model Developed by Clarence M. Shepherd, this model is neither an electrochemical model nor an equivalent circuit, but it is a mathematical expression that represents 31

56 2. DESCRIPTION OF ESS Figure Non linear circuit model for lithium ion batteries the behavior of lithium ion batteries. This model is commonly used in large power applications such as hybrid electric vehicles. Equation is the basic expression of this model. E T = E 0 R i I K i 1 1 f (2.3.4) Where: E T is the battery terminal voltage E 0 is the fully charged open circuit voltage R i is the internal resistance I is the output current K i is called polarization constant f is a measurement of the amount of charge removed, opposite to SoC f = Idt Q 0 (2.3.5) Graph of figure shows the evolution of the battery for certain R i and K i during discharge at different discharge rates. Notice that even when battery is fully charged and there is no output current according to equation the terminal voltage of the battery will be lower than the theoretical full charge open circuit 32

57 2. DESCRIPTION OF ESS voltage. Therefore, the model won t be applicable when operating at low discharge currents and high SoC levels, i.e. when E T is closer to E 0. E T (0) = E 0 R i 0 K i = E 0 K i < E 0 (2.3.6) Unnewehr Universal Model is a simplification of Shepherd model that attempts to solve the low current high SoC problem. Expression of Unnewehr model is equation E T = E 0 R i I K i f (2.3.7) Maintenance and failure prevention Lithium ion batteries failures can be classified into energetic and non-energetic modes. Energetic failure modes are those related to loss of performance, such as capacity loss, internal impedance increase and selfdischarge increase. Lithium battery packs normally include some mechanisms to permanently disable the battery if too low SoH is reached, so no longer operation is permitted to avoid dangerous situations. The main causes that lead to a premature degradation are overcharge, overdischarge and too high charging or discharging currents. Overcharge causes that when no more lithium can be intercalated in the graphite structure of the anode, it starts to create dendrites that may even trespass the electrolyte and, eventually, cause internal short circuit. Also in the cathode the overcharge may cause lack of lithium in the ionic structure, so it becomes unstable. If overdischarge is repeatedly carried out, lithium plating and dendrite growth can happen similarly than when overcharging. To avoid overdischarge damage battery management systems, or battery pack safety circuits, auto-disconnect the battery when minimum SoC is reached, connecting it back when battery is recharged. Charge or discharge current abuse may cause internal temperature increase leading to electrolyte degradation and, in a extreme case, thermal runway. The energetic failure (thermal runway) consists in a sudden selfdischarge of cell, causing a high temperature increase that may lead to further dangerous situations, as explosion due to internal pressure increase and combustion. Although thermal runway may happen in many other battery chemistries, it is especially important in lithium ion batteries due to their high energy density. One of the reasons that might 33

58 2. DESCRIPTION OF ESS cause thermal runway is a high temperature: if cell temperature rises above 90 o C, the graphite electrode can initiate this reaction. Another typical reason of thermal runway is mechanical abuse, as impacts on the battery. Risk of thermal runway and the severity of the effects it can cause increases with the SoC, as there is more energy available. Lithium cells are equipped with a pressure relief valve similar to VRLA batteries. If, due to pressure increase from thermal runway, this valve opens, released gases might ignite. [MIKO11] Charge/discharge processes Operating lithium ion batteries requires a close control of voltage and current applied, as they are very sensitive to overcharge and high rates. Discharge is controlled through battery management systems, monitoring the individual current, voltage and temperature of every cell that makes up the battery and allowing charge transfer among them to compensate differences. Battery chargers, compared to the simplicity of lead acid battery chargers as shown on figure 2.2.9, are also complex systems that watch that the charging process is carried out within the safe operation limits. Typical charge profile consists of an initial constant current and variable voltage charge followed by constant voltage and variable current charge, until current drops below a minimum value, when charge is stopped. If voltage falls due to self-discharge or battery relaxation below a given threshold value, the charge will be reactivated again. This process is illustrated on figure Advantages, disadvantages and applications The main advantage of lithium ion batteries are their low weight compared with other battery technologies. This characteristic makes them show a high specific power and specific energy compared with other battery technologies and other ESS systems in general. Also in terms of energy density and power density (per volume unit instead of per mass unit) show much better values that other technologies. Thanks to these characteristics lithium batteries are, nowadays, the best option for mobile applications, from small electronic devices to large electric vehicles. Another advantage compared with NiCd and NiHM batteries (not covered by this project) is that lithium batteries don t show memory effect. This is of great 34

59 2. DESCRIPTION OF ESS importance when applications don t allow to wait to fully discharge the battery before starting the recharge. The main disadvantages is still high price, which, as figure shows, is expected to decrease in the coming year. 35

60 2. DESCRIPTION OF ESS Figure Battery performance according to Shepherd model Figure Battery performance according to Unnewehr model 36

61 2. DESCRIPTION OF ESS Figure Lithium ion battery charge 37

62 2. DESCRIPTION OF ESS 2.4. Hydraulic energy storage systems Hydraulic energy storage systems work storing the energy through a pressurized fluid. There are some other systems, mainly CAES and hydroelectric power, that also work by applying pressure to a fluid, but the way they work is completely different. Compressed air energy storage plants (CAES) consist of gas cycles power plants (Brayton cycle or similar) equipped with a large underground impermeable cavern in with compressed air is stored during low electricity demand hours. This compressed air is applied to the cycle intake so energy consumed by the compressor is significantly reduced. Hydroelectric power consists in extracting the power of a water stream (such as a river) by means of a damn, in which water is stored and pressurized by gravity. For energy demand compensation, pumped hydroelectric power plants consist of two reservoirs at different heights; when there is electric energy excess water from the lower reservoir is pumped to the upper, so that power can be recovered when the energy demand is higher. Although both systems are very interesting for high scale energy demand and production compensation, they will be out of the scope of current project as we will focus on smaller scale applications. The hydraulic systems we are proposing will be based on hydraulic accumulators. Hydraulic accumulators are already widely used in hydraulic circuits and other systems, but most of the time their application is limited to smooth pressure peaks that may cause system damage, control the pressure and provide a fluid reservoir to ensure the continuity of the flow in the circuit pipes. We consider them to have a great potential also as energy storage systems and this project will evaluate their feasibility on energy applications and compare their performance to that of other typical systems such as the batteries already studied in the previous sections Operating principle Hydraulic accumulators are devices that store incompressible fluid under pressure on a closed (except for charging and discharging ports) variable-volume chamber. Pressure is applied to the fluid through the movable limit of the chamber by means of mechanical or pneumatic systems. The basic principle that governs the behavior the accumulators is the force balance on both sides of the barrier that limits the chamber. Mechanical accumulators are cylinders in which one of the bases is a movable piston. The movement of said piston causes the variation in the chamber volume. Fluid is pressurized by applying a force to the piston. Said force may be produced 38

63 2. DESCRIPTION OF ESS from a weight (gravity or weighted accumulators) or a spring (spring accumulators). Main advantage of weighted accumulators is that pressure doesn t vary with the state of charge. Disadvantages are the high weight and size that they require to achieve acceptable pressure and capacity compared to other accumulator types. In spring accumulators, considering that force is proportional to the compression, pressure will be proportional to the volume of stored fluid (plus the preset pressure), so energy stored is proportional to the square of the pressure. Gas charged accumulators (also called hydropneumatic accumulators) are hydraulic accumulatosr in which the fluid is pressurized through the compression of a gas. This type of accumulators is the most used in industry and it is the one we will be focused from now on, as it is also the type of accumulator we will test on chapter Structure and types Gas charged accumulators can be classified in three different configurations. Piston accumulators are closed cylinders with a movable piston inside that divides the interior of the cylinder in two different closed chambers. One of the chambers will be filled with the gas and the other chamber with the hydraulic fluid. As the volume of stored fluid increases the piston will move toward the gas side, which will be compressed in the other chamber and, therefore, the pressure will increase. The other two types are bladder accumulators and diaphragm accumulators. In both of this two accumulator types, hydraulic oil is separated from gas by a flexible barrier. Bladder accumulator said barrier consists in a bag inside the cavity of accumulator. Said bag is filled with the gas and oil is stored in the space between the walls and the bag. Diaphragm accumulator barrier consists in a membrane that divides the cavity in two halves, that membrane works like the piston in the piston type, but instead of being displaced it is deformed toward the gas side when charging or the oil side when discharging. Bladder and diaphragm accumulators offer a faster response than piston accumulators and they are better to avoid infiltrations of gas in the oil side or oil in the gas side as they are completely separated, while piston accumulators require piston rings to seal the chambers. Advantage of piston accumulator are that they normally withstand higher pressures, they show a wider precharge pressure range and offer higher flow rates. Diaphragm accumulators are usually manufactured for low volumes, up to 5 litters, while higher volume accumulators are commonly bladder type. 39

64 2. DESCRIPTION OF ESS Figure Gas loaded accumulators [PARK03] Besides the basic elements (containing vessel and separator whether diaphragm, bladder or piston), accumulators have a gas charging valve and a poppet valve that controls the maximum flow rate and keep the bladder of diaphragm from been extruded if the accumulator is excessively discharged. Figure shows these three types of accumulators and their main components. In addition to those three types, there if a fourth type of hydraulic accumulators. Metal bellows accumulators are very similar bladder accumulators, but in this case the elastomeric bladder is replaced by metal bellows. Hydraulic fluid may be inside of the bellows with the gas between the bellows and the container or opposite, the gas inside the bellows and the fluid outside. Although their response time is not as fast as for the bladder of diaphragm type, they offer a very good reliability and long life, with little or no maintenance required. So they are often applied in emergency systems, such as emergency brakes Parameter measurement and estimations State of Charge Best way to know the SoC of an accumulator is through its pressure. As seen previously, pressure varies with the charge for any type of accumulator except for weighted ones. When the gas is compressed its temperature raises and heat will be rejected to the environment causing an energy loss. Depending on charging speed and cycle time the energy losses from heat rejection will be significant or not. In order to obtain an accurate measurement of the energy stored in the accumulator, besides the pressure, the gas temperature must be measured. The 40