REGENERATIVE BRAKING SYSTEMS

Transcription

1 ESCUELA TÉCNICA SUPERIOR DE INGENIERÍA (ICAI) INGENIERO IEM REGENERATIVE BRAKING SYSTEMS Autor: Leticia Vila-Coro Fuentes Director: Dr. Patrick N. Currier Madrid Mayo 2014

2

3 AUTORIZACIÓN PARA LA DIGITALIZACIÓN, DEPÓSITO Y DIVULGACIÓN EN ACCESO ABIERTO (RESTRINGIDO) DE DOCUMENTACIÓN 1º. Declaración de la autoría y acreditación de la misma. El autor D. Leticia Vila-Coro Fuentes, como alumna 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 la obra Regenerative Braking Systems 1, 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 1 Proyecto de Fin de Grado 1

4 como 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 y ponerla a disposición del público a través de un archivo abierto institucional, accesible de modo libre y gratuito a través de internet. 2 (d) Distribuir copias electrónicas de la obra a los usuarios en un soporte digital. 3 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. 2 En el supuesto de que el autor opte por el acceso restringido, este apartado quedaría redactado en los siguientes términos: (c) Comunicarla y ponerla a disposición del público a través de un archivo institucional, accesible de modo restringido, en los términos previstos en el Reglamento del Repositorio Institucional 3 En el supuesto de que el autor opte por el acceso restringido, este apartado quedaría eliminado. 2

5 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. 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: 3

6 - retirar la obra, previa notificación al autor, en supuestos suficientemente justificados, o en caso de reclamaciones de terceros. Madrid, a 28 de Mayo de 2014 ACEPTA Fdo 4

7

8

9 SISTEMA DE FRENADO REGENERATIVO Autor: Vila-Coro Fuentes, Leticia. Director: Currier, Patrick. Entidad Colaboradora: EcoCAR 2. Introducción RESUMEN DEL PROYECTO El frenado regenerativo consiste en hacer uso de la energía derivada de hacer que un vehículo disminuya su velocidad o se pare. En este proyecto, un sistema de frenado regenerativo será desarrollado e implementado para el vehículo de la universidad Embry-Riddle Aeronautical University para la competición EcoCAR 2. El coche en el que se implementará este sistema es un Chevrolet Malibu donado por General Motors y convertido en un híbrido que se conecta a la red, entre otros con donaciones de los patrocinadores de la competición. Se alimenta de electricidad y biodiesel B20, y utiliza un motor eléctrico AM-Racing de REMY, una batería Li-ion y un motor diésel para alimentar dicha batería (Embry-Riddle EcoeEagles, 2014). Todas las modificaciones en todos los sistemas principales del coche son hechas por estudiantes, entrenados y ayudados por los patrocinadores. Los alumnos se organizan en equipos para concentrarse en el desarrollo de un área específica, como puede ser la mecánica, eléctrica, los sistemas de control La competición EcoCAR 2 busca desarrollar un vehículo que, comparado al convencional de gasolina, consuma menos y tenga menos emisiones, manteniendo la aceptación del consumidor. Quince universidades de los EEUU y Canadá compiten en el diseño, implementación e integración de los sistemas necesarios para alcanzar esas metas, desarrollando el vehículo durante un periodo de tres años. La competición también introduce algunas limitaciones que influyen sobre las decisiones tomadas para el diseño del Sistema de Frenado Regenerativo (Regenerative Braking System o RBS), mientras que a su vez, provee de los medios para su desarrollo. Estado de la técnica Hay varios sistemas que se basan en los mismos principios que el desarrollado, como pueden ser los KERS y los frenos hidráulicos regenerativos. Sin embargo, se ha elegido diseñar un sistema de control que regule la cantidad de par utilizado para frenar el coche.

10 Mientras se está frenando el coche, el motor se convierte en generador, alimentando la batería, que se carga en el proceso. Por lo tanto, mediante el uso del frenado regenerativo, el alcance del vehículo aumenta dependiendo en la cantidad de par exigida y el número de veces que se activa el sistema. Este par es negativo y es la salida del sistema implementado. Hay dos tipos de sistemas de frenado regenerativos: los que se centran en el uso del acelerador y los que se centran en el pedal de freno. Para el sistema implementado se ha elegido la última de las posibilidades, que a su vez de divide en dos corrientes: frenado en serie, en el que se decide la cantidad de frenado regenerativo y fricción a aplicar en cada caso, dependiendo de las intenciones del conductor, o en paralelo, que consiste en aplicar en todo caso fricción y frenado regenerativo además cuando se considere conveniente. El enfoque tomado ha sido diseñar un sistema en paralelo. Cada vez que el conductor presione el freno, se aplicará frenado mediante fricción y mediante freno regenerativo, si el usuario así lo decide y se cumplen las condiciones de seguridad. De esta manera, no es necesario un sistema electrónico de control de frenado que decida la cantidad de freno regenerativo que el coche y la batería pueden aportar, y lo complemente con fricción, lo que atribuye al sistema características de seguridad para frenado de emergencia y disminuye la complejidad del sistema. Motivación Como parte de la competición, el equipo (también llamado EcoEagles), decidió desarrollar un vehículo ecológico que maximizara el alcance utilizando la menor cantidad de energía posible. Este proyecto forma parte de esa meta ya que se aprovecha de energía no utilizada para cargar la batería. También fue diseñado para que la mínima cantidad de ensayos fueran necesarios a la hora de hacer que el sistema funcionara, de manera que no interfiriera con los ensayos críticos para el funcionamiento del coche. El frenado regenerativo es un tipo de sistema relativamente novedoso, en el cual todos los posibles enfoques no han sido estudiados todavía. Por lo tanto proporciona grandes posibilidades a la hora de desarrollar sistemas innovadores o más simples que los implementados en la industria. Siendo la industria automovilística tan confidencial y competitiva como es, no ha sido posible encontrar líneas generales de acción a seguir para el desarrollo de ningún tipo de frenado regenerativo. Mediante el estudio de datos, comentarios de los consumidores y características de los productos se ha podido obtener una idea de los enfoques tomados por las empresas para el desarrollo de estos sistemas. En base a esas deducciones se han desarrollado y hecho funcionar los sistemas discutidos en este documento, que da información sobre cómo se crea un RBS. Objetivos Los objetivos del proyecto son los siguientes:

11 Crear el Sistema de Frenado Regenerativo a implementar en el coche que sea suficientemente simple como para no requerir muchas pruebas, que incremente el alcance, mantenga la aceptación del consumidor y seguridad de los usuarios. Todo esto será hecho siguiendo las reglas de la competición EcoCAR 2. Implementar de manera satisfactoria el RBS contemplando todos los requerimientos. Analizar los resultados mediante una prueba de conducción y frenado con el RBS activado. Se decidió añadir un objetivo adicional: crear un sistema de frenado regenerativo innovador, llamado Sistema Incremental, que esté basado en el uso del acelerador y que presente un enfoque más simple que los sistemas similares siendo utilizados por algunas empresas, mediante el uso de menos variables y eliminando la necesidad de creación de un mapa del motor que relacione par y velocidad. Este sistema no será implementado, pero su base teórica será discutida, así como una justificación de por qué es novedoso. Metodología El sistema principal se basará en un planteamiento concentrado en el uso del pedal de freno. Sin embargo, no sólo se desarrollará esta opción, sino que también se discutirán las posibilidades de un sistema innovador concentrado en el uso del acelerador. Este último sistema será diferente a lo implementado por la industria en el momento en el que se escribe este documento, sin embargo, solamente será desarrollado, simulado y analizado. El primer paso en el desarrollo del RBS será decidir las entradas y salidas del sistema utilizadas para cumplir las metas. Esto se hará mediante el análisis de las señales disponibles y decidiendo cuáles serán críticas para su correcto funcionamiento contemplando los límites propuestos. Después, se modificará la planta del coche del que se dispone y se validará, para ejecutar simulaciones y decidir el enfoque a tomar. Dado que el Sistema de Frenado Regenerativo se desarrolló a la vez que el coche también estaba en fase de desarrollo, será necesario asegurar la mínima cantidad de test a ejecutar posibles. Este enfoque no utilizará ninguna señal que no estuviera disponible cuando se empezó a desarrollar el sistema, y por lo tanto, las señales utilizadas no serán las convencionales. Un ejemplo de esto es que no se utilizará la presión del líquido de frenos, sino que las señales utilizadas serán las observadas en la Figura 1: la posición del pedal de freno, el gradiente de dicha posición, la velocidad y una señal para la activación del sistema que será salida de una máquina de estados finitos que garantiza condiciones de seguridad.

12 Figura 1: Sistema de Frenado Regenerativo Una vez el sistema ha sido diseñado, los resultados de la implementación serán discutidos, y comparados con aquellos que habrían sido obtenidos sin la implementación del sistema, así como los beneficios obtenidos. También se analizarán las posibles modificaciones que podrían mejorar la aceptación de los consumidores. También se hará un análisis en cuanto a potencia y energía, para justificar la implementación de este tipo de sistema como una forma más simple de lo que vehículos de todo el mundo llevan incorporado, con resultados similares. El Sistema Incremental, que se puede observar en la Figura 2, un sistema innovador que deriva su funcionamiento del uso del acelerador será discutido, así como sus diferencias con los sistemas convencionalmente utilizados por la industria. Figura 2: Sistema Incremental Se hará un análisis de su comportamiento y tratamiento de las señales necesarias para implementar este sistema. La base sobre la que se apoya se explicará, y para la misma planta utilizada en el sistema anterior, se simulará. Los resultados se explicarán, así como las características del sistema y cómo éste afectará al uso del acelerador. Resultados El sistema de frenado regenerativo principal fue exitosamente diseñado e implementado, siguiendo un planteamiento conservativo para procurar aceptación del consumidor y la mínima cantidad de pruebas necesaria para proporcionar un sistema que funcionara.

13 El siguiente paso fue llevar a cabo una prueba de frenado y conducción para poder estudiar el sistema y comparar su funcionamiento con el correspondiente sin el sistema activado. Este estudio se centrará en energía y potencia, y en cómo el alcance del coche mejoraría hipotéticamente. En la Figura 3 se puede observar una comparación teórica basada en datos extraídos de la prueba. Está claro que el uso de este sistema, en este caso hipotético, mejoraría su comportamiento en más de un 42%. Figura 3: Comparación con y sin el RBS activado Conclusión Estos sistemas presentan un nuevo y simplificado enfoque a sistemas siendo desarrollados por la industria automovilística. Los sistemas de control involucrados en este proyecto proveen de innovadoras formas de afrontar problemas a los que se busca solución, con resultados positivos para variables como alcance o consumo de energía. Más que eso, el sistema implementado mantiene aceptabilidad por parte del consumidor, al ser la sensación al utilizarlo muy similar a la convencional, a la vez que alcanza el resto de metas propuestas. El Sistema Incremental presenta, a su vez, una posible solución simplificada a desarrollos complejos. Estos sistemas tienen el claro beneficio de aumentar el alcance del coche híbrido gratis, mientra s que reducen emisiones y contribuyen a la sostenibilidad del medio ambiente. El beneficio principal es que todos los conductores puedes adaptarse a estos sistemas, y estos sistemas pueden adaptarse a los conductores.

14 REGEN BRAKING SYSTEM SUMMARY OF THE PROJECT Introduction Regenerative braking consists on making use of the energy derived from slowing down or stopping a vehicle. In this project, a regenerative braking system will be developed and implemented for the Embry-Riddle Aeronautical University EcoCAR 2 competition vehicle. The car this system will be implemented and developed for is a Chevrolet Malibu donated by General Motors and converted into a plug-in hybrid, amongst others with donations from the EcoCAR 2 sponsors. It runs on both electricity and B20 bio-diesel, and uses AM-Racing, REMY electric motors, an A123 Systems Lithium Ion Battery Pack and a diesel engine to sustain the battery (Embry-Riddle EcoeEagles, 2014). All the modifications in all of the main systems in the car are done by students, which are also trained and helped by the sponsors. The students are organized into teams to develop one critical area essential for the functioning of the vehicle, such as mechanical, electrical, controls, software or communications. The EcoCAR 2 competition aims to develop a vehicle that compared to the conventional gasoline vehicle consumes less fuel and has less emissions, while maintaining consumer acceptability. Fifteen universities of the United States of America and Canada compete by designing and integrating the necessary powertrains to reach those goals, developing the vehicle during a three year period. The competition introduces some constraints that shape some of the decisions taken in the development of the Regenerative Braking System, while providing with the means to develop it. State of the Art There are several systems that perform the same function as the system developed for this project, for example those that use flywheels or hydraulics. The system here developed and implemented is a control system that regulates the amount of torque that will be used to stop the car. While regenerative braking occurs, the motor runs as a generator, feeding current to the battery, which is charged in the process. Therefore, by the usage of this system, the driving range of the vehicle will be increased depending on the amount of regenerative braking events and the torque the motor provides to stop the car in every event. This torque will be negative and, in this project, commanded as an output from the system developed. There are mainly two types of regenerative braking control systems: those focused on the use of the accelerator pedal and those focused on the use of the brake pedal. This project will concentra te on the latter, in which there are also two currents: to implement it by determining how much regen

15 braking and how much friction braking will occur depending on the intentions of the driver (series approach) and to implement both each braking event (parallel approach). The approach taken in this project is to implement a parallel regenerative braking system. This means that the regen braking system will be implemented but the friction brakes will still operate every time the user presses the brake pedal. This is an approach that does not need an additional EBS to determine the friction to regen ratio, and provides with safety measurements as the hydraulic braking will always be active in the car for emergency braking. Grounds for the Project As a part of the EcoCAR 2 competition the Embry-Riddle Aeronautical University EcoEagles aimed to develop an eco-friendly vehicle that can provide with as much driving range as possible by using the least amount of energy necessary. This project fits into that goal as it takes advantage of unused energy to recharge the battery. It was also developed to require a minimum amount of testing in order not to interfere with other critical systems being developed and implemented in the car. Regenerative braking is a relatively new type of system that has approaches that have not been completely developed and researched yet. Therefore it provides with great possibilities to develop innovative, and even simpler approaches from those that are currently being implemented by the industry. Being the automotive industry as secretive as it is, there are no guidelines of how to develop this type of system. By studying data, consumer comments and product characteristics a superficial idea of current approaches taken by car companies can be deduced. This project has its base in a deepened analysis of those deductions, which are developed and finally converted into a working system that is explained and provides with an insight into how regenerative braking is created. Objectives The objectives of the project are: To create the Regenerative Braking System to be implemented in the car that is simple enough not to require a lot of testing, increases range, maintains consumer acceptability and ensures safety of its users. All of this will be done while following the rules of the EcoCAR 2 competition. To successfully implement the Regenerative Braking System in the EcoCAR 2 vehicle meeting all of the necessary requirements. To analyze the results obtained by performing a test drive with the Regenerative Braking System enabled that focuses on braking events to ensure its correct performance.

16 An additional objective was added; to create an innovative additional regenerative braking system, called the Incremental System, based on the accelerator pedal position that presents with a simpler approach than the similar systems being currently implemented, as it will simplify most of the variables conventionally used and will eliminate the need of a very precise mapping of the motor in relation to torque and speed. This system will not be implemented in the car but its theoretical basis will be set, as well as a justification of why it is original. Methodology This main system will be based on an approach focused on the brake pedal. However, it will not only concentrate on that, but it will discuss the possibilities of an innovative regenerative braking system applied to the accelerator pedal. This second system will be different to what the industry is implementing at the time of writing of this project, however, it will not be implemented, but developed and analyzed. The first step in this project will be to decide the inputs and outputs used to achieve the goals. This will be done by analyzing the signals available and deciding which ones will be critical for the correct performance of the system and to meet the constraints imposed. Then, a model provided will be modified and validated, in order to perform various simulations and decide on the best approach to take when deciding what the relation between the inputs and the outputs should be. This decision will be made regarding consumer acceptability and safety features. Being the Regenerative Braking System developed while the car was engineered and improved, an approach to provide with minimal testing had to be created. This approach did not need any signal that was not available at the moment at which this system started to be developed, and therefore, the signals used were not be the ones that would be expected. An example of this is that the pressure of the fluid in the friction brakes will not be an input for the system that decides the amount of torque commanded, but the only signals that will be used can be seen in Figure 4 and are brake pedal position, brake pedal position gradient, speed of the car and the regen enable signal derived from a finite state machine to guarantee safety features. Figure 4: Regenerative Braking System

17 Once the system has been engineered, the results of the implementation of one of the approaches will be discussed. These results will be compared to those that would have been obtained without the Regenerative Braking System implemented, and the benefits from the implementation of this system will be analyzed, along with possible modifications that could potentially improve it or its consumer acceptability. Also, there will be a brief energy and power analysis, to justify the implementation of this type of system as a simpler and easier to develop version of what vehicles all around the world have, with similar outcomes. The Incremental System, that can be seen in Figure 5, an innovative regenerative braking system that derives its performance from the use of the accelerator pedal in a new way, will then be discussed, along with the differences with the conventional equivalent system being implemented by the industry. Figure 5: Incremental System There will be an analysis of its behavior and treatment of the signals needed to implement this type of system. The basis upon which this system stands will be set and for the same model used in the Regenerative Braking System, it will be simulated. The results will be explained and the characteristics derived from the implementation of this system, such as how it will change the accelerator pedal use of the driver, will be talked about. Results The main Regenerative Braking System was successfully designed and implemented, following a conservative approach to ensure consumer acceptability and the least amount of testing necessary to provide with a working system. The next step was to perform a test drive to provide with means to study the benefits of this system and compare its performance to that without it enabled. The main studies done focus on power and energy, and how the range would be hypothetically improved. In Figure 6 a hypothetical

18 comparison between the same test drive data without the regen torque and the power and energy generated associated to it and the equivalent with RB enabled can be seen. It is clear that introducing regenerative braking improves energy usage by more than 42% of the total value for this test drive. Conclusion Figure 6: Comparison with and without RB These systems present with a fresh and new approach to important problems currently being tried to solve by the automotive industry. The control systems developed in this project provide simple and easy-to-implement solutions with positive outcomes for variables such as range and energy consumption. Moreover, the implemented Regenerative Braking System maintains consumer acceptability and a positive driver feel while meeting all of the goals posed, and the Incremental System presents with a possible simpler solution to more complicated procedures. These systems have the clear benefit of improving the driving range for free, while reducing emissions and creating a more eco-friendly product. The main benefit of the approach implemented is that even the most conservative drivers could adapt to it without problems, as it does not change the sensation during braking. References/ Referencias Embry-Riddle EcoeEagles. (2014). EcoEagles- About Us: Our Vehicle. Obtenido de EcoEagles:

19 ESCUELA TÉCNICA SUPERIOR DE INGENIERÍA (ICAI) INGENIERO IEM REGENERATIVE BRAKING SYSTEMS Autor: Leticia Vila-Coro Fuentes Director: Dr. Patrick N. Currier Madrid Mayo 2014

20

21 Index Part I: Report... 7 Chapter I: Introduction The State of Art of Regenerative Braking Motivation Background of the project Objectives Methodology Chapter II: Development Approach Inputs: Chapter III: Model Powertrain Vehicle Dynamics Chapter IV: Simulations Curve Curve Curve Curve Curve Curve Comparison Chapter V: Results Torque analysis Voltage and current analysis Energy and power analysis Chapter VI: The Incremental System Basis of the system and grounds to consider this system innovative Conditioning of the primary signal Torque as a function of accelerator pedal position and speed Development of the system Simulation Advantages and drawbacks

22 Chapter VII: Conclusion Future development Part II: Economic Study Chapter I: Economic Study Annex I: EcoCAR 2 Sponsors Annex II: Simulink diagrams: Simulink diagram of the final Regenerative Braking System: Simulink diagram of the Incremental System: Annex III: References

23 Index of Figures Figure 1: Plug-in hybrid sales (Cobb, 2014) Figure 2: Main view of the RB system Figure 3: Regenerative Torque as a Function of Speed Curve Figure 4: Brake Pedal Position Gradient Test Data Figure 5: Brake Pedal Position Test Data Figure 6: Main view of the finite state machine Figure 7: Finite state machine Figure 8: Simulink model of the system Figure 9: Main view of the original model Figure 10: Final model for the car Figure 11: Inputs and outputs of the vehicle dynamics subsystem Figure 12: Speed comparison of the model and car before changing parameters Figure 13: Model and car speed comparison Figure 14: Torque comparison between the model and the car Figure 15: Brake Pedal Position Test Input Figure 16: Brake Pedal Position Gradient Test Input Figure 17: Speed Test Input Figure 18: Torque Command Test Input Figure 19: View of the delay of the system Figure 20: Curves for the controller Figure 21: Torque Output for Curve # Figure 22: Acceleration Output for Curve # Figure 23: Torque Output Curve # Figure 24: Acceleration Curve # Figure 25: Torque Output # Figure 26: Acceleration Output # Figure 27: Torque Output Curve # Figure 28: Acceleration Curve # Figure 29: Torque Output Curve # Figure 30: Acceleration Output Curve # Figure 31: Torque Output for Curve Figure 32: Acceleration Output for Curve Figure 33: Comparison of Outputs Figure 34: Speed, Torque and Brake Pedal Position. Data from the car

24 Figure 35: Closer look at Speed, Torque and Brake Pedal Position Figure 36: Commanded Torque and Torque Feedback main view Figure 37: Commanded and Feedback Torque closer view Figure 38: Detail of regen Torque Command and Feedback Figure 39: Theoretical and Real Torque Figure 40: Detail of Theoretical and Real Torque Figure 41: Closer look at regen part of Real and Theoretical Torque Figure 42: Torque Compared to Current Figure 43: Torque Compared to Current Detail Figure 44: Voltage comparison with torque Figure 45: Voltage comparison to torque in detail Figure 46: Power compared to torque Figure 47: Power compared to torque detail Figure 48: Energy consumed Figure 49: Energy consumed detail Figure 50: Comparison of energy with and without regen Figure 51: Energy recovered with RB Figure 52: Distance traveled by the car in the test drive Figure 53: Finite state machine that decides when positive or negative torque should be applied. 76 Figure 54: Accelerator Pedal Data Figure 55: Torque as a function of accelerator pedal position Figure 56: Curve for the regen multiplier Figure 57: Overview of the incremental system Figure 58: Simulink diagram of the incremental approach Figure 59: Positive Torque Command Figure 60: Negative Torque Command Figure 61: Torque comparison for the incremental system Figure 62: Regen torque and accelerator pedal position

25 Index of Tables Table 1: Regenerative Torque as a Function of Speed Table 2: Finite State Machine Simulation Results Table 3: Inputs and outputs of the model Table 4: Powertrain inputs and outputs Table 5: Inputs and outputs of the Powertrain Table 6: Parameters of the transmission Table 7: Parameters for the tractive force block Table 8: Formulas for "Vehicle Dynamics" Table 9: Parameters for "Vehicle Dynamics" Table 10: Input Curve # Table 11: Input Curve # Table 12: Input Curve # Table 13: Input Curve # Table 14: Input Curve # Table 15: Input Curve # Table 16: Pedal position and torque relation Table 17: Pedal position and Torque as an input Table 18: Torque multiplier

26 6

27 Part I: Report 7

28 8

29 Chapter I: Introduction With a higher consumer interest on sustainability, and an increasing tendency to cut down on non-renewable sources of energy, electric and hybrid vehicles have seen their popularity rise very fast. This increase in consumer acceptability has also come at the same time their efficiency has started to match that of conventional cars. Electric cars have been around since the 19 th century, but lost to the internal combustion engine vehicles in the 20 th century when Charles Kettering invented the electric starter and the muffler, invented by Hiram Percy Maxim, started to be used to reduce noise. This, added to the cheap price of gas, and good mileage provided by the engine, made what we now know as conventional cars far more interesting for consumers (History of the electric vehicle, 2014). The main problem that electric cars faced then was their range, which has improved since then, but cannot be compared to the range an ICE car can give with one full tank. However, range was greatly improved with the invention of the rechargeable battery and the decrease on weight of the batteries. The nickel-metal oxide (NiMH) battery was a great improvement compared to the first rechargeable battery also known as the lead-acid battery, as they are less toxic and heavy. However, studies show they are carcinogenic. The invention of lithium-ion batteries, the least toxic and heavy of them all, and used in multiple other applications, was the last step to create a consumer-acceptable electric car (Science Channel). Rechargeable batteries were a great improvement because they enable the consumer not only to recharge the battery when it is flat, but it provided with the means to develop new systems that can charge the battery by taking advantage of energy that would be wasted otherwise. Regenerative braking is one of the systems developed from this improvement. It is based on taking advantage of the energy lost during braking, for example in the form of heat and the energy derived from slowing down the car. The motor works as a generator supplying current to the battery to charge it. This creates a momentum that opposes the movement of the car and stops it in the process. This approach is not only sustainable engineering as it takes advantage of energy otherwise wasted, but it also helps increase the driving range of the motor. It has been proved that the vehicle s efficiency can be improved up to 30% of the energy demanded by using from recaptured energy (Walker, Lamperth, & Wilkings, 2002). The limits related to recovering vehicle kinetic energy through this type of system are mainly derived from the charging capability and other constraints of the battery, for example how much current it can accept and how much it can be charged. The other limit that must be taken into account is the ABS system, as combining regenerative braking and ABS can 9

30 lead to breaking of the drive shafts. Therefore, regen and ABS should be coordinated but used exclusively. The regenerative braking system developed in this project will be designed and implemented in the Embry-Riddle Aeronautical University EcoCAR 2 vehicle. This HEV is part of the EcoCAR 2 competition, which is a three-year competition in which students develop a competent vehicle that, compared to the production gasoline vehicle, reduces fuel consumption and greenhouse and tailpipe emissions while maintaining consumer acceptability in the areas of performance, utility and safety. This competition was established by the U.S. Department of Energy (DOE) and General Motors (GM), and challenges 15 universities across North America to reduce the environmental impact of a Chevrolet Malibu, following a real-world Vehicle Development Process (VDP) (ECoCAR 2, 2011). The involvement in this competition introduces constraints to the engineering of the system, as there are certain rules related to braking that have to be taken into account. An example of this is that a series regenerative braking system cannot be implemented, as the hydraulic brakes have to always be functioning as a safety feature. In this project a regenerative braking system will be engineered and implemented in the car. This system will be based on a brake pedal related approach. However, the possibilities of a system applied to the accelerator pedal will be discussed as well. 10

31 1.1 The State of Art of Regenerative Braking Regenerative braking was first implemented in a car in 1967 by American Motors. The first car to have it was the Amitron, a small compact car engineered in conjunction with Gulton Industries (Dachet, 2013). It had an autonomy of 241 km at 80 km/h and used two different batteries for different purposes: power delivery and energy storage (Vielcker, 2014). However, this car never got past the prototype stage (Lightning Strikes - Coachbuild.com). Regenerative braking was invented as a way to increase driving range in battery-powered vehicles. Electric vehicles had been around for a while when the first regenerative braking system was developed. The first electric carriage was invented in 1832, and the first practical electric car wasn t invented until 1835 by Thomas Davenport (Timeline: History of the Electric Car, PBS.org, 2009), or 1839 by Robert Anderson (Berman, 2011). Both of these were powered by non-rechargeable batteries, as the first lead-acid storage battery was invented in By 1893 models of electric cars were being exhibited, and in 1900 twenty-eight percent of the cars produced in the United States were powered by electricity. Twenty years later the interest on them decreased, as gasoline became more available and incremented the driving range. It wasn t until 1966 when U.S. congress introduced bills recommending the use of electric vehicles and interest started to rise again. The first full-powered hybrid vehicle is built in 1972 by Victor Wouk, being the program under which it was developed discontinued four years later, the same year the Electric and Hybrid Vehicle Research, Development and Demonstration Act is built. The public is starting to demand more research on the possibilities of electric and hybrid vehicles, and in 1988 the CEO of G.M. starts to develop the EV1, teaming up with California s AeroVironment. However, the first commercially mass-produced and marketed hybrid doesn t appear until 1997, in which year Toyota presents the Prius. From that year until 2000, a few thousand different all-electric cars were manufactured. In 2006 Tesla Motors presented the Tesla Roadster, to be sold from That same year, BYD, a Chinese company, releases the F3DM, the world s first mass produced plug-in hybrid compact. In 2009, the Nissan LEAF was introduced, to be sold from (Timeline: History of the Electric Car, PBS.org, 2009) Most, if not all of the HEVs (Hybrid Electric Vehicles) and EVs (Electric Vehicles) use a system similar to the one developed and implemented in this project to improve the usage of energy. There are two trends on how to implement it and coordinate the hydraulic and regenerative braking related to the brake pedal position. 11

32 The first one is the parallel approach, which will be implemented. It consists on applying both regenerative braking and hydraulic braking at all times when the brake pedal is pressed. This approach is considered to be less effective that the series approach but safer, as the hydraulic brakes are constantly functioning and there is always a system on demand for emergency braking. The second one needs to be implemented along with electronically controlled braking, ECB, a system that will determine the driver s braking intentions and adjust the ratio between regen and hydraulic braking. Both types are blended, maximizing the amount of regenerative braking according to the constraints and applying hydraulic braking to complement and reach the calculated driver s intended braking torque. This approach is more complex and not viable for this project. Also, regenerative braking can be applied in the accelerator pedal. This is another widely used approach, which consists on applying regenerative braking when the accelerator pedal is let go or progressively less pressure is applied to it. This approach usually involves a very precise mapping of the engine and torque supply, and involves an adaption from the driver to a new way of driving. This is not a problem if the design is done properly, as the learning process is not too complicated and drivers usually get used to it quickly. This approach can roughly be explained by considering the pedal- speed relation to be a curve, and if the brake pedal position falls under the speed to which it is correspondent, regenerative braking will be applied until the relation is again met. Some of the most widely- known hybrid or electric cars and their approaches to the implementation of regenerative braking are the following: - The Tesla Model S and Roadster only apply regenerative braking in the accelerator pedal (Solberg, 2007). - The Toyota Prius applies regenerative braking when using the brake pedal. It calculates the amount of braking requested and supplements regenerative braking with friction braking, unless the ECU determines it is a panic stop situation. - The Chevrolet Volt uses both approaches. It applies regenerative braking when you either lift your foot from the accelerator or when the brake pedal is pressed. - The Honda Fit EV uses regenerative braking in the accelerator pedal. It has its own setting in the lever, B for braking, that allows regen to kick in (Ryan, 2013). - The Audi E-Tron has regenerative braking incorporated in both accelerator and brake pedal. The rear wheels are computer-controlled to match the friction brakes at the front wheels while applying regen, up to 0.45g (Horrell, 2013). However, this vehicle will not be commercialized due to its cost. - The Honda Insight has two braking modes, each one of them applies regenerative braking in a different pedal (Deceleration / Regenerative Braking Mode). 12

33 - The Nissan Leaf adapts the amount of regenerative friction braking applied depending on its calculations of how much braking force the driver intends to apply (dgpcolorado, 2013). - The Toyota Camry Hybrid has braking pedal regenerative braking only, attempting to feel as much as a gas-only car as possible for consumers (US News, 2014). Therefore, it can be seen that there are two main ways of implementing regenerative braking, and it depends on what it is desired to achieve with the car. In more sports-like cars, the trend is to apply regenerative braking when the accelerator pedal is let go. This way, there would not be a coasting sensation as there is with ICE (Internal Combustion Engine) cars, but this approach results on the battery being charged instead. It is an approach that changes the way the driver uses the accelerator pedal, and therefore does not please everyone for it is more aggressive. The second approach consists on implementing regen with the friction brakes. This can be done in two ways: by calculating the amount of braking the driver intends and creating that amount by using as much regen as possible and supplementing it with the friction brakes if necessary (only in normal braking, in emergency braking it would only be friction braking); or by applying both friction and regenerative braking at all times during normal braking. With the first of these brake-pedal approaches, it is necessary a very good estimator of the driver s intentions, as too much or too little braking is a risk. It is also a requirement that an ECB (Electronically Controlled Braking) system is operative at all times, for what intense testing to ensure safety in emergency braking situations is needed. Regenerative braking is better for city driving, and some of its main drawbacks are that it is mostly used along with friction brakes (except in the first approach as explained here, which does not feel comfortable for all users as they have to adapt to a new way of driving), or needs a very precise and complicated mapping of different variables such as speed and accelerator pedal position, that do not always follow an easily obtained correlation. For the same reasons that regenerative braking was developed, other systems have become popular. Examples of this are the KERS- Kinetic Energy Recovery Systems, and hydraulic braking. KERS use flywheels to store energy from deceleration, and can provide with large amounts of power for short periods of time. It has been used in Formula One racing with good results. Hydraulic regenerative braking is based on the use of high-pressure accumulators, in which hydraulic fluid is forced by using the kinetic energy from braking. This liquid is later used to power the wheels. This system is currently used by UPS shipping company in its trucks (Parker, 2013). 13

34 1.2 Motivation Gasoline has been increasing its price. Since 2004, an approximate increase of 192% has been recorded in the average retail price per gallon in regular gas in the U.S., (Gas Buddy, 2014). With the scarcity of fuel resources and the increasing concern in sustainability, the future clearly drives an electric car. This car of the future will rely on taking advantage of all of the losses that occur in day-today driving. This means that regenerative braking will be a key feature, and although regenerative braking has been developed and implemented for quite a while now, it is nowhere near completely exploited. However, in the topic of electric cars, there are still more improvements to be made, better efficiencies and better driver feel can be achieved. New approaches are to be developed. For now, the step to be taken is to improve hybrid vehicles. While batteries remain expensive and don t provide with the same driving range gasoline cars do, hybrids provide with a still sustainable- option that meets the requirements of the public. They do not depend as much in charging infrastructures, while for EVs better charging infrastructures have to be developed efficiency, and that is a work in progress. The development of hybrids, then, is the first step into developing better EVs. However, being hybrids a temporary solution does not make them less profitable. Plug-in hybrids, which run on the motor fed by the battery (which can be recharged by connecting the vehicle to an outlet or by the engine), provide with a solution to the charging infrastructures problem. Plug- in vehicles have been increasing sales since their appearance in the U.S. in 2010 at a high rate, with vehicles sold in 2013 (Cobb, 2014). This is more than 80% more vehicles sold than in 2012, what can be observed in Figure 1, where a comparison between sales of plug- in electric vehicles in consecutive years is made. 14

35 Comparison of plug-in hybrid sales in the U.S Plug-in hybrids sold by year Figure 1: Plug-in hybrid sales (Cobb, 2014) It can be seen the importance of increasing the range of any vehicle that relies on batteries, as it is one if the main challenges faced by the industry, and the most efficient system for that is regen braking. Regenerative Braking, therefore, provides with the solution to one of the things in the way of the final EV and HEV success, as it is one of the most important problems faced by the automotive industry in relation to this type of vehicles. There is still many improvements to be made and it can still be optimized. New, easier ways that provide with better results have yet to be investigated. More can still be taken from what was usually wasted in other forms of energy, more can be driven for free. 15

36 1.3 Background of the project This project was designed as part of the EcoCAR 2 competition. The technical goals of this competition are to design a vehicle that compared to the conventional gasoline car: - Reduces fuel consumption - Reduces well-to-wheel greenhouse gas emissions - Reduces criteria tailpipe emissions - Maintains consumer acceptability in the areas of performance, utility and safety. The students are required to explore different powertrain options in the 2013 Chevrolet Malibu (EcoCAR 2, 2011). The approach chosen by Embry-Riddle Aeronautical University was to implement a Series Plug-in Electric Vehicle (PHEV), using a lithium-ion battery pack from A123 and an AM-Racing, Remy Electric Motor that gives a 56 to 68 km range. That is implemented with a 1.7 L diesel engine and a reservoir of B20 biodiesel to maintain charge in the battery pack. This design provides with an estimated 105 MPGe (Miles per Gallon Gasoline), which is 44.6 km/l (EcoEagles, EcoCAR 2, 2014). As the main goal of the competition is to produce a car with reduced environmental impact, it was decided to design and implement a regenerative braking system, which will maximize energy usage. The rules of the competition related to Regenerative Braking that must be held for the final system require the following: 1. Brake Pedal feel is normal and the primary braking system is the hydraulic brake. 2. There must have a Regen Disable switch that is functional. 3. The ABS must be functioning at all times. These constraints were held satisfactorily by the implementation of a regenerative braking system in the brake pedal. The reason for this concrete system to be designed and implemented is that there is not a generic regenerative braking system available for any car to use. The torque necessary to stop each car is different, and so the parameters will also not be the same. Additionally, the control algorithms being used in commercialized vehicles are not available to the public. There are barely a couple indications on what signals could be used to develop such a system, and therefore a new approach had to be created to implement for the competition. The system was developed by using MathWorks software: MATLAB, Simulink and Stateflow. 16

37 1.4 Objectives The study of the methods being applied now to increase the mileage, sustainability and marketability of hybrid and electric vehicles will provide with an insight to what the industry is heading towards and how it is done. It will provide with experience in testing and development of test plans, as well as with coordination with different design teams. This experience will be highly valuable when working in a development environment and procedures related to the testing and implementation field will be learnt. This project aims to develop and implement one of the mentioned regenerative braking systems, creating the necessary systems and setting up the necessary parameters. It will not be purely theoretical, but results of the implementation of the approach will be studied and commented. A second system for the accelerator pedal will be also designed, but not implemented. This system will provide a completely new approach, which has not yet been investigated or implemented in the industry. This new system accomplishes the same objective as conventional regen approaches in the accelerator pedal without the mapping of the motor. The feel for the driver will be slightly different but it will still be a quick learning process, similar to the one required for the system being currently implemented, also known as floating zero. However, the main system discussed will be the one applied to the brake, as the main objective was to create a system that is successful with recovering energy. 17

38 1.5 Methodology Through studying the different approaches that can be taken and what is possible to do with the time and test limitations on the vehicle one first optimal RBS (Regenerative Braking System) will be designed. The design will take place in multiple parts: 1. Deciding the inputs and outputs: discerning which signals could be of use and why they will be selected, as well as justifying the use or not of the conventionally used signals for this type of system. 2. Modifying the model provided for the simulations to be valid: the model should reflect the main aspects of the vehicle dynamics while maintaining simplicity. 3. Fixing the relation between inputs and outputs: it will be decided how the system should behave and how to achieve that outcome while ensuring consumer acceptability and safety features. 4. Study of the benefits of the system implemented. The second system will be studied in relation to what applications and innovations presents, and the theoretical aspects of this approach. It will start by analyzing the relationship between input and output and will progressively introduce the final theoretical system and its constraints. 18

39 Chapter II: Development There exist several RB systems that could be implemented, divided into two types, depending on where it is applied, as it has been previously discussed in the introduction chapter. The main system discussed and implemented is focused on the brake pedal. The reason to choose this system is that it provides with a more conventional driver feel, and there is no need of adaption to it. Also, it requires a lot less testing in order to have something that works than any accelerator pedal related approaches. The system will not strive to provide with an aggressive approach, but rather something that is comfortable for the driver and regenerates energy without changing the way a conventional car would feel. For more conservative drivers this will be very appealing, as there should not be any major differences in how driving feels when compared to any conventional gasoline car, and energy will be recovered. Another reason to choose this system as the main one is that all of the systems implemented up until now applied to the accelerator pedal require a very precise and complete mapping of the torque, speed and accelerator pedal position, amongst others. This mapping involves intense testing, which was not available due to the nature of the EcoCAR 2 competition and state of the vehicle while this project was being developed. It was therefore decided to implement a RBS (Regen Braking System) that would use as an input the driver s brake pedal. Some of the inputs conventionally used for the design of this type of system are the following: 1. Brake Pedal Position 2. Power limit of the battery 3. SOC of the battery 4. Torque available from the powertrain or battery 5. Pressure of the Friction Brakes 6. Friction torque applied 7. ABS on 8. Speed Another constraint added was that the vehicle was being engineered and fixed, and it was not 100% functional most of the time in which the system was being developed, as well as that the signals that were accessible were not all of the usual signals conventionally considered as inputs for this type of system. For an instance, there was no access to the pressure of the brake fluid, which is used to provide with a softer feel. This translated into a more simulation-concentrated approach for the system before implementation, as well as the usage of a different set of inputs for the system developed. These new inputs were chosen to adapt to what a normal braking feel is like and predict user intentions. The main difference with the conventional RBS is the introduction of a new 19

40 variable, the brake pedal position gradient, to distinguish between when the driver wants to slightly brake or stop the car as soon as possible. 2.1 Approach The obtaining of the signals is done through the CAN2.0A bus provided in the inverter. It follows little-endian format, which has the LSB as the lowest address. The inverter is a RMS PM150 (which has a 150kW 3-phase inverter, with a 450A peak, 450A continuous current rating at 300V input, input range VDC, (EVHANGAR, 2013)). A system to incorporate regenerative braking when pressing the brake pedal was designed, using as inputs for the system brake pedal position, brake pedal gradient, regenerative braking enable and current speed of the car, to obtain the regenerative demanded torque. Figure 2: Main view of the RB system Inputs one through three in Figure 2 are directly obtained from the data recorded from the car, and input four is generated to decide whether regenerative braking can or cannot be applied. 20

41 2.2 Inputs: Speed The speed is one of the most significant inputs to the system, as it is the variable from which the first torque command estimation will be developed. The speed of the car is received as an input in m/s, and converted into km/h. It is then used as an input for a lookup table that determines regenerative torque as a function of speed. One of the curves implemented in the car can be seen in Table 1. Table 1: Regenerative Torque as a Function of Speed Speed Torque

42 Torque as a function of speed y = x x x x x R² = Torque as a function of speed Poly. (Torque as a function of speed) Figure 3: Regenerative Torque as a Function of Speed Curve The curve in Figure 3 has its maximum negative torque around 70 km/h (43.5 mph), which is an approximated average value for optimal fuel efficiency (Fuel Economy, n.d.). In this curve, this value was chosen as the correspondent to maximum torque commanded, to shape regenerative braking as speed varies. When the car is going at high speeds, the amount of RB will increase as the car slows down, up to when the car reaches approximately 70 km/h, point at which the amount of regenerative torque commanded will start to decrease, creating a comfortable feeling in the driver. In city driving, the amount of regen torque will not be high, to provide with the same sensation as driving in a normal car Brake Pedal Position The system should take into account the brake pedal position, as the response should not be the same when a 10% pedal displacement or a 90% pedal displacement are applied. The further the braking pedal is pressed, the more regenerative torque should be applied. This will be implemented by multiplying the torque obtained as a function of speed by a factor obtained from the brake pedal displacement. It was decided to normalize the brake pedal position signal by dividing it by its maximum, 100. Then a dead zone and a saturation were applied to the signal, as it is not desired that the system shows a lot of variation at the limits of the dynamic range. The range was then limited to from 10% to 95%. This also eliminates the bias introduced by the brake pedal position sensor, which results on the signal having its minimum at approximately 4%, as it can be seen in Figure 5, and removes any problem due to SNR in low values of the signal. 22

43 The final output from the brake pedal position subsystem is then a positive number inferior to 1 which will depend on the input, giving the same output for inputs from 0 to 10%, and from 95% to 100% Brake Pedal Position Gradient To take into further account the driver s intentions, the brake pedal position gradient signal was introduced. The outcome of this will be a different final regenerative torque if the driver presses the brake pedal fast than it would be if the driver presses it slowly. The reason for this is to anticipate if the driver is going to want to stop the car fast, in which case a high amount of regenerative torque would be applied, or slow, in which case the torque applied would be significantly inferior. In this subsystem, it was decided to saturate the signal under 100 degrees/second, as the difference in the torque applied will not exist between different slow braking speeds. Also, this way the noise of the signal at its low values is removed. It is then normalized to obtain a percentage (divided by the upper limit of its dynamic range, 360 degrees/second). This percentage will then be added to a constant of 1 and multiplied to the torque obtained from the brake pedal position and speed of the car. The reason for it to be added to 1 and then multiplied is that the desired outcome of this subsystem is to increment the amount of regenerative braking torque as the braking speed increases. 23

44 Figure 4: Brake Pedal Position Gradient Test Data Figure 5: Brake Pedal Position Test Data In Figure 4 and Figure 5 it can be seen that the noise usually stays under 100 degrees/ second, as stated earlier. This can be observed as in the brake pedal position signal (Figure 5) the signal is low or close to its minimum, and the brake pedal position gradient varies. 24

45 2.2.4 Regen Enable The Regen_Enable signal will have as a main function to control whether regenerative braking is applied or not. The criteria to decide this is the following: Regenerative torque will be applied if: 1. The SOC (State of Charge) of the battery is under 99%, therefore the battery can be charged. This parameter is easily changed. The reason why this percentage was chosen is that the life of the battery will not be evaluated in the competition, therefore, it is not necessary to optimize this parameter in relation to that constraint. It will be considered that as long as the battery can be charged, it should be charged, without taking into consideration how that might affect its life. 2. The user has enabled regenerative braking. This is done for both security and user preference reasons. 3. The ABS system is off. This constraint is related to safety. The reason for this last condition is that the simultaneous use of regen and ABS can cause breaking of the driven shafts. The solution to this problem was to discontinue regenerative braking every time ABS kicked in. Another approach that was considered was to predict when ABS was about to start and turn off regen before that. However, this approach is far more complicated and unnecessary. In the finite state machine that controls whether regen is being enabled or not, the sample time is seconds, which is not enough time for the system to react and cause any damage because of the interference of both systems. This approach was implemented by the design of a finite state machine using StateFlow (Mathworks) that can be seen in Figure 6. The inputs to the system will be the signals derived from those three conditions (SOC_High, Regen_Disable, ABS_On), and the output will be Regen_Enable. Figure 6: Main view of the finite state machine As it can be seen in Figure 7, the finite state machine has two states, On and Off, that make the output be either 1 or 0, depending on the inputs. 25

46 Figure 7: Finite state machine The first state consists on having regenerative braking off, and will be the default state. This is done for security reasons and because the system should only apply regenerative braking when all of the conditions are met, and it is only guaranteed if it is done this way. The change between both states occurs when: 1. From OFF to ON: if the driver enables regen, the battery is under 99% charge and the ABS system is off. 2. From ON to OFF: if either the driver turns regen off, the ABS kicks in or the battery is charged over 99%. The finite state machine was simulated, obtaining the results in Table 2. 26

47 Table 2: Finite State Machine Simulation Results ABS_On Regen_Disable SOC_High OUTPUT It was then concluded that the behavior of the finite state machine was satisfactory. Also, the rise and settling times will not be analyzed, as the speed of the reaction of the system is of no special concern, as even adding it to the time the system will take to disable regen, it will be in every case inferior to the time the car takes to react to any change. 27

48 2.2.5 System The combination of these inputs was then modeled with Simulink. The system also has a saturation to limit the demanded torque only to negative values (superior to -600 Nm, number chosen because it is more negative than the maximum negative torque the motor can supply, around -311 Nm). This can be seen in Figure 8. Figure 8: Simulink model of the system In Figure 8 it can also be seen the previously described signal treatments. The torque command is interpolated from the values of the curve being used and the multiplied to the final value of the other two signals to conform the final regen output that will be selected every time regen braking should be enabled, otherwise zero regen will be applied. For this system it was important to consider driver feeling when deciding about the curves that were to be chosen to shape the amount of torque applied. The main signal in which driver feeling can be estimated is deceleration rates. The parameters that will be used for the deceleration rates are the following: a. Comfortable deceleration is around 3 m/s 2. For 90% of the drivers, a comfortable deceleration rate is 3.4 m/s 2, (Bureau of Local Roads & Streets, 2006). b. Maximum deceleration rates for most vehicles range from 5 to 10 m/s 2, (Halmerman & Martin, 2011, 2009). 28

49 A model of the car will be used to simulate these rates and obtain a curve to implement in the car. These simulations were done with input data recorded in the car when performing test drives. It is important to take into consideration that there are also other systems in charge of preventing and dealing with faults that can result in the destabilization of the regenerative braking system. These systems deal mainly with prevention of problems with the battery and the inverter and include from current to temperature limitations, none of which will be treated for the development of this system, as they have already been taken into account in other systems and are included in the Regen_Enable signal behavior. 29

50 30

51 Chapter III: Model In order to be able to simulate and later implement a system, it was necessary to develop a model. Instead of creating a completely new model, it was decided to modify a model for the Tesla Model S developed by Derek Bonderczuk, Steven Fox and Hasnaa Khalifi. The main reason for this is that creating a completely new model with the time constraints the project had would put to risk the quality of the regenerative braking system, the main goal of the project. Therefore there exist some inexactitudes in the model compared to how the car reacts, as some of the parameters used are not the optimal. However, the dynamic equations are the same, and this will only affect the numbers, not the trends, so it was decided that to continue using this to simulate while considering what the differences might be. The original system can be seen in Figure 9. It is divided into three blocks, and the input is a drive cycle. The three blocks are: Derived Tractive Force, Powertrain and Vehicle Dynamics. The Derived Tractive Force block generates the tractive force that the car needs to follow the speed commanded, the Powertrain subsystem incorporates all of the internal variables in the car, such as battery, motor, loads etc. The Vehicle Dynamics block calculates all of the external actuators on the speed of the car, to provide an actual velocity that will be used as feedback to both of the other subsystems. Figure 9: Main view of the original model This system was modified so that the input was the torque command, as the inverter will be running in torque mode. This was done by eliminating the Derived Tractive Force block, as the calculations performed in it are not needed. Therefore, the Velocity Command that served as an input for this system was eliminated as well. Also, this model only considers one type of braking: regenerative. This had to be modified as hydraulic braking will be applied in the real vehicle and this will modify the behavior of the regenerative braking system. For this, another input was introduced: brake pedal position, and with data from different braking events an approximate model of the hydraulic brakes was developed. The final system that was used can be seen in Figure 10. The list of inputs and outputs of the model can be seen in Table 3. The throttle command will be multiplied by maximum 31

52 torque to obtain the torque command, and it was decided to create three main outputs: torque, velocity and acceleration. The actual torque will be used as a feedback to the system for the controller, and velocity and acceleration will be used to monitor the correct functioning of the system, as well as to compare with the real data obtained from the vehicle. Another thing that had to be taken into account is that velocity is used in the Powertrain block to calculate the rpm of the motor, which will be an important constraint when simulating for positive torque because of the revolutions limiter. Figure 10: Final model for the car Table 3: Inputs and outputs of the model. Inputs Throttle Brake Pedal Position Outputs Applied Torque Velocity Acceleration 32

53 3.1 Powertrain The powertrain subsystem is what is in charge of emulating the mechanism that transmits power to the axis, and one of the most important systems in the model. Table 4: Powertrain inputs and outputs System Inputs Outputs Powertrain Actual Velocity Actual Tractive Force Commanded Force Its inputs and outputs can be seen in Table 4. It is also divided into different subsystems, as seen in Table 5. Another observation about this subsystem is that it also had to be modified by eliminating the throttle calculation from the commanded force that was in the original system, as throttle is no longer an intermediate variable but an input. Table 5: Inputs and outputs of the Powertrain Subsystem Inputs Outputs Transmission Velocity RPM Motor Voltage Current Throttle Torque RPM Maximum Possible Force Battery Low Voltage Current Voltage of the Battery Motor Current Auxiliary Loads Throttle Auxiliary Current Battery Voltage Tractive Force Torque Tractive Force Calculation Maximum Possible Maximum Possible Tractive Torque Force 33

54 3.1.1 Transmission The transmission block converts linear speed into angular speed of the motor by using Equation 1. The velocity is in m/s and the parameters can be seen in Table 6. RPM = Equation 1 Velocity Gear Ratio wheel radius Table 6: Parameters of the transmission Motor Parameter Value Wheel Radius (m) 0.35 Gear Final Drive Ratio 9.73 The motor was modelled as a zero order system because it was considered that the lag of the motor would not affect significantly the design of the RB (Regenerative Braking) system. The safety and activation modes of the RB system can be developed without taking this into consideration, because the behavior will not differ greatly, as the main factor that will introduce a lag will be the inertia of the vehicle not the dynamic behavior of the electric motor. Regen braking will therefore be considered as a steady-state operation. Equation 2 Maximum Torque = Equation 3 Power Angular velocity Torque = Throttle Maximum Torque Equation 4 Current = Power Voltage In Equation 2, Equation 3 and Equation 4 it can be seen how the outputs are calculated from the inputs, in which angular velocity is in rad/s, torque is in Newton per meter, power is in Watts, throttle is a percentage, current is in Amps and voltage is in Volts. 34

55 3.1.3 Battery In the battery block the current from auxiliary loads and from the actual speed are used to obtain the voltage in the battery. The model used for the battery is based on a model developed by Olivier Tremblay, Louis- A. Dessaint and Abdel-Illah Dekkiche in A Generic Battery Model for the Dynamic Simulation of Hybrid Electric Vehicles (Tremblay, Dessaint, & Dekkiche). This model uses only the battery SOC (State-Of-Charge) as a state variable, as it has been shown that this model can represent multiple types of batteries. It consists of a controlled voltage source in series with a resistance. The voltage of the battery is calculated from the current, assuming the same characteristics for the charge and discharge cycles, and calculating the open voltage source from the actual SOC of the battery. The voltage in this model follows a non-linear relation that only depends on battery charge Auxiliary Loads In this subsystem, other possible electrical loads are considered, by introducing a power load factor, derived from the power being consumed by starting the engine, igniting the fuel, operating the lights, etc. Due to simplicity and time constraints, it was decided to introduce this value as a constant load. The final output of this system, auxiliary current, is calculated as shown in Equation 5. Equation 5 Auxiliary Load Power Consumed (W) Auxiliary Current (A) = Battery Voltage (V) The existence of auxiliary loads will be one of the reasons why the system will not be modelled for current or voltage, as it would add time to develop a system that takes into account a variable auxiliary load factor. It was considered that the importance of those variables is not the same as the value of having the right torque output, therefore the variables derived from the battery will have the only function of contributing to the final and right torque estimation. Therefore, this will be a model for torque, not for voltage or current. 35

56 3.1.5 Tractive Force The tractive force is calculated following Equation 6, using the parameters in Table 7. It will be an output to the system that will be used to calculate a balance of forces in order to estimate the total force and velocity of the vehicle. Equation 6 Tractive Force = Equation 7 Torque at axle Wheel radius Torque at axle = Torque Gear Drive ratio Gear Drive Efficiency Table 7: Parameters for the tractive force block Parameter Value Gear Drive Efficiency 0.98 (per unit) Gear Drive Ratio 9.73 Wheel Radius 0.35 (m) 36

57 3.2 Vehicle Dynamics The main function of the vehicle dynamics block is to provide with a balance of the forces interacting in the vehicle to obtain speed and acceleration. Inputs Outputs Tractive Force Velocity Acceleration Figure 11: Inputs and outputs of the vehicle dynamics subsystem It calculates the rolling resistance, aerodynamic drag, slope force and braking force, what along with the tractive force output a sum of forces, and divided by the vehicle mass gives the acceleration. Table 8: Formulas for "Vehicle Dynamics" Force Formula- Units needed in SI or units indicated in Table 9 Rolling Resistance Rolling Resistance Force Force = Rolling Resistance coeficient Vehicle Mass Gravity Aerodynamic Drag Aero Drag = 0.5 drag coeficient rho frontal area Force velocity 2 Slope Force No slope in the driving circuit is considered Braking Force Braking Force = Brake Pedal Position Gain Table 9: Parameters for "Vehicle Dynamics" Parameter Value Rolling resistance coefficient Vehicle mass (kg) 2170 Drag coefficient (Cd) 0.24 Rho (Air density, kg/m3) Frontal Area (m2) 2.82 Theta (slope angle) 0 Gain The final force is calculated as shown in Equation 8. The gain that is part of the braking force will be subject to modifications to adapt the behavior of the model to that of the car and will be later discussed. 37

58 Equation 8 Total Force (N) = Tractive Force Rolling Resistance Aero Drag Braking Force By using the Equation 9, the acceleration is calculated, and by integrating it velocity is obtained. Equation 9 Acceleration ( m s 2) = Hydraulic Braking and validity of the model Total Force (N) Vehicle Mass (kg) As the objective of this project is to develop a system that stops the car, it was decided to make sure that the hydraulic brakes had similarities with respect to the real brakes in the vehicle. Since the car that has been modeled and the EcoCAR 2 vehicle are not the same one, in the first simulations the comparison of the speeds from both were not similar, as it can be seen in Figure 12, even after achieving similar slopes for the part of the curve in which the car was braking. Figure 12: Speed comparison of the model and car before changing parameters 38

59 It can be seen in Figure 12 that the model does not reach the same speed as the vehicle when the speed is low, and when the speed is high, it has a superior value. It was decided to change the parameters of the model to try to approximate the speed of the model as much as possible to that of the car. Although the new values are not real, a more accurate simulation is generated from them. The new speed comparison can be seen in Figure 13. Figure 13: Model and car speed comparison It can be observed in this new simulation that speeds match much better. However, there are still some differences: the model stops faster than the car. Comparisons in the speed simulated will be generated as an auxiliary factor of decision to establish parameters. They will not have much importance and will not be further discussed. The main and most important thing it will have related will be accelerations, in which the real acceleration experimented in the car will be significantly inferior to that of the model, as the hydraulic braking system developed for the simulation is stronger than the one in the vehicle. Therefore although there are some differences in the simulation, it will be considered positive that the model reaches a stop before the car does. It was decided that the current similarity was enough for the development of the system, as for the simulations the speeds used as inputs will be those recorded from the car. In torque, however, since it is the main input with which the system will be developed, it is important that both correspond almost to perfection. This can be seen in Figure

60 Figure 14: Torque comparison between the model and the car The accuracy of the new simulations and validity of the model can be seen in the similarity of both curves in Figure 14. This model will therefore be used to simulate and develop a regenerative braking system. Voltage and current have not been analyzed in this comparison as their values will not match those of the vehicle. The reason for this is that the model of the battery does not reproduce with exactitude the battery used. Also, there will be auxiliary variable loads that have been simplified in this model. 40

61 Chapter IV: Simulations In this chapter the model with the inputs described will be simulated. The main objective of this is to observe an approximation of how the car is going to react to the speed-torque curves, taking into especial account the torque commanded and accelerations. The data seen in Figure 15, Figure 16, Figure 17 and Figure 18, was extracted from the car and used to feed the inputs of the system. Figure 15: Brake Pedal Position Test Input Figure 16: Brake Pedal Position Gradient Test Input 41

62 Figure 17: Speed Test Input The Regen_Enable signal is not being used, as it is not of importance for the behavioral study of the system, but to enable or disable it. Also, the torque command from the driver was also used as an input, since regenerative braking is only used to stop the car, and in order to stop the car it must be running first. It can be observed that this data is only positive, as no regenerative braking system has been implemented in the car yet. Figure 18: Torque Command Test Input The car has an internal controller to command torque, that is part of the inverter and it will be considered that its output is going to be the commanded torque with a lag. A continuous 42

63 PID controller was developed instead of modelling this complicated controller. Its main function will be to set the model to the same torque that was commanded in the car. This will also influence the outcome of the simulations, but will not introduce great changes as its time constant will be in any case inferior to that introduced by the inertia of the car, therefore the error could be disregarded. The reason to choose this method is that developing a complex controller that simulates the behavior of the internal controller would take a long time and many variables. This does not make sense, as the objective of this project is not to elaborate a precise model of the vehicle. The objective of this simulations is to give a general idea of how the car will behave in order to implement the best solution. The only constraints considered for the design of this controller were to make a stable controller with a reasonable response time, which was at least similar to that of the inverter. It was considered that a rise time of 0.07 seconds and a settling time of seconds was enough, as when simulating this for torque command and comparing it to the actual response of the car the result was very close. This can be seen in the positive part of all of the simulations for torque. Figure 19: View of the delay of the system These inputs were simulated for 500 seconds, for the following curves shaping the controller that can be seen in Figure

64 Torque (Nm) REGENERATIVE BRAKING SYSTEMS 0-50 Curves for the controller Curve 1 Curve 2 Curve 3 Curve 4 Curve 5 Curve Speed (km/h) Figure 20: Curves for the controller It was decided to match the maximum regenerative torque with two different speeds, for future implementations and comparisons in the car. The two speeds selected where 70 km/h, because that is an approximate speed at which the car reaches optimal fuel efficiency, and 50 km/h, because that is a common speed for urban driving. The two basic curves created were scaled by different factors to create two sets of curves with different regenerative braking torque to speed ratio (1 through 4 and 5 and 6). The aim for all of these simulations was to find out which curves could have a wider consumer acceptance, and therefore implement it. However, this can also be seen as a set of curves that could be available for the driver to choose, and therefore adapt the regenerative braking system to the user s preference. In order to analyze the outcome of the simulations, it should be considered that this data was generated trying to stop the car as fast as possible, therefore it will reflect the behavior of the system for its maximum values. 44

65 4.1 Curve 1 For the creation for this curve the aim was to create a conservative curve that would establish the first values for which regenerative braking was possible, and on which other curves would be based upon. The shape of the curve was created so that, depending on the speed correspondent to maximum torque, if the speed was superior, it would provide with more regenerative torque as the speed got closer to the maximum to then start decreasing as the speed went towards zero. That way, at the beginning, not too much regen torque would be applied so it would provide with a comfortable feeling to the driver, and then progressively more regen would be applied until the car reached a stop. Table 10: Curve #1 Speed Torque The results in torque (compared to the torque commanded in the test data) and acceleration can be seen in Figure 21 and Figure

66 Figure 21: Torque Output for Curve #1 Figure 22: Acceleration Output for Curve #1 In Figure 21 it can be seen that in the event of hard braking, regenerative braking torque up to -180 Nm is applied. In Figure 22 it is observed that deceleration with regenerative braking reaches -6.5 m/s 2, what is more than the -3.5 deceleration. This will, however, not be considered too high since in this set of data the driver was intending to stop the car as quickly as possible, and the maximum absolute value is less than 10 m/s 2. It can also be seen that in the non-hard braking events the maximum deceleration rate is greater than -3 m/s 2, which is a comfortable deceleration rate. This means this approach 46

67 should be comfortable for the driver and an even more aggressive approach should still feel good. Also, as it has already been mentioned, these accelerations will be in every case inferior to those of the real car, as acceleration was modelled so that the model provides with the worst case scenario. 4.2 Curve 2 This curve was generated by incrementing the amount of torque demanded for each one of the speeds proportionally to Curve 1. When using Curve 2 as an input it can be seen that the result is more regenerative braking torque than by using Curve 1, with the amount of regenerative braking being close to -60 Nm more in this case. The deceleration rate also increases by roughly -1 m/s 2, as it can be seen in Figure 24. This approach should still be comfortable for the driver. Table 11: Input Curve #2 Speed (km/h) Torque (Nm)

68 Figure 23: Torque Output Curve #2 Figure 24: Acceleration Curve #2 This approach increases the regenerative braking to acceleration ratio. We are achieving more regenerative braking without increasing the acceleration in the same amount. This trend continues to increase through the next curves. However, the increase in acceleration has a limit to maintain consumer acceptability and that limit is reached in curve #3, #4 and #6, being approximately the same as with this one in curve #5. 48

69 4.3 Curve 3 This curve shows an increased relation between speed and torque following the same trend as in Curve 1 and 2. It can be seen that with the data in Table 12, both the regenerative torque and acceleration increment, with the regenerative torque close to the theoretical maximum (311 Nm) as it can be seen in Figure 25. In Figure 26 it can be seen that the difference between the acceleration of Curve #2 and Curve #3 is not too significant compared to the increment in torque, but significant enough to forecast that this curve should not be as comfortable as the previous one. Table 12: Input Curve #3 Speed (km/h) Torque (Nm)

70 Figure 25: Torque Output #3 Figure 26: Acceleration Output #3 This curve would nearly reach the theoretical maximum value that should be implemented, and the expectation for this curve when implemented would be not to be comfortable for the driver, as a very high amount of negative torque is being applied very fast. However, a curve 50

71 with a higher torque demands will be simulated to observe if the relation between the torque and acceleration variables changes and what would the maximum acceleration be. 4.4 Curve 4 This curve has the same shape and the three previous ones, but the torque demanded for each speed increases again. It can be seen that since there is a limitation in the amount of torque to be applied in the system, the final result does not differ a lot when comparing Curve# 4 and Curve #3, but if the behavior of the system when the intent was not to stop the car but to reduce its speed was compared, it would be seen that in this last one the absolute value of the amount of torque applied for the same speeds would be greater. Table 13: Input Curve 4 Speed (km/h) Torque (Nm) The maximum amount of regenerative torque is applied (Figure 27), and with that maximum the correspondent in acceleration is reached. It can be observed in Figure 28 that the maximum acceleration is not too high (its absolute value is inferior to 8.5 m/s 2 ), and that there is not a big improvement between this approach and the previous one. The only thing that would differ is that more regenerative torque would be applied for the same speed, and that would alter negatively the driver s feel. 51

72 The maximum amount of regenerative torque is applied (Figure 27), and with that maximum the correspondent in acceleration is reached. It can be observed in Figure 28 that the maximum acceleration is not too high (its absolute value is inferior to 8.5 m/s 2 ), and that there is not a big improvement between this approach and the previous one. The only thing that would differ is that more regenerative torque would be applied for the same speed, and that would alter negatively the driver s feel. Figure 27: Torque Output Curve #4 Figure 28: Acceleration Curve #4 52

73 4.5 Curve 5 This curve belongs to the second set of curves, that is not shaped like the first one. Instead of designing it with the maximum regenerative braking torque for 70 km/h, it was decided to center it around 50 km/h. Table 14: Input Curve #5 Speed (km/h) Torque (Nm) It can be seen that the maximum regenerative torque that results from this approach is greater than for Curve #1 in Figure 29, with a value around -250 Nm, which was expected as the maximum value of torque that can be commanded using this curve is greater than in the very first one. The output is also similar to the one obtained with Curve #2. Deceleration doesn t reach the maximum values for this curve stated earlier, so the driver feel should not be aggressive. 53

74 Figure 29: Torque Output Curve #5 Figure 30: Acceleration Output Curve #5 When comparing this approach to the one in Curve #2, Curve #5 provides with a softer feel as if the speed from which braking occurs is greater than 50 km/h, less regenerative torque is applied at the beginning, as it can be seen in Figure 33: Comparison of Outputs. 54

75 4.6 Curve 6 It can be seen that this approach is less conservative than the previous one, although the curve is shaped in the same way. This approach is more alike the approach taken in curve 4, as it can be seen in both Figure 31 and Figure 32. A driver would, however, be able to distinguish between both, as for example, if the speed from which braking starts is greater than 50 km/h, this approach would provide with a slower increase in the amount of braking torque, while with the other one, the maximum braking torque for that speed would be applied at the beginning, decreasing progressively afterwards. Table 15: Curve 6 Input Speed (km/h) Torque (Nm)

76 Figure 31: Torque Output for Curve 6 Figure 32: Acceleration Output for Curve 6 56

77 4.7 Comparison Putting all of the simulated results in the same plot as reflected in Figure 33, it can be seen the two different shapes in curves. The first shape is the one that comprises curves 1 through 4, being the other two from the other approach. This can be observed as the maximum regenerative torque is applied at different speeds in both trends. A comment in the values of this regenerative braking torques should be made, as these are not the real ones that will be applied in the car when stopping. The system will have an impact on how the driver uses the brake pedal, varying the position if too much braking is applied. This being said, the trend of the output will be the same, so to first simulate, compare and evaluate the validity of this curves this approach will be valid. Figure 33: Comparison of Outputs Another way some of these curves could be used would be to enable the driver to choose between them. That way the user would have control over how the car should feel and what amount of regenerative braking feels comfortable, as a too conservative or too aggressive approach would not be of the liking of all of the drivers. 57

78 58

79 Chapter V: Results It was decided to implement a RBS in the car. The first implementation was made with Curve 1. The reason for this is that it was decided that it was important to ensure that the system worked and was not too aggressive, and once tested, change it accordingly to driver feel if it was necessary. Also, no regenerative braking system had been implemented in the car prior to this, and trying first the most conservative approach seemed reasonable. The system composed by the combination of what can be seen in Figure 2 and Figure 6 was implemented in the EcoCAR, and a dynamometer test to ensure its safety and a test drive were performed, from which the data to be analyzed were extracted. The main observation to be made is that the driver feel was very comfortable and unless the people in the car were told that regenerative braking was on, it was hard to guess that something other than what usually takes place in an ICE car was happening. The driver could feel that more braking torque than usual was being applied, but it wasn t a problem. Therefore, the most conservative drivers would be comfortable with this approach, what is positive, as one of the main objectives of the EcoCAR 2 competition is to maintain consumer acceptability. The data from the test drive was studied, starting at making sure that regenerative braking had been effectively working and comparing the different results. There will be three main parts of study: 1. Torque 2. Voltage and Current of the battery 3. Energy and Power Before getting into each one of them, the whole of the system will be analyzed. In Figure 34 it can be seen a plot of the speed, torque and brake pedal position in a portion of the drive. It can be observed how those main variables correspond, and how the speed decreases as regen kicks in when the brake pedal is pressed. It is important to consider that all of the data recorded for this chapter is concentrated on creating braking events, and will not reflect exactly what would happen in a normal driving cycle. The speed is maintained each time around 5 seconds, sometimes more and sometimes less, and then braking until a reduced speed or zero speed occurs. The maximum speed is also low, it does not reach 50 km/h. Therefore the energy consumption will be inferior to 59

80 that of a normal driving cycle with the RBS implemented, as more braking than usual will take place. 60

81 Figure 34: Speed, Torque and Brake Pedal Position. Data from the car. The maximum of the speed is reached between the end of torque application and the brake pedal being pressed. It can also clearly be seen that there is a bias in the brake pedal position signal, as was taken into account when designing the system. There is a correspondence between regenerative torque and brake pedal position, in what the first one is shaped accordingly to the behavior of the second signal. Figure 35: Closer look at Speed, Torque and Brake Pedal Position 61

82 5.1 Torque analysis Torque analysis in the inverter To have a better understanding of the system, the torque command and torque feedback data was plotted, as it can be seen in Figure 36. This is done in order to analyze if an adaption to what the inverter is doing when the torque decided in the system is commanded is needed. The difference between these signals is that the Torque Command is the torque that the inverter thinks is being applied, while the Torque Feedback is the actual measured torque applied. Figure 36: Commanded Torque and Torque Feedback main view In Figure 37, it can be observed that there is a difference between torque commanded and feedback. However, in Figure 38 it can be seen that there it is not a significant delay between both signals in the rising (for positive torque) and falling (for negative torque) edges. It is interesting to point out that the feedback signal tends to fall apart from the command as the torque increases in positive torque and decreases in negative torque commands, tending to go to zero and anticipating what the signal was going to do. This could be either due to noise in the sensor that increases with the absolute value of the signal, or due to the revolutions limiter. Even though this occurs in nearly every command, the signal tends to follow the same shape in the curve as the command, so this inexactitude will not be considered as a problem to be fixed, but as a consequence of the internal controller of the inverter that cannot be changed. Also, the torque feedback signal has a lot of noise and it is hard to determine the exact amount of measured torque being applied to the car. A significant observation is that, in the 62

83 regenerative braking part, the measured torque being applied has quite a lot of overshoot, reaching more than -90 Nm when the commanded torque was around -60 Nm. This does not look like it is part of the noise, as there are several points that follow the same trend. Figure 37: Commanded and Feedback Torque closer view The overshoot, however, only lasts for one or two sample times, being therefore not noticed when driving, as the car takes longer to respond to such a large torque being demand than the time that torque feedback is applied. 63

84 Figure 38: Detail of regen Torque Command and Feedback In Figure 38 it can be more clearly observed the difference between the input torque to be commanded at the inverter and the torque commanded by the inverter as an output, which, as said before, is a difference influenced by the internal controller of the inverted, which cannot be changed. It will mainly affect the regenerative part of the torque, as it does not reach the same values commanded. Although there is an obvious difference, it was decided not to take it into account, as by applying less regen than expected, no safety restrictions are compromised, and the final outcome will not be significantly different for the driver s feel. 64

85 5.1.2 Real and theoretical torque command comparison The difference between what the Simulink system would consider as torque to be applied for the input signals from the car and the real command, derived from the analysis of the inputs and the Torque Feedback signal, will be analyzed. It is important to make sure that what has been simulated until now does not differ greatly from the real command. This is done mainly to reconfirm the validity of the simulations. Figure 39: Theoretical and Real Torque Figure 40: Detail of Theoretical and Real Torque This comparison was generated by introducing the inputs used in real time in the car in the model, and comparing the results obtained from that with the data recorded in the car. The difference is only in the final values, and it can be seen in Figure 39, Figure 40 and Figure 65

86 41 that the rate of change is similar in both. There is a difference, as expected, due to the effect of the feedback signal. Figure 41: Closer look at regen part of Real and Theoretical Torque With this analysis it is concluded that the system works as it had been predicted, and that, although there are differences between the theoretical and real values, it will not affect the behavior of the system significantly. 66

87 5.2 Voltage and current analysis The next step is to analyze the impact of this approach, from a power and energy point of view. Analyzing the current in Figure 43 it was observed that, as it would have been expected, current switches sign as torque does. When regen is active, current is fed to the inverter to charge the battery, while in the opposite case, the battery functions as a supplier of energy for the rest of the systems. In this section an analysis on current and voltage in the battery will be done. This, however, will not be compared to the results obtained in the model, as it was modelled for torque, not for intensity or voltage. One of the reasons for this is that there will be a current derived from auxiliary loads that will be variable and therefore difficult to predict with exactitude. It is also important to consider regarding this analysis that there already exist current and voltage limitations that have been taken into consideration in other systems implemented in the car, as well as other type of limitations such as temperature of the battery. These limitations include limiting the current input to that which the battery can accept, as it depends on the SOC and other variables. The case in which this or any other safety requirement was not met would raise a fault, and the adequate system to stop the car or perform other actions would be enabled (or disabled, such as the Regen_Enable signal). Therefore, this section will only comment the general aspects regarding current and voltage being applied at the battery when regenerative braking is enabled. Figure 42: Torque Compared to Current 67

88 Figure 43: Torque Compared to Current Detail It can be seen in Figure 42 and Figure 43 that the current increases (the sign criteria is: positive current exiting the battery, negative otherwise) as the load increases when more torque is commanded, while current turns negative when regen is being applied, and therefore the battery is charged. It is surprising the great correspondence that can be seen in the amount of torque applied and current in the battery, as the shape and values nearly correspond in both cases. Figure 44: Voltage comparison with torque 68

89 The voltage maintains a more or less constant value of 290 V as it can be seen in Figure 44 and Figure 45, with small fluctuations (the biggest ones are around 20 V, less than 10%) as different amounts of torque are demanded. It can also be seen a correspondence in Figure 45, as whenever more torque is commanded, the system reacts varying the voltage. Figure 45: Voltage comparison to torque in detail 69

90 5.3 Energy and power analysis It is important to generate an energy and power analysis, as this is mainly an energy recovering system. It is also important to comment on the efficiency of this system, to prove that its development and implementation is justified. However, it must also be taken into consideration that this system is not the optimal system for energy recovery, but a system that could be implemented for that purpose with the constraints posed by the EcoCAR 2 competition. The main goal of this RBS is not to maximize energy recovery, but to provide with an energy recovery system that is simple enough not to need constant testing that maintains safety and optimizes energy usage within the constraints posed. It must also be considered that the driving cycle performed with regenerative braking enabled will not be a regular driving cycle, as the speeds reached and the abundance of braking events will not the norm in regular driving. First, the power and energy signals were calculated using Equation 10 and Equation 11. Equation 10 Power of the battery(w) = Voltage (V) Current (A) Equation 11 Energy of the battery (Wh) = Power (battery)(w) dt = Power (Battery)(W) time (sample time in seconds/3600) Analyzing the power signal it was found that, as expected, power was generated (negative according to the sign criteria chosen) when negative torque was applied, reaching values of 70 kw in this portion of data as it can be seen in Figure

91 Figure 46: Power compared to torque It was also observed in Figure 47 that power is more efficiently generated the more regenerative braking torque is applied, being the power very low for small amounts of torque and not following a linear correspondence. Figure 47: Power compared to torque detail 71

92 Looking at the energy consumed in Figure 48 and Figure 49 there can be seen that it increases, as it would be expected, since the vehicle is consuming every time torque is being applied. Also, the system developed will not balance all of those losses, and that is not its aim. Its aim is to recover as much energy that would otherwise be lost as possible. However, looking closely at the correspondent energy consumption when regen is active it can be concluded that every time negative torque is applied the total value of the energy consumed decreases. The amount of energy generated by the system is enough to be observed with no effort compared to the amount of energy consumed, as small peaks can be seen in Figure 48, which correspond with energy being recovered. Figure 48: Energy consumed Figure 49: Energy consumed detail 72

93 To compare the amount of energy saved by the use of this system, it was decided to plot the energy consumed when using regenerative braking against the energy consumed not taking into account when negative torque was being applied (energy was being recovered), as it can be seen in Figure 50. The conclusion to be extracted from this is that kilowatts per hour more were consumed in the case in which RB (Regenerative Braking) was not taken into account. This is 42.55% of the total energy consumption in the case where RB is ignored, and therefore a significant improvement in energy usage 1. Also, the total running time of this set of data is minutes. This is important because it means that watts per hour are being recovered in average every minute in this driving cycle. Figure 50: Comparison of energy with and without regen If that difference is plotted along with both consumptions, it can be seen in Figure 51 that, for this driving cycle, it means an increase of approximately (if the energy consumption is considered more or less constant) 546 seconds (9 minutes) in range if regenerative braking is still enabled. 1 It should also be taken into account the fact that in this driving cycle, more braking events than in a normal driving cycle were performed. Therefore, the energy recovered will not be a real indicator of what would happen in a city-driving cycle, but it can be used as an indicator that regen braking introduces a clear improvement in the vehicle s range. 73

94 Figure 51: Energy recovered with RB Figure 52: Distance traveled by the car in the test drive Since the main objective of this system is to increase range, as that will reduce emissions and create an eco-friendly car that takes advantage of all of the losses that occur in day-today driving, it can be said that this system is successful, as the increase of range achieved in this driving cycle in less than 14 minutes is of approximately 600 meters (again, considering that the next cycle driven was exactly the first part of the same cycle), a 71.6% of the total distance traveled. 74

95 Chapter VI: The Incremental System 6.1 Basis of the system and grounds to consider this system innovative As the objective of the competition that establishes the reason for this project is to develop a vehicle as ecological as possible, it was decided to start with the design of other system based on the same principles to achieve this same goal. This additional system will supply regenerative torque, with an approach based in the accelerator pedal position. After all of the research done on different aspects of what regen is and how it is developed and implemented (without obtaining any specific results on concrete approaches taken in any of the vehicles in which this type of system is used) it was decided to create an innovative and simpler procedure to achieve the same outcome. The incremental system is what will decide whether negative or positive torque needs to be applied by extrapolating the driver s intentions through the use of the accelerator pedal. This unprecedented system will work similarly to a floating zero regenerative system, but presents a different approach in what signals will be considered to determine the output. A floating zero consists on applying positive or negative torque depending on where the position of the accelerator pedal is, compared to the position needed to maintain the speed at which the car is going. This means that if the speed of the car is 60 km/h and the accelerator pedal position needed to maintain that speed is 50%, if the pedal position changes to 60%, more torque will be applied. However, if the pedal position drops to 40%, regenerative braking torque will be applied until the speed matches the speed that is maintained with a 40% displacement (considering the other conditions, such as torque, as set, which is not real, but provides with an easy example to understand how it works). Being this the basis for the new system, it was decided to simplify it. The reason for this is that a good and very precise mapping of the accelerator pedal position, speed and torque is needed. The current constraints under which the system is being developed do not make possible to do so. Also, one of the drawbacks of this type of system is that it does not consider the possible changes in torque necessary to maintain a certain speed depending on the slope of the road and other factors. This simplification consists on deciding whether to apply positive or negative torque depending only on the position of the accelerator pedal. The signal will be delayed and compared to the previous value of the signal to see if there has been an increment or decrement on its value (that is what it gets its name from), and depending on that positive or negative torque will be applied. 75

96 6.2 Conditioning of the primary signal The accelerator pedal position is going to be the basic signal for the system. The previous accelerator pedal position is going to be subtracted to the current one each sample time. This way, each sample time the increment or decrement of the accelerator pedal position is obtained. If the position is incrementing, the approach will be to apply positive torque and if it is decrementing, regenerative (negative) torque. However, this signal will not only be positive and negative, and the sign of the increment will not always reflect the intentions of the driver. Because the incremental signal can also be zero, and the driver intention will not necessarily be to apply zero torque but to maintain the level of power that the car provides, a finite state machine was designed. Its function is to decide when positive or negative torque is to be applied. Figure 53: Finite state machine that decides when positive or negative torque should be applied It has two states: up and down. The finite state machine will go to the state named up when the incremental signal is positive, and therefore, the amount of torque to be applied is positive. The state will be down when the incremental signal is negative, and the torque to be applied is regenerative. The states will not change unless the sign of the incremental signal does. This guarantees that when the incremental signal is zero but the driver wants to apply positive torque, the output of the system will be positive torque. In the same way, if the driver started decrementing the accelerator pedal position, the car will start decelerating, and if the driver maintains the accelerator position, the torque applied will still be negative until the driver decides to increment the accelerator position. To reflect the intentions of the driver, a dead zone was introduced in the system. Observing the accelerator pedal data in Figure 54 it was seen that there are variations in the position of the pedal when there is no intention of accelerating. This dead zone will be set so that these variations will not make the finite state machine switch states. 76

97 Figure 54: Accelerator Pedal Data The dead zone will have two positive consequences: it will make the system disregard the noise generated both by the sensor and the driver (that cannot be expected to maintain an exact foot position but will make slight changes without being its intention to change the output of the system), and it will reduce the set of values with an influence on the output. This last consequence is considered positive in this system because it is not desired that if the driver by mistake decrements the accelerator pedal position the car immediately falls into the regen mode. The driver should need to make a conscious decision or foot movement to enable regenerative braking, otherwise this system will not be accepted by consumers and therefore, it would be a waste of time to even think of developing or designing it. Therefore, the dead zone will not be symmetrical. The main function of the accelerator pedal is to command positive torque, so the value for which the system will switch to positive torque should be greater than the absolute value for which the system will enable regen. This value will also not be a set accelerator pedal position point, but a rate of variation or increment. The amount of torque commanded will be extracted from a set of curves, which consider accelerator pedal position (up to 95% and 95% increase because it was considered that there should be no difference between 95 and 100%) and speed as inputs to determine the torque. The curves that use pedal position as an input will be used to interpolate the torque commanded, both positive and negative, and the curve that is a function of speed will be used as a multiplier to scale the final value of regenerative braking according to what would be comfortable to the user. Both curves will be calculated at all times, and the finite state machine will decide which one is to be applied. 77

98 Torque (Nm) REGENERATIVE BRAKING SYSTEMS 6.3 Torque as a function of accelerator pedal position and speed Several driven test where made in the car to obtain sets of data. The ones that incorporated both torque command, speed and accelerator pedal position were analyzed in order to see how the variables correlated. Looking at the data, it was seen that the correlation between accelerator pedal and torque applied was very good, and speed is not strongly connected. Therefore, it was decided to eliminate speed as a variable of how much torque should be applied, and use this curve to create one of the inputs to the system Torque as a function of accelerator pedal position Pedal Position (%) Figure 55: Torque as a function of accelerator pedal position Data from the car Data applied Poly. (Data from the car) The equation of the trend line is y = x x. That means that the data recorded from the car tends to follow the equation Being x the accelerator pedal position in %. Equation 12 Torque (Nm) = x x The data that was finally applied for the simulations follows that trend line and was the following for positive torque: 78

99 Table 16: Pedal position and torque relation Pedal Torque Position REGENERATIVE BRAKING SYSTEMS This data will be an input for the system. The system will interpolate between those values to obtain the positive torque commanded in every moment, and that will be applied when the finite state machine is in the state that commands positive torque. This, however, will not be the curve finally used in the car, but an approximation to the torque command in the car to use in simulations. For negative torque, it was decided to create a curve based on the positive torque. However, this approach will not be sufficient, as the maximum regenerative torque should not be applied at the same speeds as with positive torque to ensure driver comfort. It was decided to create another curve to obtain values to be multiplied by the ones derived from this one, so that the amount of regenerative braking applied was also a function of speed. The combination of these curves will have the result of, considering fixed pedal position, decreasing the amount of negative torque applied as the car slows down when the car is at low speeds and when the car is at higher speeds, it will increment the amount of regenerative torque as the car slows down approaching the optimal speed (considered around 50 km/h), and decrease it as the speed tends to zero. The first approach to this was to create a non-aggressive curve, in which not a lot of regenerative torque was applied, and test it. The curve created and tested to obtain the amount of negative torque can be seen in Table

100 Table 17: Pedal position and Torque as an input Pedal Position (%) Negative Torque (N m) REGENERATIVE BRAKING SYSTEMS Figure 56 and Table 18 show the polynomial-like curve chosen for the multiplier. It can be seen that the maximum is 50 km/h, which is a speed low enough for the driver to feel comfortable with high amounts of regenerative braking, and that will shape the final amount of regenerative braking applied so that as the car slows down it decreases, creating a comfortable feeling. Table 18: Torque multiplier Torque Speed Multiplier

101 Multiplier (-) REGENERATIVE BRAKING SYSTEMS Multiplier for Regenerative Braking Speed (km/h) Figure 56: Curve for the regen multiplier It can be seen that the maximum amount of regenerative torque is around half of what the system can provide, what may seem slightly aggressive. However, this combined with the torque multiplier scales it down. It could be argued that with the torque multiplier it would still give the same theoretical value as the maximum of that curve is 1. However, that will not be applied as it holding a speed of 50 km/h with a pedal position of 95% or more, and then letting go completely is highly unlikely. 81

102 6.4 Development of the system The system was developed from two inputs, accelerator pedal position and speed. The output of the system will be torque commanded, either positive or negative. A block diagram of what the system looks like can be seen in Figure 57. In this diagram it can be seen how the inputs relate to the output and the different parts in which the system can be divided. Figure 57: Overview of the incremental system The translation of Figure 57 to Simulink, with data from the car as an input, can be seen in Figure 58. (It can also be seen in Annex II). Figure 58: Simulink diagram of the incremental approach 82