Medida de la temperatura superficial de combustión de carbonizado mediante pirometría con cámara digital para estudiar el efecto del CO2 en su oxi-conversión en reactores de lecho fluido.

TESIS DOCTORAL MEDIDA DE LA TEMPERATURA SUPERFICIAL DE COMBUSTIÓN DEL CARBONIZADO MEDIANTE PIROMETRÍA CON CÁMARA DIGITAL PARA ESTUDIAR EL EFECTO DEL CO2 EN SU OXICONVERSIÓN EN REACTORES DE LECHO FLUIDO POR PARTE DE JESÚS SALINERO GONZÁLEZ PARA OBTENER EL GRADO DE DOCTOR INTERNACIONAL DIRECTOR: DR. ALBERTO GOMEZ BAREA Departamento de Ingeniería Química y Ambiental ESCUELA SUPERIOR DE INGENIERÍA, UNIVERSIDAD DE SEVILLA Sevilla, España 2017 THESIS FOR DEGREE OF DOCTOR OF PHILOSOPHY MEASUREMENT OF CHAR TEMPERATURE IN A FLUIDIZED BED REACTOR USING PYROMETRY WITH A DIGITAL CAMERA: APPLICATION TO OXY-FUEL COMBUSTION BY JESÚS SALINERO GONZÁLEZ DIRECTOR: DR. ALBERTO GOMEZ BAREA Chemical and Environmental Engineering Department SCHOOL OF ENGINEERING, UNIVERSITY OF SEVILLE Seville, Spain 2017 A MIS PADRES i MEDIDA DE LA TEMPERATURA SUPERFICIAL DE COMBUSTIÓN DEL CARBONIZADO MEDIANTE PIROMETRÍA CON CÁMARA DIGITAL PARA ESTUDIAR EL EFECTO DEL CO2 EN SU OXI-CONVERSIÓN EN REACTORES DE LECHO FLUIDO JESÚS SALINERO GONZÁLEZ Departamento de Ingeniería Química y Ambiental ESCUELA SUPERIOR DE INGENIERÍA, UNIVERSIDAD DE SEVILLA Sevilla, España 2017 RESUMEN A pesar del trabajo en oxi-combustión de partículas de carbón/carbonizado en reactores de lecho fluido, el efecto de cambiar la atmósfera de O2/N2 a O2/CO2, con una elevada concentración de O2, sobre el proceso de conversión no está totalmente claro. Ello se debe (i) a que la mayoría de los trabajos experimentales utilizaron concentraciones de O2 menores al 21 %v/v; (ii) a que algunas de las técnicas de medida de temperaturas empleadas, tales como los termopares, podría haber alterado la conversión; y (iii) a que la mayoría de los estudios teóricos asumen que la partícula de carbón/carbonizado está únicamente en la fase emulsión, obviando las observaciones experimentales que la sitúan tanto en la fase burbuja como en la zona splash (chapoteo) del lecho durante parte del tiempo de fluidización. Por lo tanto, el objetivo principal de esta Tesis es aclarar la oxi-combustión de partículas de carbonizado (varios milímetros) en lecho fluido con una elevada concentración de O2 en CO2. Para ello, se diseñó y construyó un reactor de lecho fluido de geometría bidimensional (escala laboratorio, 18 x 50 x 1.8 cm) equipado con una ventana de cuarzo (20 x 20 cm) que permitió realizar medidas ópticas. Además, se utilizó un segundo reactor de lecho fluido cilíndrico (escala laboratorio, 7 cm diámetro interno x 50 cm). En todos los tests, tanto de combustión (O2/N2) como de oxi-combustión (O2/CO2), se alimentó al reactor una única partícula de carbonizado, la cual en algunos ensayos tenía un termopar insertado. Las partículas de carbonizado utilizadas se generaron a partir de partículas irregulares de madera de haya, y carbón bituminoso y subbituminoso, cuyos tamaños y pesos se fijaron en función del test donde fueran a ser utilizados. Como gente de fluidización se utilizó mezclas de O2 y N2 u O2 y CO2, así como aire, con una concentración de O2 del 11, 21, 30, 40, y 50 %v/v. En el primero de los estudios se desarrolla un método de medida de temperaturas superficiales del carbonizado, que se aplica a la combustión en un reactor de lecho fluido. La técnica se basa en la captura e interpretación de la radiación térmica contenida en las bandas espectrales visibles roja, verde, y azul que hace una cámara digital (Pirometría con Cámara Digital). Se muestra que el uso secuencial de la radiación contenida en los tres colores mencionados mejora la precisión de la temperatura medida respecto el uso tradicional simultáneo. Además, el método permite medir temperaturas del carbonizado menores a las del fondo (limitación en el uso simultáneo de dos colores). En el segundo estudio se analiza el efecto del termopar sobre la combustión del carbonizado en lecho fluido, encontrando que el termopar altera el movimiento del carbonizado a través del lecho, variando el tiempo de residencia en ii cada una de las fases del lecho. En concreto el termopar aumenta el tiempo de la partícula en la fase burbuja por la mayor resistencia del conjunto termopar-partícula al arrastre de las partículas del lecho. Este aumento en el tiempo del carbonizado en las burbujas aumenta la temperatura de combustión, reduce el tiempo de consumo y hace que el consumo superficial no sea homogéneo (la rotación de la partícula está impedida). En el tercer trabajo se analiza el efecto del movimiento del carbonizado a través de las diferentes fases del lecho sobre su temperatura, mostrándose que la temperatura superficial de combustión varía decenas de grados en tan solo centésimas de segundos. Este movimiento es experimentalmente caracterizado y se implementa en el modelo teórico de combustión de una partícula de carbonizado en un lecho fluido. Las predicciones ajustan bastante bien las oscilaciones de temperatura y tiempos de consumos. En el último estudio se analiza la influencia del CO2 y O2 en la oxi-combustión de las partículas de carbonizado, concluyéndose que la oxi-combustión del carbonizado en un lecho fluido está controlada por la trasferencia de O2 desde el seno de la fase fluida hacia el carbonizado, y que el consumo por gasificación no es despreciable. Se encuentra que el consumo por gasificación del carbón compensa la menor difusividad del O2 en CO2 reduciendo las diferencias entre los consumos aparentes medios del carbonizado (g/m2s) en O2/CO2 y en O2/N2 para temperaturas del carbonizado menores a los 925 °C, y haciendo el consumo en O2/CO2 mayor para temperatura del carbonizado mayores a 925 °C. Finalmente, las temperaturas registradas en el carbonizado durante conversión en O2/CO2 son menores a las registradas en la atmósfera equivalente (igual concentración de O2) de O2/N2, registrándose las mayores diferencias cuando las partículas están en la fase burbuja o zona splash (capoteo) del lecho. PALABRAS CLAVE: Lecho fluido, Oxi-combutión, Temperatura del carbonizado, Pirometría, Carbón iii MEASUREMENT OF CHAR TEMPERATURE IN A FLUIDIZED BED REACTOR USING PYROMETRY WITH A DIGITAL CAMERA: APPLICATION TO OXY-FUEL COMBUSTION JESÚS SALINERO GONZÁLEZ Chemical and Environmental Engineering Department SCHOOL OF ENGINEERING, UNIVERSITY OF SEVILLE Seville, Spain 2017 ABSTRACT Despite all work carried out on coal/char oxy-combustion in fluidized bed (FB), the effect of changing the atmosphere from O2/N2 to O2/CO2 at high O2 concentration is not well understood yet. Most of the experimental work has been carried out with O2 concentration lower than 21 %v/v and some of the techniques used to measure the fuel particle temperature, such as thermocouples, could interfere with the conversion. Theoretical studies assume that the coal/char particle is always in the emulsion phase, although experimental observations confirm that the particle is also in the bubbles and splash zone. These different char particle’s location in the bed could involve oscillations in the char particle’s temperature during the oxy-combustion process. On the other hand, controlling and measuring the char particle's temperature is one of the main challenges, but neither thermocouples nor pyrometry with an optical probe are accurate enough to be used for that purpose. Addressing those questions, a bi-dimensional fluidized bed reactor (lab scale, 18 x 50 x 1.8 cm) was designed and constructed with a quartz window in its bottom for optical measurements. Moreover, another cylindrical fluidized bed reactor (lab scale, 7 (i.d) x 50 cm) was occasionally used. All tests carried out were single char particle combustion (O2/N2) or oxy-combustion (O2/CO2), most of them the particle was freely fluidized (without an embedded thermocouple). Char particles from three different fuels were used: Beech wood, and Sub-bituminous and Bituminous coal, whose sizes and weights were set as a function of the tests where they were going to be used. The fluidization flow was either a pre-heated mixture of O2 and N2 or O2 and CO2, where the O2 concentration could be 11, 21, 30, 40 and 50 %v/v. In the first study, a method is presented to measure the temperature of the char particle's surface during conversion in a FB using a digital camera. The method applies one-color pyrometry sequentially for the three spectral bands red, green and blue (instead of the widespread used two-color pyrometry), changing from one band to another automatically as a function of the radiation intensity received by the digital camera’s sensor. In the second study, the pyrometric measurement is used to analyze the impact of using an embedded thermocouple into a char particle in a FB, finding that the thermocouple increases the combustion temperature and involve a non-homogeneous surface carbon consumption. In the third study, it is experimentally measured oscillations in the char’s temperature of tens degrees in hundredths of second resulting from its movement through the different bed's phases. That char's movement is experimental characterized and implemented in a theoretical model of a single char particle conversion in a FB, predicting quite well the experimental oscillations of the temperature and ix RESUMEN DE LA TESIS DOCTORAL EN CASTELLANO 1. ANTECEDENTES Y JUSTIFICACIÓN La necesaria reducción en las emisiones de CO2 es de tal magnitud, que no es posible conseguirla únicamente mediante la mejora en la eficiencia de las actuales plantas de energía, haciéndose necesario el desarrollo de soluciones complementarias basadas en tecnologías de captura y almacenamiento de CO2. La oxi-combustión es una las alternativas para la captura de CO2 en consideración, mediante la cual se obtiene una alta concentración de CO2 (del orden del 90% en volumen) en los gases de combustión, facilitando el secuestro y almacenamiento del CO2 contenido en la misma. Para ello se realiza la combustión con oxígeno puro y parte de los gases de salidas, ricos en CO2, recirculados para mantener el control de la temperatura en la caldera. De esta forma se consigue una razonable eficiencia neta en la planta, siendo la tecnología con mayores emisiones de CO2 evitadas. La 1ª generación de oxi-calderas consiste en el retrofitting de las plantas de producción de energía, donde la concentración de O2 en el interior de la caldera: 25–35 %v/v, se mantiene en condiciones similares a las de combustión ordinaria con aire (cambiando el N2 por CO2). Aunque esta 1ª generación supone una reducción de las emisiones de CO2, esta tecnología podría utilizarse con una mayor concentración de O2 en la caldera. Sin embargo, existen una serie de desafíos antes del diseño y construcción de este tipo de calderas de oxi-combustión, también llamadas de 2ª generación. Entre otras cuestiones, está la de aclarar la influencia del CO2 sobre el modo de conversión de la partícula de carbón, y el control de la temperatura en un reactor con menor volumen para intercambiadores de calor y donde la concentración de O2 es elevada. Las principales investigaciones se centran en calderas de carbón pulverizado al ser la caldera más común en la mayoría de las plantas eléctricas. Sin embargo, los reactores de lecho fluido pueden ser mucho más adecuados para la 2ª generación de oxi-calderas debido al control de la temperatura que brinda la circulación de sólidos y su capacidad de trabajar con combustibles de bajo rango, como son los residuos y la biomasa. Las investigaciones fundamentales de oxicombustión en reactores de lecho fluido se han centrado en condiciones de 1ª generación, y aunque recientemente se ha tratado de dilucidar el efecto del CO2 para condiciones de 2ª generación, la técnica de medida de temperatura empleada en estos trabajos se basaron en métodos intrusivos como termopares, de tal forma que el proceso de oxi-combustión podría verse afectado. Por otro lado, los principales estudios teóricos asumen que la partícula de x carbón/carbonizado reside únicamente en la fase emulsión, sin tener en cuenta observaciones experimentales que la sitúan en la fase burbuja y zona splash del lecho durante parte de la fluidización. En estas zonas, la combustión puede ser mucho más intensa (transferencia de masa) y las condiciones de transferencia de calor muy diferentes, por lo que las velocidades de consumo de las partículas de carbonizado y su temperatura pueden ser mucho mayores. Los modelos desarrollados hasta la fecha no han tenido presente explícitamente estas consideraciones, por lo que existen significativas discrepancias entre las predicciones y los resultados experimentales (típicamente superiores al 15 %). El control de la temperatura de las partículas de carbón/carbonizado es especialmente importante para evitar la aglomeración y defluidización del lecho. Esta importancia se incrementa al aumentar la concentración de O2 dado que se podrían alcanzar temperaturas muy superiores a la del lecho que produzcan la fusión de las cenizas. Es por ello que se han realizado algunos trabajos para aclarar la velocidad de combustión en condiciones de oxi-combustión con altas concentraciones de CO2 con distintas concentraciones de O2. Sin embargo, la mayoría de los trabajos se han centrado en condiciones de oxi-combustión de primera generación, con concentraciones de O2 típicamente por debajo de 35%V/V. Durante las investigaciones realizadas para esclarecer la velocidad de combustión en estas nuevas condiciones de reacción, una de las principales técnicas de medida consiste en introducir uno o varios termopares en el interior de la partícula de combustible, que posteriormente se arroja al lecho para medir la temperatura de la partícula con el tiempo, hasta que la partícula alcanza un tamaño aproximado de 3 mm o se desprende del termopar. Además, de la limitación en el tamaño mínimo cuya temperatura puede medirse y de los problemas de desprendimiento partícula-termopar, el movimiento de la partícula fluidizada con un termopar se ralentiza. Sin embargo, su uso se justifica asumiendo que la temperatura de una partícula fluidizada con un termopar es similar a la de una partícula libremente fluidizada, y que el proceso de conversión no se ve afectado. Se han empleado otras técnicas basadas en medidas ópticas como la pirometría, donde se solventan estos inconvenientes, pero si la captura de la radiación térmica emitida por la partícula se realiza por medio de una sonda óptica inmersa en el lecho, los resultados presentan una gran dispersión que los hace difícil de interpretar. Esta captura e interpretación de la radiación podría realizarse por medio de una cámara digital si el lecho es visualmente accesible desde el exterior del reactor. De esta forma, además de no afectar al movimiento de la partícula, los resultados serían mucho más claros y fáciles de interpretar. No obstante, esta técnica tan solo se ha utilizado recientemente para analizar el efecto del CO2 sobre le temperatura de llama en la devolatización de partículas de distintos tipos de carbón aisladas en una corriente de gas. Es decir, hasta la fecha no se disponen de medidas fiables y precisas de la oxi-combustión de partículas de carbonizado en reactores de lecho fluido y, a pesar de que la técnica de pirometría con cámara digital podría ser adecuada, ésta aún no se ha aplicado debido a las complejidades de visualizar adecuadamente las partículas de combustible durante el proceso de combustión lecho fluidizado. xi 2. OBJETIVOS El objetivo de esta tesis es aclarar el proceso de oxi-combustión de partículas de carbonizado en lecho fluido, y para ello, desarrollar y poner a punto la técnica de medida de la temperatura superficial de carbonizado con pirometría con cámara digital en un lecho fluidizado. Para ello, se ha diseñado y construido una instalación experimental basada en un lecho fluidizado bidimensional donde se ha puesto a punto la técnica anterior, realizándose experimentos con distintas partículas de carbonizado de carbón y biomasa en condiciones de oxi-combustión. El trabajo experimental se ha completado con el desarrollo de un modelo teórico que tiene en cuenta de forma explícita el movimiento del carbonizado a través de las diferentes fases del lecho (movimiento que también se ha caracterizado experimentalmente). Adicionalmente, la investigación ha permitido aclarar el efecto del termopar sobre la medida de la temperatura de la partícula, así como mostrar la influencia que el movimiento de la partícula de char a través del lecho tiene sobre la temperatura superficial del char, y en particular, sobre la frecuencia y amplitud de las fluctuaciones de ésta durante la combustión en lecho fluidizado. 3. EQUIPOS Y METODOLOGÍA EXPERIMENTAL La medida de la temperatura del carbonizado durante los ensayos (tanto de oxi-combustión (O2/CO2) como de combustión (O2/N2) se basa en la captura e interpretación de la radiación térmica contenida en las bandas espectrales visibles roja, verde y azul que realiza la cámara digital. El software proporciona tres valores para cada uno de los píxeles de la imagen, conocidos como números digitales (DN), que cuantifican la radiación que recibieron en cada una de las bandas espectrales mencionadas (ver Fig. 3.1, Capítulo 3). Si la radiación visible que alcanza al sensor se debe a la temperatura de los objetos presentes en el campo de visión de la cámara (Ley de Planck), es posible utilizar la información contenida en esos números digitales junto con el modelo matemático de la radiación que alcanza el sensor para determinar la temperatura de dichos objetos. La temperatura pirométrica de las partículas de carbón/carbonizado o de llama en reactores de lecho fluido se miden en la literatura mediante la relación simultanea de la información contenida en dos bandas espectrales o colores (pirometría en dos colores, P2C). De esta forma, se determina la temperatura de la partícula de combustible si esta es mayor a la del fondo (background) y su emisividad no cambia con las longitudes de ondas de los colores utilizados (superficie gris). Sin embargo, si se conoce la emisividad de la partícula en esos colores es posible utilizar la información contenida en cada uno de los colores de forma secuencial y se pueden medir temperaturas del carbonizado menores a las del fondo. Por otro lado, la relación entre los pixeles ocupados por el diámetro de la partícula en los primeros instantes del ensayo y su medida en mm permite estimar el tamaño del carbonizado durante toda la conversión. En este trabajo se utilizaron partículas de carbonizado de madera de haya y carbón subbituminoso y bituminoso (análisis aproximado en la Tabla 3.1, Capítulo 3), cuya forma y masa dependen del tipo de estudio y objetivo. En cada uno de los ensayos se realizó la oxi-combustión (O2/CO2) o combustión (O2/N2) individual de una de estas partículas. Para desarrollar este trabajo se diseñó y construyó la planta FBC 2D (ver Fig. 3.2, capítulo 3), compuesta por un reactor de lecho fluido de geometría bi-dimensional (18 x 1.8 x 50 cm) equipado con una ventana xii de cuarzo en el bottom. Este se situó en el centro de un horno eléctrico (11 kW, 50 x 30 x 70 cm) rodeado por una superficie negra que homogeniza la radiación procedente de las resistencias del horno. El horno eléctrico también está equipado con una ventana de cuarzo desde la cual se realizaron las grabaciones con una cámara digital comercial (JVC Everio HD, 2 Mpx). También se utilizó una planta de lecho fluidizado de laboratorio (FBC 3D), con un reactor de lecho fluido cilíndrico (51 mm de diámetro interno en la parte inferior) situado en el centro de un horno eléctrico circular (9 kW, 30 cm de diámetro x 70 cm altura). En ambos reactores se utilizó arena de sílice como material del lecho (di: 0.32-0.5 mm, umf: 0.18 m/s), cuya temperatura, así como la composición y velocidad del flujo de fluidización dependieron del estudio dentro que se llevase a cabo (ver descripción de los equipos en la Tabla. 3.1). 4. ESTUDIO TEÓRICO (MODELO) DE LA COMBUSTIÓN DE UNA PARTÍCULA DE CARBONIZADO EN UN LECHO FLUIDO Las predicciones de temperatura y conversión realizadas en este trabajo tienen en cuenta el movimiento del carbonizado a través de las diferentes fases del lecho (emulsión, burbuja y splash). Para ello se caracteriza experimentalmente este movimiento cuantificando el tiempo que la partícula de carbonizado está en cada una de las fases en función de su tamaño. De esta manera se añade una variable adicional a la simulación que indica la posición del carbonizado en cada instante. Básicamente, lo que se realiza es la descripción del movimiento aleatorio de la partícula durante su fluidización como un movimiento cíclico que se repite durante la conversión, cuyo periodo y tiempo en cada una de las fases del lecho varía con el tamaño del carbonizado (ver Fig. 2.2, Capítulo 2). Dado que el tamaño de la partícula varía durante la simulación, la caracterización del ciclo también lo hace. Este enfoque, que podemos denominar dinámico, permite simular el cambio en la trasferencia de masa y calor desde/hacia la partícula de carbonizado en función de su posición en el lecho. Este enfoque contrasta con el “tradicional” o “estático” donde se asume que la partícula reside siempre en la fase emulsión. Dada las características experimentales bajo las cuales se obtienen las correlaciones de trasferencia de masa y calor de la literatura, se utilizan correlaciones teóricas (ver Ecs. 2.22-2.27, Capítulo 2). 5. METODOLOGÍA ESPECÍFICA Y CONDICIONES DE OPERACIÓN EN LOS ENSAYOS REALIZADOS En el presente trabajo se realizaron diferentes estudios para investigar aspectos de la combustión de partículas de carbonizado en lecho fluido. Todos ellos utilizan la técnica pirométrica con cámara digital pero con propósitos diferentes. A continuación se resumen las condiciones de operación y ajuste de cada uno de los cuatro estudios principales llevados a cabo. 5.1. CALIBRACIÓN Y VALIDACIÓN DE LA MEDIDA DE TEMPERATURA MEDIANTE PIROMETRÍA CON CÁMARA DIGITAL EN LECHO FLUIDO Se realizó la combustión (O2/N2) de partículas de carbonizado de madera de haya con forma de paralelepípedo en el centro del horno eléctrico de la planta FBC 2D. En estos ensayos la partícula tenía un termopar alojado en una ranura superficial que permite medir la temperatura de una determinada región de la superficie del carbonizado durante parte del tiempo de combustión. De esta forma fue posible relacionar los números digitales que la cámara asociaba a xiii una radiación/temperatura conocida (ver Fig. 3.3, Capítulo 3 o Fig. 4.1 capítulo 4), necesarios para la calibración de la cámara digital. Una vez calibrada la técnica de medida se compararon sus resultados (temperaturas pirométricas) haciendo uso de la información contenida en los diferentes colores de forma simultanea (P2C) y de forma secuencial (P1C) con la medida de la temperatura superficial mediante un termopar. 5.2. INFLUENCIA DEL TERMOPAR SOBRE LA TEMPERATURA DE COMBUSTIÓN Y CONSUMO SUPERFICIAL DEL CARBONIZADO EN LECHO FLUIDO En este estudio se realiza y graba la combustión (O2/N2) de partículas individuales esféricas (~ 10 mm) de carbonizado de madera de Haya, carbón Sub-Bituminoso y Bituminoso en el reactor de lecho fluido FBC 2D fluidizadas libremente (sin termopar), y con un termopar insertado en su centro. Del análisis (clasificación) de las imágenes de estos ensayos (ver fig. 3.6, Capítulo 3 y Fig. 4.9, Capítulo 4), se cuantifica el tiempo de la partícula en cada una de las fases del lecho (emulsión, burbuja y zona splash), y como ese tiempo varía por efecto del termopar. La medida de la temperatura por pirometría con cámara digital, permitió comparar las temperaturas superficiales registradas durante la conversión, y mediante la extracción de las partículas de carbonizado del reactor FBC 2D y del reactor FBC 3D a distintos tiempos de combustión se analizó su efecto sobre el consumo superficial del carbonizado. 5.3. OSCILACIÓN DE LA TEMPERATURA DEL CARBONIZADO DURANTE SU COMBUSTIÓN EN LECHO FLUIDO En este estudio se realiza la combustión (O2/N2) de partículas individuales esféricas (~ 8 mm) de carbonizado de madera de haya y carbón sub-bituminoso en el reactor de lecho fluido FBC 2D. La temperatura del lecho se fijó a 800 °C, y la concentración de O2 fue del 11 o 21 %v/v. La medida de la temperatura de las partículas de carbonizado en secuencias de imágenes donde el carbonizado cambia de posición en el lecho: fase emulsión ↔ fase burbuja ↔ zona splash, permitió cuantificar la amplitud de las oscilaciones de las temperaturas y el tiempo en el cual se producen. Además, se implementó la caracterización experimental del movimiento del carbonizado a través de las diferentes fases del lecho en el estudio teórico de conversión del carbonizado en lecho fluido, mostrando su capacidad predictiva de las variaciones de temperatura. 5.4. INFLUENCIA DEL CO2 EN LA OXI-CONVERSIÓN DEL CARBONIZADO EN LECHO FLUIDO En este estudio se realiza la oxi-combustión (O2/CO2) y combustión (O2/N2) de partículas individuales esféricas de carbonizado de madera de haya y carbón sub-bituminosos en el reactor de lecho fluido FBC 2D. El peso de las partículas de carbonizado se fijó de forma que las partículas tuvieran un tamaño inicial aproximado de 8 y 13 mm. La temperatura del lecho de arena fue de 800 ó 850 °C, y la concentración de O2 entre 11 y 50 %v/v en N2 o en CO2. Todos los ensayos fueron grabados lo que permitió medir la temperatura superficial del carbonizado por pirometría con cámara digital (P1C). Además, se midió la conversión mediante el análisis del cambio de tamaño del carbonizado durante todo el proceso de conversión. Se compararon las predicciones xiv del modelo con los resultados experimentales de temperatura, tiempos de consumo y tasa de consumo aparente (ver Eq. 3.7, Capítulo 3) de este estudio y de otros estudios de la literatura. 6. DISCUSIÓN DE RESULTADOS Una vez calibrada la cámara digital, esto es, conocido como los parámetros ópticos y electrónicos de la configuración experimental se relacionan con los números digitales generados por el software, se midió la temperatura superficial del carbonizado de madera de haya mediante el uso simultaneo (pirometría en dos colores, P2C) y secuencial (pirometría en un color, P1C) de los números digitales (DN), y mediante un termopar alojado en su superficie (ver Fig. 4.5, Capítulo 4). El resultado muestra que el uso secuencial propuesto mejora la precisión de la temperatura medida, y aumenta el rango de temperaturas que puede medirse, incluyendo temperaturas del carbonizado inferiores a las del fondo. Aunque el principal inconveniente de esta medida (P1C) es que la emisividad del carbonizado debe conocerse, la incertidumbre en el resultado (temperatura medida) cuando la emisividad varía entre los principales valores de la literatura, es inferior al 1 % si la temperatura del carbonizado es mayor a la del fondo (ver Fig. 4.8b, Capítulo 4). No obstante, si la temperatura del carbonizado es inferior a la del fondo es importante realizar una buena estimación de la emisividad (ver Fig. 4.8a, Capítulo 4). La clasificación de las imágenes de los ensayos de combustión de las partículas de carbonizado fluidizadas con y sin termopar mostró que su uso hunde la partícula en el lecho, aumentando el tiempo que ésta está en la fase burbuja (ver Fig. 4.11, Capitulo 4). El conjunto carbonizadotermopar resiste el arrastre de las partículas del lecho (que tienden a mantener al carbonizado en la fase emulsión) y mantiene al carbonizado enfrentado al flujo ascendente de las burbujas (ver Fig. 4.13, Capítulo 4). El aumento en el tiempo dentro de la fase burbuja (superior al 40, 60, y 100 % para un termopar de 0.25, 0.5, y 0.75 mm de vaina) lleva consigo el aumento de la temperatura de combustión durante la fluidización, reduciendo el tiempo de consumo en más del 5 % respecto del registrado para una partícula de carbonizado fluidizada libremente (ver Fig. 4.14, Capítulo 4). Además, dado que la rotación libre de la partícula está impedida por el termopar, es la misma superficie de la partícula la que enfrenta el flujo de gas que asciende desde el plato distribuidor (ver Fig. 4.16, Capítulo 4). Por lo tanto, esta región es la que sufre las mayores temperaturas y consumos superficiales durante toda la conversión y produce un consumo superficial no homogéneo del carbonizado (ver Figs. 4.16 y 4.17, Capítulo 4). La interpretación de las imágenes de los ensayos de combustión por pirometría detectó variaciones en la temperatura superficial del carbonizado de decenas de grado en tan solo centésimas de segundo (ver Fig. 4.19, Capítulo 4). Estas oscilaciones de temperatura que se producen durante toda la conversión, son consecuencia del cambio de posición del carbonizado en el lecho, y su amplitud aumenta con la concentración de oxígeno y disminuye con el tamaño de partícula (ver Fig. 4.20, Capítulo 4). Los resultados indican que la temperatura de la partícula de carbonizado es más estable cuando ésta reside en la fase emulsión que cuando está en la fase burbuja o en la zona de chapoteo (splash). Por otro lado, la implementación del movimiento del carbonizado a través del lecho (caracterizado experimentalmente, ver Fig. 4.12 en Capítulo 4) en el modelo, mostró la gran capacidad predictiva del modelo para describir las oscilaciones de temperaturas observadas experimentalmente. xv La medida experimental de la temperatura y consumo de una partícula de carbonizado durante su oxi-combustión (O2/CO2) y combustión convencional (O2/N2) en el reactor FBC 2D ha mostrado que la transferencia de O2 desde el seno de la fase fluida a la partícula controla el proceso global de conversión, y que el consumo del carbonizado por gasificación no puede despreciarse en ningún caso. Este consumo adicional por gasificación compensa la menor difusividad del O2 en CO2 respecto la del O2 en N2 reduciendo las diferencias en el consumo en ambas atmósferas (O2/N2 y O2/CO2) cuando la temperatura del carbonizado es inferior a los 925 °C, o haciendo que el consumo en la atmósfera O2/CO2 sea mayor si la temperatura del carbonizado es mayor a los 925 °C. Además, cuando el carbonizado reside en la fase emulsión la temperatura es similar en ambas atmósferas, registrándose las mayores diferencias cuando la partícula está en la fase burbuja o zona splash. Por lo tanto, las principales diferencias en la conversión se deben al tiempo durante el cual la partícula está fuera de la fase emulsión, lo que muestra la importancia de considerar de forma explícita el movimiento del carbonizado a través de las fases del lecho en el estudio teórico. La comparación de los resultados experimentales y de otros trabajos de la literatura con las predicciones del estudio teórico dinámico planteado muestra una incertidumbre menor al 15 %, mejorando algunos de los resultados que pueden encontrarse en la literatura. 7. CONCLUSIONES El trabajo realizado y discutido puede agruparse en cuatro grandes estudios que aun compartiendo un mismo objetivo general, que es aclarar el proceso de oxi-combustión de partículas de carbonizado en lecho fluido, tuvieron objetivos particulares. A continuación se resumen las conclusiones y contribuciones más destacables de cada uno de los cuatro estudios (publicados en 4 artículos y 2 aportaciones/proceedings a congresos internacionales, ver § PAPERS). El primero de los estudios consiste en el desarrollo de un método de medida de temperaturas superficiales del carbonizado durante su oxi-combustión y/o combustión en un reactor de lecho fluido, basado en la captura e interpretación de la radiación térmica en las tres bandas espectrales visibles roja, verde, y azul que hace una cámara digital (pirometría con cámara digital), mostrándose que: I. El uso secuencial de la radiación contenida en los tres colores mencionados mejora la precisión de la temperatura medida respecto el uso tradicional simultáneo (uso simultáneo de dos colores). II. Se pueden medir temperaturas del carbonizado menores a las del fondo (cosa que no se puede hacer con el uso de varios colores simultáneamente), siempre que se conozca la emisividad del carbonizado. En el segundo estudio analiza el efecto del termopar sobre la combustión de una partícula en lecho fluido: temperatura y consumos superficiales del carbonizado, mostrándose que: III. El termopar altera el movimiento del carbonizado a través del lecho, variando el tiempo de residencia en cada una de las fases. El hundimiento de la partícula en el lecho causado por el termopar incrementa su tiempo en la fase burbuja, dada la mayor xvi resistencia que el conjunto termopar-partícula ofrece al arrastre de las partículas del lecho. IV. El aumento del tiempo en la burbuja incrementa la temperatura superficial de combustión y reduce el tiempo de consumo. V. El consumo superficial no es homogéneo, dado que la rotación de la partícula está impedida. El tercer trabajo aborda el efecto del movimiento del carbonizado a través de las diferentes fases del lecho sobre su temperatura de combustión, e implementa su caracterización experimental en el estudio teórico VI. Se muestra que la temperatura superficial de combustión varía decenas de grados en tan solo centésimas de segundos como consecuencia de las diferentes posiciones de la partícula en el lecho. Por otro lado, las predicciones del modelo permiten describir estas oscilaciones durante la conversión, mejorando los modelos existentes. El último estudio analiza la influencia del CO2 y O2 en la oxi-combustión de las partículas de carbonizado, concluyendo que: VII. La oxi-combustión del carbonizado en un lecho fluido está controlada por la trasferencia de O2 desde el seno de la fase fluida hacia el carbonizado, y que el consumo por gasificación no es despreciable. VIII. El consumo adicional por gasificación del carbón compensa la menor difusividad del O2 en CO2 reduciendo las diferencias entre los consumos aparentes del carbonizado (g/m2s) en O2/CO2 y en O2/N2 para temperaturas del carbonizado menores a los 925 °C, y hace mayor el consumo en O2/CO2 que en O2/N2 para temperatura del carbonizado mayores a 925 °C. IX. Las temperaturas registradas en el carbonizado durante conversión en O2/CO2 son menores a las registradas en la atmósfera equivalente (igual concentración de O2) de O2/N2, registrándose las mayores diferencias cuando las partículas están en la fase burbuja o zona splash del lecho. MEASUREMENT OF CHAR TEMPERATURE IN A FLUIDIZED BED REACTOR USING PYROMETRY WITH A DIGITAL CAMERA: APPLICATION TO OXY-FUEL COMBUSTION CHAPTER 1 6 1) It is developed and demonstrated the application of pyrometry with a digital camera to measure the surface temperature of char particles during combustion in a FB, discussing the capability of the technique to detect further details such as surface temperature gradients or particle’s shrinkage. 2) Using this technique a comprehensive analysis of the effect of the thermocouples on the char particle’s conversion and the experimental characterization of the char particle's movement through the bed are carried out. 3) Moreover, the amplitude of the temperature oscillations caused by this movement is analyzed both experimentally and theoretically by the implementation of the mentioned experimental characterization of the char particle's positions in the bed during fluidization in the conversion model proposed by Biggs et al. [43]. 4) The temperatures and consumptions of char particles during their single combustion (O2/N2) and oxy-combustion (O2/CO2) in a FB, working at different conditions (O2 concentration, char particle’s size, bed temperature, and types of chars), are measured and analyzed. Moreover, these experimental measurements are compared with our own theoretical predictions and with those from published work in literature, concluding favorably about the ability of the theoretical treatment developed in this thesis with respect to existing models published up to date. 7 CHAPTER 2. THEORETICAL PREDICTIONS OF A SINGLE CHAR PARTICLE OXY-COMBUSTION IN A FLUIDIZED BED The theoretical approach for the conversion of the char particle in a fluidized bed (FB) under oxy-fuel conditions taking explicitly into account the movement of the char particle through the bed is presented. The experimental characterization of this movement between different phases of a FB and its effect on the transfer phenomena are established and discussed. 2.1. THEORETICAL APPROACH Single char/coal particle conversion in a FB has been widely studied over the past decades, where have been proposed different theoretical approaches attending to the way in which the carbon conversion proceeds inside the char particle [70]. Among of these theoretical models and it is justified below (see § 2.3), in this work is assumed that the char consumption (by carbon oxidation and gasification) proceed in a very narrow char layer that separates the unreacted char particle from the bulk phases of the bed since the ash generated is detached (shrinking unreactive particle model). Although a single char particle conversion in a FB is carried out, involving the same concentration of the different species (O2, N2, CO2, and CO) in all bed phases: emulsion phase, bubble phase, and splash zone; the heat and mass transfer from/to the char particle varies depending on char’s place/location in the bed. Therefore, taking into account the char's place in the bed for each time, the mass transfer is solved providing the diffusion of the different species (fluxes) from/to the char. These fluxes are then used to calculate the char consumption and temperature using the heat transfer conditions in the corresponding bed phase. It should be noted that before solving the theoretical approach, the characterization of the char particle's movement through the emulsion phase, bubble phase, and splash zone has to be set (see § 2.4). 2.2. MOLAR FLOWS The char particle consumption in an O2/N2 or O2/CO2 atmosphere is given by oxidation C(s)+2+ξ 2 (1+ξ) O2(g)→ ξ (1+ξ)CO(g)+1 (1+ξ)CO2(g) (2.1) and by gasification C(s) + CO2(g)→ 2CO(g) (2.2) CHAPTER 2 8 where ξ=Aco/co2 exp(−Eco/co2 RgTc) (2.3) is the CO/CO2 molar ratio produced by the carbon oxidation [71]. As it is shown in Fig. 2.1, the char consumption occurs in a multicomponent gas atmosphere where equimolar counterdiffusion is likely not established [72-74]. Although this phenomenon is often neglected during calculation of single char particle combustion [26, 36-38], errors is introduced in the estimation of mass transfer [72-74] (errors up to 10 % for a 21 %v O2 concentration in N2 have been found [74]). To consider the multicomponent diffusion, the Stefan-Maxwell equations are used: dxi dr =∑1 CDij n j=1 j≠i (xjNi−xiNj) (2.4) relating the molar fluxes of species i (Ni) from/to the char particle’s surface and the molar fractions (xk) at any point of the gas film surrounding the particle. The boundary conditions are: xi(r=rc)=xi,c xi(r=rc+δ)=xi,bed (2.5) where xi,c and xi,bed are the molar concentration at the char particle surface and that in the bulk gas and Dij=εmf τdij (2.6) the effective mass diffusivity, being dij the gas binary diffusivity (of the i-th compound through j) [75] and τ tortuosity (Dij = dij when the char particle is in the bubble phase and splash zone). Fig. 2. 1. Sketched of the Molar fractions of the gaseous compounds around a char particle during its conversion in an O2/N2 (a) or O2/CO2 (b) atmosphere Taking the mass-transfer resistance into account, the thickness of the gas film around the particle (𝛿) has been estimated by [76] δ= rc θShcdeq Sho−1 (2.7) where Sho is the Sherwood number during equimolar counter-diffusion without no massconvective flux (Sho = 2εmf if the particle is in the emulsion phase and Sho = 2 if the particle is in the bubble phase or splash zone). Shcdeq is the Sherwood number when there are equimolar THEORETICAL PREDICTIONS OF A SINGLE CHAR PARTICLE OXY-COMBUSTION IN A FB 9 counter-diffusion and mass-convective fluxes. When there is no equimolar counter-diffusion, a correction factor θ have been estimated elsewhere [76] by θ=ξ+2 ξ1 xi,bedln (1 + ξ ξ+2xi,bed) (2.8) assuming that the O2 concentration at the char/coal surface [76] is negligible and equal diffusivities of CO2, CO and O2 through N2 [72, 74] (note in Eq. 2.7 θShcdeq instead of Shcdeq). Applying the conservation of the i-th compound through the gas-film (considering no chemical reaction) r2Ni=rc2Ni,c (2.9) where Ni,c is the molar flux at the char particle’s surface, and taking into account that the oxygen is consumed at the external char surface at a rate given by a first-order kinetic combustion reaction rO2,cb=− 2+ξ 2 (1+ξ) KcbCxO2,c (2.10) and by gasification reaction rCO2,gf=−KgfCxCO2,c (2.11) where the kinetic coefficients (based on external particle surface) are given by kcb=AcbTcexp(−Ecb TcRg) (2.12) for combustion and kgf=AgfTcexp(−Egf TcRg) (2.13) for gasification, the flux of the various species (CO, CO2 an O2) at the char particle’s surface can be calculated by NCO,c=−2ξ 2+ξrO2,cb+2rCO2,gf (2.14) NCO2,c=− 2 2+ξrO2,cb−rCO2,gf (2.15) NO2,c=rO2,cb (2.16) It should be noted that the flux of N2 is null at any point of the gas-film surrounding the particle since there is no consumption and production. 2.3. SHRINKING PARTICLE MODEL As it will be shown in the experimental results (see § 4.2), for the various char tested in this work, the ash detaches rapidly during conversion in fluidized bed and the char particle’s density remains approximately constant. Assuming that the temperature gradient inside the particle is small [30], an isothermal shrinking particle model can be applied to a spherical particle to predict char consumption with time. This theoretical model has been successfully applied for different biomasses [77, 78] and coals [53]. The change of char particle’s size and temperature with time is predicted by solving the two-coupled equations CHAPTER 2 10 drc dt = − 1 ρmc(rCO2,gf−2 (ξ+1) 2+ξ rO2,cb) (2.17) dTc dt =3 ρm,cCP1 rc(((−∆Hgf)rCO2,gf−(−∆Hcb)rO2,c) − (Nucdeq j Kdg2rc(Tc−T∞)+ σωc(Tc4−T∞4))) (2.18) with the initial condition at t = 0 rc=rc,o Tc=298 K (2.19) where ∆Hcb=( 1 (1+ξ)∆HCO2+ξ (1+ξ)∆HCO)−(∆HC+2+ξ 2 (1+ξ) ∆HO2) (2.20) is the combustion heat (exothermic) and ∆Hgf=(2∆HCO)−(∆HC+∆HCO2) (2.21) the gasification one (endothermic), and Nujcdeq is the Nusselt number when the char particle is in the emulsion phase (j = ePh), bubble phase (j = bPh), or in the splash zone (j = sPh) as it is explained below (see § 2.4). It could be argued that the gasification of the char particle with CO2, which could occur during oxy-fuel combustion with high CO2 concentration and char particle's temperature higher than 900 °C, is controlled by intrinsic kinetic [24] and takes places throughout the entire particle as the particle is porous and the reaction relatively slow. However, in the treatment given here, char gasification is assumed to effectively occur in a narrow layer near the external surface of the char particle and therefore an externally surface-based kinetic equation rate is used (see Eq. 2.11, 2.13, and Table 2.1). Consistently, the intrinsic kinetic parameters commonly obtained in a Thermo gravimetric Analyzer [18] are extrapolated to apparent ones (externally-based and taking diffusion effects into account) [77]. As a result, apparent kinetic used elsewhere [80-82] for different types of fuel particles has been taken for gasification reaction. It is known that the gasification reaction weakens the internal structure of the char particle by pore enlargement [24], and this could change the conversion rate and density of particles especially at high conversion. However, here is assumed that the char particle density remains approximately constant since the conversion conditions handled involve that the gasification contribution is much lower than oxidation one. Therefore, the possible char particle's density change is assumed to be small supported by our experimental observations. In summary the approach followed in the present work by implementing the gasification as a heterogeneous reaction taking place in parallel with oxidation in a thin shell layer close to external particle, is justified. 2.4. CHAR PARTICLE’S MOVEMENT THROUGH A FLUIDIZED BED As it is pointed out above (see § 1.1), there are experimental observations indicating that the char particle is in the emulsion phase a large part of the conversion time, but that it is also in the bubble phase and splash zone [32-34, 82]. This period is a fraction of the fluidization time not negligible (it could be higher than 20 % [32, 33]), during which the heat and mass transfer THEORETICAL PREDICTIONS OF A SINGLE CHAR PARTICLE OXY-COMBUSTION IN A FB 11 from/to the char particle vary, and possibly other phenomena such as the CO oxidation close to the char particle. This idea is sketched in Fig. 2.2a by “zooms” in the different char particle’s locations in the bed. As it has been theoretically proposed elsewhere [43], the “random” char particle’s movement through the bed could be implemented into the conversion model treating this “random” movement as a cyclic one (see Fig. 2.2a). It links the downward motion of the char particle through the emulsion phase until a certain depth from the bed surface, followed by its upward motion from there to the bed surface where it resides some time before starting a new cycle. During the upward motion of the char particle, it stays most of the time in the emulsion phase with intermittent penetration into the bubbles. It should be pointed out that the char particle in the emulsion phase is assumed to be stopped or moved downward because that phase is assumed to be under minimum fluidization condition. The bubbles move sand particles upward and it is assumed that the char also moves upward when it is in the bubble phase or close to it. Fig. 2.2. Sketch (a) and temporal characterization (b) of the char particle’s movement through the emulsion phase (sPh), bubble phase (bPh) and splash zone (sPh) and the heat and mass transfer and CO/CO2 ratio in each place As it is shown in Fig. 2.2b, the characterization of this cycle by fluid-dynamic parameters required the upward and downward char particle’s velocities through the bed (ur and ud), the depth reached by the char (hp), the bubble velocity and diameter (ub and db) among others [33]. However, if the depth reached by the char and the fractions of the time (𝜃i) during which it is in the emulsion phase (i = ePh), bubble phase (i = bPh), and splash zone (i = sPh) can be experimentally established (see § 3.2.2), the cycle’s characterization can be calculated by estimating the bubble velocity and the upward char velocity (see Appendix I). It should be noted that a new time-dependent variable becomes involved in the presented theoretical description: the location of the char particle in the bed. This approach, which can be called “dynamic”, CHAPTER 2 12 contrasts from the classical approach used in the literature where the char particle is assumed to remain always in the emulsion phase (see § 1.1), here called “static”. Therefore, it is needed the estimates of the mass and heat transfers for each one of the char particle’s positions in the bed. 2.5. HEAT AND MASS TRANSFERS AND CO/CO2 RATIO Mass transfer estimates in a fluidized bed (for experimental conditions handled in this work, see Table 3.1) applied techniques based on the consumption of mass or drying of an active particle, such as naphthalene [83, 84], wet porous particle [85], or coal [51], quantifying the change in the particle mass and/or in the concentration of gaseous compounds, such as naphthalene, water vapor, or CO and CO2. Moreover, CO oxidation in a gas fluidizing a sand bed with few Pt catalyst spheres immersed, has also been used to estimate mass-transfer coefficients [86]. Heat transfer has been estimated by an energy balance during the cooling of warm metal particles [55, 84, 87], or during water evaporation from a wet porous particle [85]. In all these studies, the particles were freely fluidized, or with an embedded thermocouple, having likely stayed in the emulsion phase most of the time, but also in the bubble/splash regions. Therefore, all these experimental heat and mass transfer estimations implicitly lumped the effect of the active particle’s movement through the bed, and none of them can be formally used for just one of the char particle's positions in the bed. Consequently, here it is used theoretical correlations for the mass and heat transfer estimates [39-41, 43] when the char particle is in the emulsion phase (ePh) Shcdeq ePh =2εmf τ2+0.3465(ReePhCd,ePh)0.33(τ εmf ⁄ )0.66((1+ReePhSc)0.33−1) (2.22) Nucdeq ePh =2kd,ePh kd,g +0.3465(ReePhCd,ePh)0.33(τ εmf ⁄ )0.66((1+ReePhPr)0.33−1)+ (dc kd,g) ( di ψKd,g+0.5√πtePh,T Kd,ePhρmfCpi)⁄ (2.23) bubble phase (bPh) Shcdeq bPh =2+0.3465(RebPhCd,bPh)0.33((1+RebPhSc)0.33−1) (2.24) Nucdeq bPh =2+0.3465(RebPhCd,bPh)0.33((1+ReePhPr)0.33−1) (2.25) and splash zone (sPh) Shcdeq sPh =2+0.69(ResPh)0.5Sc0.33 (2.26) Nucdeq sPh =2+0.69(ResPh)0.5Pr0.33 (2.27) The detailed description of these correlations is in Appendix II. It could be said that the dependence of mass transfer on the minimum fluidization velocity and not on the fluidization velocity, claimed in several investigations [83, 84, 86, 87], is related to the fact that the char particle is mainly in the emulsion phase [83, 86]. However, using one of the experimental correlation could involve taking into account twice the effect of the char particle in the bubble/splash phase on the mass transfer, since that the proposed model account for separately the mass transfer when the particle is in each of these phases. Moreover, other investigations found influence of the fluidization velocity on mass transport rates [39, 85]. This THEORETICAL PREDICTIONS OF A SINGLE CHAR PARTICLE OXY-COMBUSTION IN A FB 13 contradiction could be understood if the effect of the fluidization velocity on the position of the coal/char particle in the bubbling bed changes slightly when the sand bed is at bubbling conditions. In such cases, an increasing fluidization velocity enhances the mass transfer rate because the convective flow in the splash zone and bubble phase is increased (emulsion phase is at minimum fluidization conditions). If the time in these positions for the experimental conditions used in the mentioned works (da/di: 1 – 200; Ar: 10 – 108) varies just a little, it is possible that the measurement techniques used in the estimation are not accurate enough to detect the corresponding small change in the mass transfer [86]. In chapter 4 (§ 4.2.1), the change in the char particle's location in the bed for two different fluidization velocities is quantified, proving that in such conditions (bubbling regime) these locations are barely affected. Publications addressing the CO/CO2 ratio generated by carbon oxidation agree on the difficulty of avoiding the oxidation of CO before measuring it. There is also a dependence on the fuel temperature, but if the combustion takes place in a FB, the ratio is independent of the bed temperature and the fluidization velocity [32, 42]; the inert bed particles inhibit the CO oxidation close to the char particle [42]. The CO/CO2 ratio estimated by Arthur [71] is used here, because in his study different coal ranks and an effective inhibitor of the CO oxidation were investigated. This estimation (Eq. 2.3, Table 2.1) is assumed independent of the position of the char particle in the bed, because there is a slight cloud of bed particles around the char particle when it is in the bubble phase and splash zone, as it is discussed in chapter 4 (see Fig. 4.14 and Fig. 4.18). In summary, the model used here to estimate the temperature and burnout times includes Eqs 2.1 – 2.27, along with the values of the parameters shown in Tables 2.1, and Eq. AII.1 in Appendix II. 2.6. NUMERICAL METHOD AND SOLUTION PROCEDURE In addition to the values of the char particle’s and fluidization flow’s properties summarized in Table 2.1, and the operating conditions of the bed for each test, the theoretical approach requires the char particle's movement through the bed. This movement, whose characterization is in Appendix I, gives the needed time to complete a cycle (T, see Fig. 2.2) and the residence time in each bed phase (ti,T) as a function of the char particle's size (diameter). Therefore, given the initial size of the char particle, it is known the place occupied for it during the next T seconds. For every 0.04 seconds, it is calculated all fluxes from/to the char particle's surface (Ni,c) solving Eqs. 2.3 - 2.16, along with 2.22 if the char is at the emulsion phase, 2.24 if the char is in the bubble phase, and 2.26 if the char is in the splash zone. Eq. 2.17 and 2.18 are then simultaneously solved giving the size of the char particle (rc) and temperature (Tc) after 0.04 seconds of conversion. Noted that the energy balance (Eq. 2.18) used the heat transfer correlation corresponding to the char’s place in the bed at this moment time: 2.23 for emulsion phase, 2.25 for bubble phase, and 2.27 for splash zone. CHAPTER 2 14 Table 2.1. Parameters of the model Symbol Value Unit Ref. Char Particle Density ρc 405 @ BW* kg/m3 1090 @ SBC* Emissivity 𝜔c 0.85 @ BW - [60] 0.9 @ SBC - [93] Specific heat cp 13.3 @ BW J/mol K [28] 12.6 @ SBC J/mol K [28] Enthalpy ∆Hc cp (Tc - 298) J/kg - Fluidized bed Bed particles’ density ρi 2560 kg/m3 [24] Bed’s density ρb 1440 kg/m3 - Bed’s Porosity εmf 0.44 - Bed particles’ specific heat cp,i 798 J/kg K [41] Bed particles’ conductivity ki 0.33 W/mK [41] Minimum fluidization velocity umf 0.18 m/s - Diffusivitties** O2 diffusivity in N2 dO2-N2 2.15∙104 (T/1073)1.81 m2/s [101] O2 diffusivity in CO2 dO2-CO2 1.67 104 (T/1073)1.81 m2/s [101] O2 diffusivity in CO dO2-CO dO2-N2 m2/s [72] CO2 diffusivity in N2 dCO2-N2 1.7 104 (T/1073)1.81 m2/s [101] CO diffusivity in N2 dCO-N2 dO2-N2 m2/s [72] CO2 diffusivity in CO2 dCO-CO2 dO2-CO2 m2/s [72] Kinetic Pre-exp. factor in Eq. 2.3 ACO/CO2 2512 - [71] Activation energy in Eq. 2.3 ECO/CO2 51880 J/mol [71] Combustion pre-exp. factor Acb 0.658 m/Ks [102] Combustion activation energy Ecb 74800 J/mol [102] Gasification pre-exp. factor Agf 3.42 m/Ks [79] Gasification activation energy Egf 129700 J/mol [79] * BW: Beech wood, SBC: Sub-bituminous coal, **Diffusivities evaluated at T = (Tc + Tbed)/2 15 CHAPTER 3. EXPERIMENTAL In the present Chapter the experimental set-up developed in this work is presented, as well as the theoretical basis of the method used for the design and development of optical measurements, including the processing of images. 3.1. PYROMETRY WITH A DIGITAL CAMERA Pyrometry is based on the capture and interpretation of the thermal radiation emitted from a body whose temperature is to be measured. A common digital camera can carry out this capture and interpretation if the thermal radiation includes the visible spectral band. The light from the scene is separated into three spectral bands or colors: red (0.62-0.7 μm), green (0.49- 0.58 μm), and blue (0.45-0.49 μm) by a filter place over the sensor [60]. When the radiation touches the sensor, each pixel generates three values known as Digital Numbers (DNk) that measure the radiation belonging to each mentioned spectral bands. By an image program three matrices for each image are obtained, whose elements (m,n) contain the digital numbers associated with the radiation received from the viewed scene. Fig. 3.1 shows the recreation of the capture and interpretation of the radiation from the viewed scene by the digital camera (a) and the digital numbers’ example from an image of a wave on the lake's water surface (b). It should be known that the sensor has a lower limit energy level (W/m2) and for radiation below it, the digital number assigned by a pixel is null. It also has a higher limit energy level, above from which the pixel becomes saturated and a maximum value is assigned (saturation) [88]. The water temperature shown in Fig. 3.1b cannot be measured using the visible spectral band because the light captured by the digital camera is from the sun. However, if a body emits thermal radiation within the visible spectral band due to its temperature, its temperature could be measured if the digital numbers from image are used (properly implemented in the heat transfer model). This is the basis for the temperature measurement by pyrometry with a digital camera, whose detailed discussion is given in Appendix III. Reviewed works from literature that measure the coal/char particles' temperatures in fluidized beds by pyrometry use simultaneously the information contained in two spectral bands [32, 56- 60, 65-68]. This application, named two-color pyrometry (P2C), allows calculating the temperature by Eq. 3.1. The method works provided the temperature of the object to be measured is higher than that of the background and when the surface is grey (see § AIII.2) [32, 56, 58]. It has significant advantages over other techniques used in FB, but there is a disadvantage CHAPTER 3 22 where dc,o is the initial char particle’s diameter, quantified in both perpendicular ( k = ⊥) and parallel ( k = //) directions with respect to the insertion of the thermocouple into the particle from the surface to the center. The results from the FBC 3D set-up were compared with those from FBC 2D. According to the definitions in Fig. 3.6 a char particle on the bed’s surface is associated with the splash zone when more than half of its volume is out of the bed (otherwise it is considered to be in the emulsion phase, see zooms in Fig. 3.6). By another definition, for example, a char particle on the surface of the bed is in the splash zone if part of its volume is out the bed (not necessarily more than half of its volume), the interpretation of the images would be different, and therefore, the fractions of the combustion time when the char particle is in these places (𝜃i, Eq. 3.3). However, the differences between the fractions calculated for the cases with and without a thermocouple would be equivalent. Moreover, even if the definition is changed, it would not influence the percentages associated with the bubble phase as Chapter 4 discusses (see § 4.2), since one of the differences between the conversion of a char particle with and without a thermocouple is that the thermocouple increases the presence of the char particle in the bubble phase. Fig. 3.6. Definition of the char particle’s position in the bubbling fluidized bed reactor TESTS AND PROCEDURE FOR EXPERIMENTAL MEASUREMENT AND THEORETICAL PREDICTION OF CHAR TEMPERATURE OSCILLATIONS Combustion tests of single char particles freely-fluidized into the FBC 2D were carried out to study the effect of the char particle’s movement through the emulsion phase, bubble phase, and splash zone on its combustion temperature. The bed temperature was set at 800 °C and the fluidization agent was a mixture of O2 and N2, where the O2 concentration was either 11 or 21%v/v. The char particles were from beech wood and sub-bituminous coal with spherical form (dc,o ~ 8 mm), whose weights were fixed at 0.09 and 0.28 g respectively. The images analyzed by pyrometry were in the form of sequences shorter than one second, during which the char particle changes its location. These experimental results (not just the char particles' combustion EXPERIMENTAL 23 temperatures, but their oscillation as consequence of their movement through the bed) were compared with the theoretical prediction of the conversion model presented in Chapter 2. As Chapter 2 explains (see § 2.3), the theoretical analysis needs the char particle’s trajectory and velocity passing through the bed as input data. Characterizing this char particle’s movement involves the fitting of the combustion time fractions during which the char is in each bed phase (θi, by Eq. 3.3) with its size (dc, by pyrometry): θi,T=midc+ni (3.5) where mi and ni are fitting constants (note that in this study m in Eq. 3.3 is the number of minutes whose images are classified and the char particle’s size is approximately similar). In this way, as Appendix I describes, the time spent for the char particle to complete a cycle (T) and its distribution among the different phases (ti,T) can be calculated. Fig. 3.7 shows an example for a char particle from beech wood (8 mm), where the residence time in each phase has been lumped (Fig. 2.2b vs. Fig. 3.7). Therefore, the theoretical analysis assumes that the particle is in the emulsion phase at the initial time and with its initial size calculates its movement through the bed (the time spent in each phase). The conversion is simulated until the end of the first cycle. Then the simulation continues with a new description of the cycle by the char particle’s size reached after T seconds of conversion. The process is repeated until the conversion reaches 99 %. Fig. 3.7. Time and location of a char particle from beech wood (8 mm) during a cycle comprising its movement through the bed: emulsion phase (ePh), bubble phase (bPh), and splash zone (sPh) under the experimental conditions shown in Table 3.1 Note that the cyclic movement is calculated for a given char size. This size gives the residence time in each bed phase (tj,T = θi,T T, see Eq. AI.I) where the conversion takes place. However, during these T seconds, the char size is reduced, changing tj,T (since θi,T changes with the char size, see Eq. 3.5), and that change is not taken into account until the new cycle. Because the cycle time is much shorter than the total combustion time (burnout time), this simplification seems reasonable. On the other hand, lumping the residence times of the char particle in each bed phase reduces the char/emulsion and char/bubble interruptions to a single one: once the char leaves the emulsion phase, the bubble time starts (Fig. 2.2b vs. Fig. 3.7). This simplification makes the conditions for char conversion to change instantaneously. The results in Chapter4 (see § 4.3.1) prove that the temperature variation as result of changing the location of the char in the bed is produced in hundredths of a second. 0 0.5 1 1,5 2 2,5 3 3.5 ePh bPh sPh time, s Char particle's place in the bed CHAPTER 3 24 TESTS AND PROCEDURE FOR THE EXPERIMENTAL AND THEORETICAL ANALYSIS OF THE CO2 EFFECT ON THE CHAR PARTICLE‘S TEMPERATURE AND CONVERSION Oxy-fuel combustion and combustion tests of single char particles in FBC 2D were carried out to address their conversion in an O2/CO2 atmosphere with high O2 concentration. The bed temperature was either 800 or 850 °C, and the O2 concentration in the fluidization agent were either 11, 21, 30, 40, and 50 %v/v in N2 or CO2. The char particles were from beech wood and sub-bituminous coal with spherical form. For each type of char, two different weights were fixed at 0.09 and 0.28 g to the char from beech wood, and 0.28 and 0.66 g to the char from subbituminous coal, where the diameters were 8 mm for the lighter weight and 13 mm for the heavier one. The images analyzed by pyrometry are sequences where the char particle changed its positions in the bed. In this way, the variation (oscillations) in the temperature of the char particle’s surface and its size (external diameter) could be tracked. Note that the char particle’s size could be measured just for some of the images where the particle is completely seen (see Fig. 4.10). The ratio of these diameters (dc) and the initial one (dc,o) were used to estimate the experimental conversion by xc=1−(dc dc,o)3 (3.6) The last image of the char particle is the last time when it was seen, whose time was set as burnout. Since that the char particle’s size in this image was lower than two millimeter (see Table 4.2), the experimental conversion in all the tests was higher than 99 %. On the other hand, the apparent consumption rate can be estimated using the burnout time, and the char particles’ content of carbon (xc,fx), masses (mc,o) and initial external surfaces (sc,o) by Γc=xc,fx mc,o sc,otburnout (3.7) to analyze the CO2 effect on the char particle conversion. The theoretical predictions by the model are checked comparing them with both experimental temperatures and burnout times from this work and others from the literature. The chosen works used different measurement techniques and experimental oxi-firing and air-firing conditions, covering different ranks of coals, fluidization velocities, bed temperature, etc. (see Table 4.4 where there are the main characteristics of these works). EXPERIMENTAL 25 Table 3.2. Experimental and theoretical studies carried out in this thesis Calibration and validation of the pyrometry with a digital camera Combustion of single char particles placed in the oven (FBC 2D, without FB) with an embedded thermocouple (dth = 0.75 mm) into the char's surface gap Rectangular char from beech wood: 7 x 7 x 14 mm O2/N2 atmosphere @ [O2]: 0-50 %v Toven: 800 °C Impact of using thermocouples on the char particle combustion in a FB Combustion of single char particles in FBC 2D and in FBC 3D with an embedded thermocouple and freely fluidized Spherical char from beech wood, and sub-bituminous and bituminous coal: dc,o ~ 10 mm dth: 0.25, 0.5, and 0.75 mm uf /umf: 2 (bubbling conditions) Air as fluidization flow Tbed: 800 °C Oscillation in the char particle’s temperature in a FB Single char particle combustion in FBC 2D Spherical char from beech wood and sub-bituminous coal, dc,o ~ 8 mm O2/N2 fluidization flow, [O2]: 11 and 21 %v uf /umf: 2 (bubbling conditions) O2/N2 atmosphere @ [O2]: 11 and 21 %v Tbed: 800 ºC Influence of CO2 gas concentration on the char conversion during oxy-fuel combustion in a FB Single char particle combustion in FBC 2D Spherical char from beech wood, and sub-bituminous coal: dc,o: 7-8 vs. 12-13 mm uf / umf: 2 (bubbling conditions) O2/N2 vs. O2/CO2 fluidization flows @ [O2]: 11, 21, 30, 40, 50 %v Tbed: 800 ºC vs. 850 ºC 27 CHAPTER 4. RESULTS AND DISCUSSION This chapter presents the results organized into four parts: in the first part, the method developed to measure the temperature of the char particle's surface during conversion in a FB using a digital camera are discussed (calibration and validation tests and results). In the second part, the method is applied to analyze the impact of using an embedded thermocouple into a char particle in a FB, by comparing the conversion characteristics (temperature and rate of conversion/burnout times) from experiments using freely-fluidized particles with those using char particles with an embedded thermocouple. In the third part the method is applied to capture oscillations in the char temperature during conversion, and the measurements are compared with a model developed taking explicitly into account the movement of char particle through the phases of the FB. In the fourth part, the measurement method is used in dedicated experiments to study the influence of the CO2 concentration in a gas with high O2 concentration atmosphere (up to 50%v/v) useful for the analysis of FB oxy-fuel boilers. 4.1. MEASUREMENT OF THE CHAR PARTICLE’S SURFACE TEMPERATURE BY PYROMETRY WITH A DIGITAL CAMERA 4.1.1. CALIBRATION OF PYROMETRIC TEMPERATURE Measuring the char particle’s temperature needs the value of the calibration parameter: βi/j in Eq. 3.1 for P2C and βk in Eq. 3.2 for P1C, that is, the DNk - Tc - βk relation of the digital camera (see § 3.1). Therefore, images from ten combustion/calibration tests (1273 images) were analyzed relating the digital numbers from the char particle’s surface to the corresponding temperatures measured by the thermocouple as sketched Fig. 4.1. The temperature (radiation) detected by the sensor in each spectral band or color is shown in Fig. 4.2a, where it can be seen that the digital numbers (DNk) increase with the surface temperature. The bands at longer wavelengths have lower saturation temperature (temperature when the maximum value of DNk is reached), since the emitted and/or received radiation (W/m2μm) increases with temperature and wavelength [91]. The DNr dispersion at temperatures 600-725 °C is caused by the fast increment of surface temperature at the beginning of the combustion. The temperature measured by thermocouple and the images were recorded with a rate of one per second, so the small time- CHAPTER 4 28 difference between two measurements explains this dispersion. There is no dispersion in DNg or DNb because at these temperatures the sensor does not detect the thermal radiation. The relation between the temperature of char particle’s surface and the ratio of the red/green (DNr/DNg) and green/blue (DNg/DNb) digital numbers for temperatures higher than the background (800 °C), is shown in Fig. 4.2b. The DNg/DNb ratio versus temperature lower than 850 °C is not plotted because of the high noise-to-signal ratio in the blue digital numbers at lower temperature than 850 °C. Fig. 4.1. Relation between the average digital numbers of the char particle’s surface whose temperature is measured by thermocouple Introducing the relations shown in Fig. 4.2a and Fig. 4.2b into the Eq.AIII. 17 and Eq. AIII.18 respectively, are obtained the points ln(DNi/DNj) – ln (βi/j) and ln(DNk) – ln(βk) shown Fig. 4.3, whose fitting by ln(βi j ⁄)=ai j ⁄(ln(DNiDN ⁄j))2+bi/jln(DNiDN ⁄j)+ci/j (4.1) and ln(βk)=ak(ln(DNk))2+bkln(DNk)+ck (4.2) give the calibration curves (also shown in Fig. 4.3) to measure the pyrometric temperature by two-color pyrometry (P2C) and one-color pyrometry (P1C) respectively. The coefficients values of Eq. 4.1 and Eq. 4.2 are in Table 4.1, where the of the wood char’s emissivity in Eq. AIII.18 (P1C) was assumed to be 0.85 [54, 60]. Table 4.1. Calibration curves’ coefficients in Eq. 4.1 for P2C and in Eq. 4.2 for P1C a b c r2 P2C red/green - 0.0098 - 0.6949 3.1392 0.9969 green/blue - 0.0464 - 0.6599 2.2199 0.9958 P1C red 1.2998 - 12.475 22.978 0.8280 green 0.1665 - 1.2089 - 5.4118 0.8178 blue 0.3136 - 2.4066 - 5.3567 0.8634 RESULTS AND DISCUSSION 29 Fig. 4.2. Char particle’s surface temperatures measured by thermocouple versus digital numbers (a) for red (DNr), green (DNg), and blue (DNb) and versus ratio of the digital numbers (b) for red/green (DNr/DNg), and green/blue (DNg/ DNb) Fig. 4.3. Calibration curves by fitting {ln(DNk) – ln(βk)} and {ln(DNi/DNj) – ln (βi/j)} for P1C (a) and P2C (b) respectively 600 650 700 750 800 850 900 950 1000 1050 0 50 100 150 200 255 Temperature, oC Digital number, - 600 650 700 750 800 850 900 950 1000 1050 0 2,5 5 7,5 10 Temperature, oC DNi/DNj ratio, - (b) (a) red green blue green/blue red/green 0 1 2 3 4 5 6 -10 -8 -6 ln (DNk) ln (k) 0 1 2 3 4 5 6 -2 0 2 4 ln (DNi/DNj) ln (i/j) green blue red green/blue red/green (b) (a) CHAPTER 4 30 As shown Fig. 3.4, once the calibration curve are known (the relation between the radiation received on the sensor and the digital number generated), the char particle’s surface temperature can be measured. The results of measuring it on the validation tests (429 images) using the information contained in the red spectral band (P1C R), in the green spectral band (P1C G), and simultaneously both (P2C R/G) are shown in Fig. 4.4. The measurement of the char particle's surface temperature by P1C R agrees with the one by thermocouple for temperatures from 600 to 800 °C. Above 800 °C, the pixels are saturated (see Fig. 4.2a) and P1CrR cannot measure temperature (radiation) higher than the radiation that saturates the pixels. Measurement by P1CgG has a lower limit approximately at 700 °C since the weak captured radiation at lower temperature than that (DNg < 1, Fig. 4.2a). From this temperature (radiation) to 950 °C, the accuracy of P1C G is good, but then saturation occurs. The temperature measured by P2C RG can measure surface temperatures higher than the background (see § AIII.2), and its accuracy is good enough until the most energetic spectral band is saturated at 950 °C (see Fig. 4.2a). On the other hand, if the measurements by P1C G and P2C R/G are compared, it can be seen a similar predictive capability. This is due to the saturation of the pixels in the red spectral band (DNr, see Fig. 4.2a), while the associated digital numbers of the green (DNg, see Fig. 4.2a) remain active at higher temperatures. Two-color pyrometry relates the thermal radiation from two spectral bands; once the radiation associated with the less energetic spectral band saturates, the digital numbers of the most energetic spectral band determine the measurement [69]. Fig. 4.4. Char particle’s surface temperature by thermocouple and by one-color (P1C R and P1C G) and two-color (P2C R/G) pyrometry 4.1.2. SEQUENTIAL ONE-COLOR PYROMETRY Fig, 4.4 shows that if the measurements by P1C R and P1C G are combined sequentially, the temperature range that can be measured and its accuracy are improved compared to the traditional simultaneous use of P2C R/G. The result is shown in Fig. 4.5, where P1C R is used until the radiation received by the red spectral band was close to saturation (DNr ~ 245). After reaching this value, the algorithm calculates the temperature by P1C G. Note that while P2C RESULTS AND DISCUSSION 31 R/G uses the information from thermal radiation contained in different colors simultaneously, the algorithm proposed here (P1C RG) employs this information in a sequential way according to the digital numbers. Fig. 4.5. Char particle’s surface temperature by thermocouple and by pyrometry using the red and green spectral band sequentially (one-color pyrometry, P1C RG) or simultaneously (twocolor pyrometry, P2C R/G) As shown Fig. 4.5, within the most interesting temperature range for char particle combustion in a fluidized bed (800-950) °C, both methods (P1C RG and P2C R/G) are similar. However, the char particle’s surface temperatures during oxy-combustion could be higher than the maximum temperature reached and measured in this work 950 °C. At temperatures higher than that, DNg is saturated (saturation depends on the camera as well as on the working conditions; WD, FOV). One solution to measure even higher temperatures could be the use of the information from the blue digital numbers (P1C B). Fig. 4.6 shows the result from the blue spectral band (even when the green digital numbers are not saturated) validating the proposed solution because the blue digital numbers at this temperature were around 130 (see Fig. 4.2a), so there are still more than 100 values until saturation and higher temperature than 950 °C can be measured. Another solution could be to use simultaneously green and blue (P2C G/B). In this way, the algorithm would change from P2C R/G to P2C G/B according to the DNr/DNg value. In order to avoid the reduction in the accuracy of the measurement, the change from P2C R/G to P2C G/B should be made when DNr/DNg is close to unity. If the change is made at higher values than that, the thermal radiation belonging to the blue spectral band is weak (see Fig. 4.2a), and the associated DNg/DNb value is very large and it is not reflected in 𝛽 g/b , as shown in a mathematical way by (DNg DNb)1≫(DNg DNb)2→ln(DNg DNb)1~ln(DNg DNb)2→ln (βg/b)1~ln (βg/b)2 (4.3) CHAPTER 4 38 by an optical probe during combustion in sand particles (182 and 1090 μm) fluidized at two velocities (uf/umf = 1.3 and 2), and at bed temperatures between 700 and 827 °C, noticing that the burning particles are associated with the emulsion phase for only 60 % of the time in the FB. Agarwal et al, [33] theoretically estimated a residence time in the emulsion of 80 % for the conditions conducted by La Nauze et al, [95] (size of the sand particle of 655 μm, gas velocities uf/umf from 2 to 3.5 and bed temperatures of 700, 800 and 900 °C). The above discussion makes questionable the usual hypothesis assumed in FB models that the active particle is only converted in the emulsion phase [96], particularly if the char particle is fluidized with an embedded thermocouple. 4.2.2. IMPACT ON THE CHAR PARTICLE’S SURFACE TEMPERATURE The effect of the thermocouple on the char particle’s movement and, particularly, the increase in the residence time in the bubble phase (where the combustion temperatures are higher [32]), have consequences on char conversion that must be studied and quantified. To study this effect, sub-bituminous char particles (~10 mm, 0.56 g) were burnt with a thermocouple (w th) and without it (w/o th). The result is shown in Fig. 4.14, where the temperature in the center of the char particle measured by thermocouple (Tth) is compared with the surface temperatures measured by one-color pyrometry (TP1C). Moreover, the burn-out time is shown. Fig. 4.14. Temperatures by thermocouple (Tth) and by one-color pyrometry (TP1C) during combustion of two single char particles fluidized with an embedded thermocouple (dth = 0.25 mm, w th) and without it (w/o th) in a fluidized bed (Tbed = 800 °C, uf/umf = 2). The times indicated are the burnout time of char with (1250 s) and without (1336 s) thermocouple The temperature recorded by the thermocouple reveals one of the main problems associated with a contact technique: the detachment of the char particle from the thermocouple. The thermocouple reads the end of combustion at 840 seconds, when the recorded temperature becomes equal to the bed temperature (Tth = Tbed = 800 °C). However, the char combustion is RESULTS AND DISCUSSION 39 far from finished, as shown by the surface temperature recorded by pyrometry. The differences between the surface temperatures measured by pyrometry when a thermocouple is embedded or not (TP1C in Fig. 4.14, w th vs. w/o th) reveals that the temperature is always higher when a thermocouple is used. This difference between the two cases is more pronounced when the thermocouple is attached (tc< 840 s). This is due to the longer time the char particle with a thermocouple spends in the bubble phase where the combustion temperature is higher because of the higher mass transfer [83]. Once the detachment char particle– thermocouple has occurred (tc > 840 s), the difference between the surface temperatures of the char with and without a thermocouple is reduced, because the fluid-dynamic effect of the bed (quenching) is similar on both particles. However, the surface temperature of the char particle, which had a thermocouple, remains higher than that without a thermocouple, and also the end of the combustion is reached earlier: at 1250 vs. 1336 s (reduction higher than 5 % in the burnout time). A similar result has been published in [42], where a reduction of 8 % was recorded, using a thermocouple with a 0.5 mm sheath diameter. The char particle with a thermocouple attained has a higher temperature than the freely-fluidized particle, because its size is smaller due to the higher surface temperature during the time when the thermocouple was attached. This is shown in Table 4.2, where the size of both char particles is measured just after the detachment, and at the end of both combustion processes (in the last image where the char particle is seen). Table 4.2. Char particle’s sizes during combustion with an embedded thermocouple (dth = 0.25 mm, w th) and without it (w/o th), at detachment time (15 min), and at the end of both combustion events (20 and 22 min) dp*,mm 15 min 20 min 22 min w th 4.5 1.8 - w/o th 5.2 2.8 1.7 *Sizing method’s uncertainty: ±0.3 mm,- Complete consumption The temperature-gradient on the char particle’ surface, which ranges from TP1C,max to TP1C,min in Fig. 4.14, is larger when the char particle is attached to the thermocouple (tc< 840 s) than on the freely fluidized char particles: 50-80 vs. 20-50 °C respectively. Once the detachment has taken place (tc > 840 s) this difference is reduced, and the gradients fall within the same range for all particles, which is due to the higher surface temperatures caused by the longer time in the bubble phase (see Fig. 4. 11). This is proved in Fig. 4.15 where is shown the surface temperatures of the char particles with (a) an embedded thermocouple and without it (b) at some combustion times taken from Fig. 4.14. On the other hand, Fig. 4.15 points out that the higher temperatures are always in the same surface region: the one opposite to where the thermocouple is inserted. Because of the thermocouple, the char particle is not allowed to rotate freely, and the same part of the char particle’s surface always faces the fluidization flux and bubbles. 4.2.3. IMPACT ON THE CHAR PARTICLE’S SURFACE CONSUMPTION The char particle’s orientation is characterized by the angle  between the direction of insertion of the thermocouple into the char particle and the direction of the fluidization flux (from the reactor's distributor to upwards). The effect of the thermocouple on this orientation is sketched in Fig. 4.16a, where the fraction of the char particle’s surface with higher temperature CHAPTER 4 40 is shown. The restriction of the particle's free rotation means that the same surface is always at higher temperature during the entire conversion and, as shown in Fig. 4.16b, this affects the consumption of the surface. Therefore, the reduction of the char particle’s size in the perpendicular direction (⊥) relative to the insertion direction of the thermocouple should be higher than the one in the parallel direction (//). Moreover, if the char particle is attached to the thermocouple long enough, the temperature recorded by the thermocouple would attain the bed temperature (indicating the end of the combustion Tth = Tbed), still attached to the particle. Fig. 4.15. Impact of a 0.25 mm thermocouple on the surface temperature of a sub-bituminous char particles fluidized with (a) and without (b) it (Tbed = 800 °C, uf/umf= 2) RESULTS AND DISCUSSION 41 Fig. 4.16. The effect of the thermocouple on the char particle’s orientation (a) and surface consumption at perpendicular (⊥) and parallel (//) direction relative to the insertion direction of the thermocouple into the char particle (b) To prove that, char particles fluidized with an embedded thermocouple are burnt. The particles are extracted from the bed at different combustion times (tc), and their diameters are measured in the perpendicular (⊥) and parallel (//) direction with respect to the directions of the insertion of the thermocouple. Fig. 4.17 shows the sub-bituminous (a) and bituminous (b) char particles after 12 minutes from the beginning of the combustion, and a char particle from beech wood (c) extracted after 6 minutes where the thermocouple is fixed with enough hightemperature resistant sealant to avoid the detachment at high char conversion. Moreover, the direction of the fluidization air is drawn. Higher consumption in the perpendicular direction than in the parallel one can be seen, supporting the hypothesis of non-homogeneous consumption caused of the thermocouple illustrated in Fig. 4.16b. Fig. 4.17. Char from sub-bituminous (a), bituminous (b), and beech wood (c) fluidized with a 0.25 mm thermocouple, and extracted from the bed at different combustion times (tc) (bituminous char is cut through the middle, and the direction of the fluidization air is shown) To discard the possibility that the results obtained are affected by the rectangular geometry of the CFB 2D reactor (see Fig. 3.4), char combustion tests were carried out also in a 3D cylindrical FB reactor (see Fig. 3.5b). Char particles (from beech wood, sub-bituminous and CHAPTER 4 42 bituminous coal) were extracted from the reactor at various combustion times (tc), and their diameters in the perpendicular (⊥) and parallel (//) directions with respect to that of the thermocouple insertion were measured (𝛿k, by Eq. 3.4). The sheath diameter of the thermocouple was 0.25 mm. Fig. 4.18 compares these results and shows that there is a nonhomogeneous surface consumption in both set-ups: FBC 2D, and FBC 3D, and therefore, it is verified that the thermocouple affects the conversion of the char. In particular, the size reduction in the perpendicular direction (⊥) is more than 10 % faster than the reduction in the parallel direction (//) in all cases. The results shown in Fig. 4.18, involve that if the geometry of the 2D reactor is affecting to the place of the char particle and its movement through the bed, such extra restriction (the one introduced by the width of the reactor 2D) is negligible with respect to the restriction to the movement of the char particle by the embedded thermocouple, at least in the experimental conditions of this work. Although the evolution of the size of a freely-fluidized char particle was not systematically followed during conversion in this study, the shrinkage of the char is qualitatively assessed below (see Fig. 4.26) by visualizing the volume of the particle at different instants (see Fig. 4.10), verifying that the size reduction was uniform along time. These observations are in agreement with previous work [53, 97], indicating that the non-uniform consumption of the char particle reported in Fig. 4.18 is produced by the use of the embedded thermocouple. Fig. 4.18. Beech wood (a), Sub-bituminous (b), and Bituminous (c) char particle’s reduction in the perpendicular (⊥) and parallel (//) directions to the thermocouple insertion at different combustion times (tc) in the FBC 2D and 3D set-ups (0.25 mm sheath diameter) 4.3. MEASUREMENT AND THEORETICAL PREDICTION OF THE OSCILLATION IN THE CHAR PARTICLE’S TEMPERATURE IN A FB 4.3.1. EFFECT OF THE CHAR PARTICLE’S MOVEMENT THROUGH THE BED ON THE COMBUSTION TEMPERATURE A sequence of images during 0.6 seconds from the combustion of a char particle from beech wood and its surface temperature are presented in Fig. 4.19. These images show the position of the char particle in the bed and the intermittent residence in the emulsion and bubble phases RESULTS AND DISCUSSION 43 when the char moves upward through the bed (74.60″ - 74.80″). After that, there is a period in the splash zone (74.84″ - 75.20″) where the char particle is affected by the bursting of the bubble. This bubble collapse can throw the char a few centimeters above the bed (74.84″ - 74.88″; 75.00″ - 75.08″), or cover it with sand projected (74.96″; 75.20″). The measurement of the char’s surface temperature allows analyzing and quantifying its variation during the movement through the bed: higher char temperature in the bubble and/or splash zones than that in the emulsion phase (74.68″ vs. 74.7″; 75.16″ - 75.20″) and temperature oscillation of tens of degrees in periods shorter than eight hundredths of a second (74.64″ vs. 74.68″ vs. 74.72″). Since these tests concern a single char particle, there is no difference in the O2 concentration in the emulsion phase, bubble phase, and splash zone. Therefore, the temperature variations are due to the different heat and mass transfer from/to the char particle depending on its position in the bed: the mass transfer is decreased when the particle is immersed in the bed of particles where the heat transfer rises [55]. The heat generated at the char’s surface could also change with the char position in the bed, because the sand particles of the bed inhibit the CO oxidation to CO2 [42], but this is not considered the main reason for the measured temperature oscillations. If this happened, the measured temperature variations would be higher (317 kJ/mol for C + O2 → CO2 vs. 133 kJ/mol for C + O2 → 2CO, @ Tc = 800 °C), and it can be seen in Fig. 4.19 that the char particle is surrounded by a cloud of sand particles. After applying the analysis discussed above describing the time intervals during which the char particle changes its location in the bed, emulsion phase (ePh) ↔ bubble phase (bPh) ↔ splash zone (sph), the surface temperature oscillations during the entire conversion can be quantified. Fig. 4.20 shows these oscillations from the conversion of char particles from beech wood (a) and sub-bituminous coal (b) (O2/N2, at an O2 concentration of 11 and 21 %v/v), where the length of the time intervals analyzed was shorter than 0.4 seconds. The figure bars show the temperature range measured for particles of the same size and in the same phase hundredths of second before, or after changing their place in the bed. Therefore, it is concluded that the temperatures are more stable when the particle is in the emulsion phase than in the bubble and/or splash zone. This result, which has already been pointed out elsewhere [32], is due to the stronger heat convective transfer by the sand particles in the bed, making the temperature of char staying in the emulsion phase uniform, as well as to the different conditions for convective mass transfer to a char particle outside of the emulsion phase. Note that the gas velocity surrounded the particle, ug, is different if the char is reached by a bursting bubble at the bed’s surface, ug >> umf (74.80″ in Fig. 4.19), or if it is on the bed’s surface, ug = uf (75.12″ in Fig. 4.19), or inside a bubble, ug = uth (74.68″ in Fig. 4.19). On the other hand, Fig. 4.20 shows that the temperature oscillations as result of changing the location of the char in the bed, i.e. the difference in the char temperatures in different phases, vary from 10 to 100 °C. The amplitude of these oscillations increases with higher O2 concentration and smaller char particle’s size because external mass transfer controls the char consumption and these changes increase the O2 molar flux to the char surface. CHAPTER 4 44 Fig. 4.19. Image (grey scale) sequences of 0.6 seconds from combustion of beech wood char in fluidized bed (Tbed = 800 ºC, uf/umf = 2, 21%v O2 in N2) and its surface temperature by onecolor pyrometry with digital camera (the dashed line is drawn to mark the bed’s surface and bubbles) RESULTS AND DISCUSSION 45 Fig. 4.20. Surfacetemperatures of char from beech wood (a) and sub-bituminous coal (b) in the emulsion phase (ePh), bubble phase (bPh), and splash zone (sPh) during combustion at 11 %v/v (empty symbols) and 21 %v/v (filled symbols) O2 concentration in N2 in a bed at 800 °C As discussed in Chapter 1 (see § 1.2), there are publications showing temperature oscillations during coal/char combustion [32, 35, 56, 58], some of them are even associated to different char positions in the bed (emulsion or bubble phase) [32, 35]. However, those oscillations were recorded by thermocouple and are slower than the ones presented here; that is, the time from the minimum to the maximum temperature in an oscillation is longer than that shown in Fig. 4.19: seconds vs. hundredths of a second. It is possible that a thermocouple records the effect of a complete cycle of char particle’s movement on the combustion temperature (see Fig. 2.2a). Therefore, the maximum temperatures would be from the time when the char particle is on the bed surface, or close to that (at the highest location of the movement shown in Fig. 2.2a) and the minimum one would be from the moment when the particle is close to the lowest point of its downward motion through the emulsion phase. On the other hand, the results from pyrometry with an optical probe show different combustion temperatures because of the intermittent char residence in the bubble and emulsion phases, and the size distribution of the char/coal particle in the bed. Although it is difficult and uncertain to estimate the particle size by pyrometry with optical probe [57], results related to the maximum particle size of 1 mm reveal a temperature oscillation of 100 °C [56]. 4.3.2. PREDICTION OF THE OSCILLATION IN THE CHAR PARTICLE’S TEMPERATURE The combustion model needs the implementation of the char particle’s movement through the bed (see § 3.2.2). This movement is characterized using the fractions of the combustion time during which the particle is in each bed phases (see Fig. 4.12), whose sizes range from 10 mm to 1 mm. The fractions is related to the char particle’s size by Eq. 3.5 whose fitting coefficients values are in Table 4.3. Using these fitting lines, along with the estimations of the bubble velocity and CHAPTER 4 46 the upward char velocity, the cycle time (T) and the time spends for a char particle with a given size in each bed phases (see Eq. AI.1) are calculated. These cycles for char particles from beech wood (a) and sub-bituminous coal (b) are shown in Fig. 4.21. Note that the period increases slightly as the char size decreases, which is due to the increase in the residence time in the emulsion phase as consequence of lower velocity of the descending char (Eq. AI.10, Appendix I). Table 4.3. Fitting coefficients in Eq. 3.5 Beech wood char Sub-bituminous char mePh - 46.26 - 15.81 nePh 1.00 9.50∙10-1 mbPh 9.81 9.75 nbPh 1.91∙10-2 0.0111 msPh 39.69 6.06 nsPh -5.47∙10-2 3.85∙10-2 Fig. 4.21. Periods of the char particle’s movement through the bed: char from beech wood (a) and sub-bituminous (b) The theoretical temperature for a single char particle (from beech wood, 0.09 g) combustion in a sand bed fluidized with a mixture of O2 and N2 (21 %v/v O2) is shown in Fig. 4.22. The higher and lower temperatures of the oscillations are delimited by curves. The zoom from 301 to 309 seconds shows how the temperature is affected as the position of the char particle in the bed changes during two cycles. During the downward movement of the char through the emulsion phase, the combustion temperature decreases because of the weaker mass transfer and stronger heat transfer compared to those in the bubble/splash zone. Note that, if the temperature of the char is lower than the bed temperature, it increases during the first seconds after the feed. Once the particle is in the bubble phase and/or in the splash zone, the higher mass transfer increases the carbon consumption and the heat generated on the char surface makes the char temperature to increase. In this way, the maximum and minimum temperatures delimited by the curves in Fig. 4.22 meet the char temperatures at the highest and lowest point of the char cycle in the bed [43] (see Fig. 2.2). Moreover, the model result predicts lower char temperature in the emulsion phase, something that was already mentioned above (see § 4.3.1). RESULTS AND DISCUSSION 47 Fig. 4.22. Predicted temperature (—) of a char particle from beech wood (8 mm) during combustion in the fluidized bed described in Table 3.1, where the ∆Tmax and ∆Tmin curves (−∙−) meet the maximum and minimum oscillation temperature. The zoom shows the timetemperature oscillation during two cycles of movement Fig. 4.23. Predicted and measured combustion temperature oscillation: ∆Tc and Tc respectively, from beech wood (a) and sub-bituminous (b) coal char particles (8 mm) at 11 %v/v (empty symbols, upper diagrams) and 21 %v/v (filled symbols, lower diagrams) O2 concentration in N2 The experimental and predicted temperature oscillations for the combustion of the char particle from beech wood (a) and sub-bituminous coal (b) are shown in Fig. 4.23, testing the model capability to explain and/or predict the experimental results. The agreement is seen from the above experimental results (Fig. 4.19 and Fig. 4.20). The theoretical analysis predicts fast CHAPTER 4 54 which could be due to the estimation of a faster carbon consumption by oxidation at the end of the conversion. Fig. 4.28. Predicted temperature (Tc) of the single char particle conversion (from subbituminous coal, dc,o = 8 mm) in the sand bed described in Table 3.1 with an O2/CO2 fluidization flow at 21%v/v O2 concentration The comparison of some experimental results from literature: maximum combustion temperature (a) and burnout times (b) versus the model’s predictions are in Fig. 4.29. The results predicted are from different bed-particle sizes and temperatures, type of carbon/char particles and sizes, O2 concentrations in N2 and CO2, and fluidization velocities. Moreover, they were measured by different techniques, such as thermocouples, pyrometer with optical probe or digital camera, photographs, and visual analysis. The main experimental conditions of these works are in Table 4.4. The predictions of results from investigations using thermocouples were carried out until the particle reaches the size of 2-3 mm, given that this is the maximum size whose temperature can be recorded using thermocouples (see § 4.2). The present model fits the experimental results quite accurately, improving the predictions given by some authors [30]. The maximum temperature (Fig. 4.29a) seems to be over-predicted since most of the points are close to the +15 % line. This could be a consequence of an overestimation of the carbon consumption by oxidation. It should be pointed out that the measured RESULTS AND DISCUSSION 55 temperatures from the present work, whose predictions show uncertainties greater than 15 %, are from beech-wood char at 50 %v/v O2 concentration. In those tests, the char particle reached the maximum temperature measurable by the method used (1020 ºC, see Fig. 4.2a), and possibly it is higher. The results support the significance of taking into account the movement of the char through the bed, which could be an explanation of the uncertainties found in literature related to the prediction of temperature and burnout times in oxy-combustion [31]. Fig. 4.29. Comparison of the maximum combustion temperatures (a) and burnout times (b) from published experimental work (see Table 3.2) versus model predictions. The lines show ± 15% deviations in the estimation CHAPTER 4 56 Table 4.4. Observations and experimental conditions from works where the char particles conversion in FB are addressed Ref. FB Fuel Conditions studied Techniques Measurement Andrei MA et al., [99] Silica sand, di: 0.2 mm umf: 0.02 m/s hmf: 0.12 m LC O2-N2, [O2]: 2, 5, 8 %v Tbed: 750, 900 °C uf/umf: 3.2 Batch, dc: 1.8, 3 mm Photography Char particle weighing Burn-out time Bu C et al., [29] Silica sand, di: 0.3-0.5 mm umf: 0.19 m/s hmf: 0.15 m AC BC LC SBC WC O2-N2 vs. O2-CO2, [O2]: 10–40 %v Tbed: 810 °C uf/umf: 1.8 dc: 6 mm Thermocouple, 2 x dth: 0.25 mm CCD camera Char particle’s temperature and burnout time Macek A et al., [68] Silica sand di: 1.6 mm umf: 0.8 m/s hmf: 0.9 m LC O2-N2, [O2]: 12.5-23 %v Tbed: 780-880 °C uf/umf: 1.6 Batch, dc: 3.7 mm Two-color pyrometry by optical probe Flue gas analyzer Char particle’s temperature Roy B et al., [26] Quartz sand di: 0.13 mm Umf: 0.05m/s hmf: 0.1 m BrC O2-N2 vs. O2-CO2, O2]: 5–15 %v Tbed: 890 °C uf/umf: 1.4 dc: 8-10 mm Thermocouple, dth: 0.5 mm Flue gas analyzer Char particle’s temperature and consumption rate Scala F et al., [24] Silica sand di: 0.55 mm Umf: 0.13 m/s hmf: 0.1 m BC O2/N2 (air) Tbed: 850 ºC uf/umf: 2.3 dc: 6-7 mm Thermocouple, dth: 0.5 mm Flue gas analyzer Char particle’s temperature and consumption rate Stubintong JF, [35] Silica sand di: 0.46 mm umf: 0.08 m/s hmf: 0.23, 0.27 m CK O2/N2 (air) Tbed: 900 ºC uf/umf:1.7, 1.3 Batch, dc: 2.5-5 mm Photography Thermocouple Char particle’s temperature AC: Antracite Coal, BC: Bituminous Coal, BrC: Brown Coal, CK: Coke, LC: Lignite Coal, SBC: Sub-bituminous Coal, WC: Wood Char 57 CHAPTER 5. CONCLUSIONS Oxy-fuel combustion in a fluidized bed (FB) is an interesting technology with high potential for reducing CO2 emissions. In spite of all international groups working on this field and the number of pilot plants constructed and fundamental researches published, there are still a number of challenges to face before safe design of a 2nd generation FB oxy-boilers. Among these questions are clarifying the impact of changing from O2/N2 (air-firing conditions) to O2/CO2 (oxy-firing conditions) atmosphere on the conversion of coarse char particles. Moreover, the high O2 concentration in the fluidization agent could produce unacceptable temperature levels both globally on the reactor and locally in a fuel particle. In this work, the latter issue was investigated. A 2D FB device equipped with a quartz window at the bottom and freeboard was designed and coupled with a digital camera, allowing optical measurements of single char particle during conversion. Four main tasks were carried out to answer the most relevant questions related with the increase of char temperature under oxy-fuel conversion in 2nd generation oxyboilers (conversion in O2/CO2 gas atmosphere at high O2 concentration):  In the first task a method to measure the char particle's temperatures and sizes by pyrometry with a digital camera is developed, based on the sequential use of the thermal radiation belonging to the red, green and blue spectral bands (colors).  In the second task the impact of using thermocouples on the char particle conversion was addressed. The main driving force for this investigation is to set clear whether the use of a thermocouple(s) attached to the coal/char particle to measure its temperature in a FB during conversion affects or not the accuracy of the measurement. Although it has been assumed up to now that the temperature of the char with an embedded thermocouple is similar to that of a freely fluidized char, it has shown by dedicated measurements that the thermocouple does influence the free movement of the char throughout different bed phases/zones and this considerably affects the rate of char conversion and therefore its temperature history during combustion.  In the third task the effect of the char particle's movement through the bed (emulsion phase, bubble phase, and splash zone) on its combustion temperature was studied. These temperatures are also predicted by a model, which is developed taking into account the different locations at which the char particle stays in the bed during conversion. Comparison of interpretation of model outputs together with experimental measurements allowed a better understanding of the temperature variations. CHAPTER 5 58  In the fourth task both the measurement technique and a theoretical model developed to predict single char particle conversion in a FB under oxy-firing conditions (taking into account heterogeneous reactions of char with O2 and CO2) was sued to investigate the influence of a high O2 concentration in CO2 on the rate of char conversion and temperature profile along the time. 5.1. SUMMARY OF CONTRIBUTIONS FROM THIS WORK 5.1.1. ONE-COLOR PYROMETRY WITH A DIGITAL CAMERA To develop a measurement technique of temperatures by pyrometry with a digital camera, it was carried out combustion tests of single char particles from beech wood inside of an electric oven. The temperature of the oven was set at 800 °C or 850 °C, and the O2 concentration ranged from 0 to 50 %V/V in N2. 1. A method that use the thermal radiation belonging to the visible spectral bands red, green, and blue in a sequential way is proposed (one-color pyrometry, P1C), in contrast with the traditional method based on the simultaneous use of two bands (two-color pyrometry, P2C). The color is selected automatically according to the intensity of the radiation reaching the digital camera, 2. For the optical and electronic configuration of the digital camera used in this work, the red color is used to measure char particle’s temperature from 600 to 800 °C, the green color for the range of 800-925 °C, and the blue color for temperatures from 925 to 1020 °C. 3. P1C allows measuring the char particle’s surface temperature below the background (sand bed and reactor’s wall), overcoming the limitation of the traditional P2C, where only temperatures higher than that (800 and 850 °C, typical in a FB combustion) can be measured. 4. The method allows visualizing further conversion details, such as char particles’ surface temperature gradient and size during combustion. 5. The apparent limitation of P1C, which is that the char’s emissivity has to be known, was shown to be a minor inconvenience since the error caused by a guess between 0.85 and 1 is lower than 1%, provided that the temperature to be measured is higher than that of the bed (background). When the temperature is lower than the bed, the error can be higher and careful determination of the emissivity is required. IMPACT OF USING THERMOCOUPLES ON THE CHAR PARTICLE COMBUSTION IN A FLUIDIZED BED The effect of an embedded thermocouple on the char conversion was addressed by analyzing combustion (O2/N2) tests of single char particles from beech wood, and bituminous and subbituminous coal. These particles were fluidized either with an embedded thermocouple or without it, in a sand bed at 800 °C using air as fluidization agent. 6. The images analysis of the burning char in a 2D FB shows that a freely-fluidized char particle is not always in the emulsion phase, questioning one of the main assumptions in modeling of a bubbling FB. Depending on the density of the char particle and the fluidization velocity, CONCLUSIONS 59 the presence of char particles in the bubble phase ranges between 8 % and 10 %, and light particles (wood char) spend more time on the surface of the bed. 7. The thermocouple resists the drag of the bed, increasing the time that the char particle stays in the bubble phase compared to the case without thermocouple. This increase is stronger with the thickness of the sheath thermocouple: more than 40 %, 60 %, and 100 % for 0.25, 0.5, and 0.75 mm of sheath diameter respectively. 8. The longer time in the bubble phase results in higher combustion temperatures, reducing the char’s burnout time by around 5 % for an embedded thermocouple of 0.25 mm sheath diameter. 9. The char particle’s free rotation is restricted by the thermocouple, and the same parts of their surface always attain higher temperature. This results in a non-homogeneous surface consumption: the reduction of the char particle’s size perpendicular to the direction of the insertion of the thermocouple is more than 10 % faster than the reduction in the parallel direction. This non-homogeneous size reduction of the char particle was also obtained by measurements in a 3D FB combustor. 10. The use of thermocouples is contestable if accurate char temperature’s measurements are aimed, since the temperature and consumption of the char surface are affected. Although the method can be used by keeping these deviations in mind, its effect on the char particle conversion could be higher if more than one thermocouple is used, or a thicker sheath diameter, and/or the O2 concentration around the char particle is different from the one used here (air). 5.1.3. OSCILLATION IN THE CHAR PARTICLE’S TEMPERATURE IN A FLUIDIZED BED The effect of the char particle’s movement through the bed on the combustion temperature is experimentally and theoretically studied by combustion tests with single char particles from beech wood and sub-bituminous coal in the 2D fluidized-bed reactor at a constant temperature (800 °C) and for two O2 concentration (11 and 21%v/v) in N2. 11. The images analysis by pyrometry detects variations in the temperature of the char particle’s surface of tens of degrees as consequence of its movement through the bed. The temperatures oscillations take place in hundredths of second after the char changes its location in the bed. The amplitude of these oscillations increases with O2 concentration, while it is lower for larger char size (external mass transfer controls the char consumption). 12. It is proven that char particle’s temperatures is more stable (differences between temperatures of the char in the same phase of the bed and with the same size) when it stays in the emulsion phase than when it is located in the bubble phase or splash zone. 13. The results explain the scattered temperatures from investigations using pyrometry by optical probes, since that technique measures particles’ temperatures with different sizes that are both in the emulsion phase and in the bubble phase. Moreover, the present work can explain the combustion temperatures measured by thermocouples showing similar temperature oscillations as the ones registered here: tens of degrees, despite their larger time constants. Since this longer time is, as long as the time needed to the char particle to complete one cycle of circulation through the bed, the maximum and minimum CHAPTER 5 60 temperatures reached could be those of the particle at the highest and lowest points of the cycle. 14. Predictions of the experimental results by the combustion model (proposed by Biggs et al. [43] and completed in the present work by experimental characterization of the char particle’s movement) shows the capability of ta FB combustion model to describe the temperature oscillation during combustion measured in this work. 5.1.4. THE INFLUENCE OF CO2 GAS CONCENTRATION ON THE CHAR TEMPERATURE AND CONVERSION DURING OXY-FUEL COMBUSTION IN A FLUIDIZED BED It has been experimentally and theoretically analyzed the CO2 effect on the temperatures and conversion of the char particle from beech wood and sub-bituminous coal in the 2D fluidized-bed reactor with different bed temperature and O2 concentration in N2 and CO2 as fluidization flow. 15. The analysis of the digital images from these tests has shown that the mass transfer from the bulk gas in the bed phases to the char controls the char particle conversion in the O2/CO2 atmosphere. 16. Gasification contributes to char conversion even at a bed temperature of 800 °C provided the O2 concentration is high enough, corresponding to a char temperature higher than 850 °C. This makes the apparent char consumption rate roughly equal in both O2/N2 and O2/CO2 for char temperatures between 850 and 925 °C, and higher in O2/CO2 than in O2/N2 for char temperatures higher than 925 °C 17. The main differences on the carbon consumption and temperatures recorded are from the time during which the char particles is out of the emulsion phase (shorter than 25 % of the fluidization time). 18. Taking into account the char particle’s movement through the bed as an input of the theoretical model of the char particle conversion in fluidized bed reactors results in an important improvement to describe the effect of the O2/CO2 atmosphere on the char particle conversion. 19. The model developed gave good predictions of temperatures and burnout times measured in this work and also compared well with measurements from others sources, giving uncertainties smaller than 15 %. 5.2. LIMITATIONS OF THE RESULTS AND FURTHER RESEARCHES Various aspects of the experimental devices and configuration used in this work, as well as results, deserve discussion, pointing out their weaknesses and how they can be improved by future research. 5.2.1. MEASUREMENTS BY ONE-COLOR PYROMETRY The maximum temperature measured with the new pyrometric technique was 1020 °C, which limited the analysis to operating conditions (mainly the bed temperature and O2 concentration) where the char particle's temperature is below 1020 °C. The actual temperature range measured in this work (600 - 1020 °C) is a function of the set-up's optical and electronic configuration (experimental calibration of the digital device). A filter that attenuates the thermal radiation CONCLUSIONS 61 before reaching the sensor would allow changing this range easily. In that way, keeping the same digital camera's working distance and field of view, the char particle's temperature (radiation) that saturated the pixel (without the filter) will not saturate it (with the filter). Obviously, the lower temperature measured here could not be measured anymore, since that with a filter that temperature (radiation) could not be detected by the sensor. Nevertheless, the char particles' temperature in oxy-combustion is much higher than that of the bed (typically between 800 and 850 °C), and if a new-calibration of the digital camera is carried out, char particle's temperatures higher than 1300 °C could be measured. This is an ongoing work. 5.2.2. 2D FLUIDIZED BED REACTOR One of the main criticism that could be argued is that the bi-dimensional geometry of the FB reactor could impact, to some extent, the conclusions derived in the present work. To clarify this aspect, the results from this work have been compared with those from the literature, and some tests were conducted with 3D lab-scale reactor. Full agreement was obtained and therefore, the conclusions of this works seems to be general to oxy-fuel and air-fuel combustion in fluidized bed reactors. Another aspect to be discussed is that when the char particle is close to the window (made of quartz) the heat exchange is different to the one produced when the char is close to the metal reactor's wall. However, there is a metal box (black surface) that surrounding the reactor, and when the particle is close to the quartz window, this metal surface is located just a few centimeters, so the differences between the heat exchanges are negligible. On the other hand, the bed temperature is controlled by the thermal radiation from this black metal box (heated by the electric oven) to the reactor, so the quartz window in the bottom is at least at the same temperature than the bed. Therefore, the char particle is likely not cooled down when touching that part of the reactor during fluidization. 5.2.3. MODELLING APPROACH FOR SINGLE PARTICLE CONVERSION A shrinking particle model has been used to model the conversion of a single char particle in a FB under oxy-combustion. It could be extended to shrinking core model, where a new mass transfer resistance will be introduced since there is an ash-layer around the char core. Although there are not inconvenienced in the theoretical study, it is important to note that the experimental measurement carried out here is based on the thermal radiation emitted from the char particle's surface, whose difference with respect to the one from the background (bed and reactor's walls) allow seeing the particle. If there is an ash layer on the char's surface, it is possible that these differences between thermal radiations (the one from the particle and the one from the surrounding) involve difficulties measuring the char-ash temperature since could be much difficult to see the particle. 5.2.4. FURTHER RESEARCH It has been pointed out the importance of the time the char particle is in each bed phases on its conversion: temperature oscillation of tens of degrees in just hundredth of second and the lower O2 flux to the char particle's surface when it is in the emulsion phase. Since the conditions of the emulsion phase (close to minimum fluidization velocity) are mainly a function of the CHAPTER 5 62 properties of the bed particles (size, density, shape, etc), it would be interesting to analyze the effect of these parameters on the oxy-fuel conversion. On the other hand, the maximum temperature that can be measured was limited to the O2 concentration in the gas atmosphere (either in N2 and CO2) up to 50 %v7v, when the bed temperature was set at 800 °C, and up to 40 %v/v when it was set at 850 °C. It would be interesting to analyze gas atmospheres with higher O2 concentration and bed temperature since it seems that the char's temperature in oxycombustion is not linearly related with the O2 concentration when the temperature is higher than 950 °C. Moreover, it would be interesting to feed a batch of char particles into the bed in order to analyze if there are differences in their temperatures as a function of height of the bed. Wet recirculation of the flue gases from the exit involves H2O concentration inside the boiler. It is expected that the increase in water vapor in the gas atmosphere increases the carbon consumption rate by gasification and reduce the char particle's temperature. However, the H2 produce could be oxidized close to the char's surface an increase its temperature. Therefore, it will be interesting to study the conversion when the fluidization flow is a mixture of O2, CO2, and H2O. 63 AGRADECIMIENTOS / ACKNOWLEDGEMENTS Como personas individuales vivimos todos juntos, y por lo tanto, ni estamos, ni hacemos las cosas solos. Este trabajo no fue diferente, y durante su realización he contado con la ayuda, apoyo y consejo de familiares y amigos. Por ello, quiero dar mi más sincero agradecimiento a los que forman mis circunstancias y me acompañan siempre. De manera explícita, quiero agradecer a Alberto Gómez Barea el haberme dado la oportunidad de investigar y de guiarme en el camino; el haberme permitido descubrir la utilidad de trabajar en una idea, aunque ésta fuese equivocada; y mostrarme que la satisfacción está en el trabajo que se realiza y no al final de él. Agradecer a Bo Leckner, un genio con una asombrosa capacidad de escuchar, sus consejos, propuestas y correcciones, y el haberme mostrado como mejorar un trabajo manteniendo la personalidad del estudiante. A David Fernández su afecto puro y desinteresado, prestándome su ayuda siempre que lo he necesitado. A mis compañeros en esta “aventura doctoral”, en especial a María Tripiana y Cristina López su disposición a discutir y escuchar cualquier idea. A mis compañeros “experimentales” Ana Berdugo, Eva Sánchez, y Manuel Luna su ayuda y paciencia en la preparación de las pruebas, y sobre todo su compromiso durante la realización de las mismas. A Israel Pardo su ayuda en la logística y control del material del laboratorio, si la cual ningún ensayo hubiese sido posible. También quisiera agradecer a Pedro Haro su ayuda, consejos profesionales y académicos, y sus sugerencias de socialización. Agradecer a todos los miembros del departamento de Ingeniería Química y Ambiental su colaboración y apoyo, así como a la Universidad de Sevilla su labor en nuestra formación académica y personal. Igualmente quisiera agradecer a mis padres Jesús y Conchi, a quienes dedico este trabajo, el amor y esfuerzo que me han dedicado durante toda mi vida; y sobre todo el haberme transmitido el coraje necesario para afrontar cualquier reto. A mis hermanos Inma y Adrián su alegría, paciencia y compañía de todos los días. En último lugar quisiera agradecer a Leonor el ser el viento a favor en mi vida, sin el cual no hubiese podido realizar este viaje. NOMENCLATURE 70 SUBSCRIPTS ⊥ Perpendicular direction regarding the insertion direction of the thermocouple into the char particle // Parallel direction regarding the insertion direction of the thermocouple into the char particle ABBREVIATIONS bPh Bubble Phase B Blue spectral band CCS CO2 Capture and Store ePh Emulsion Phase fps Frame per second FBC 2D Experimental set-up with a rectangular fluidized bed reactor (18 x 1.8 x 50 cm) FBC 3D Experimental set-up with a cylindrical fluidized bed reactor (7 (i.d.) x 50 cm) FOV Field of view of the digital camera G Green spectral band P1C One-color pyrometry P2C Two-color pyrometry R Red spectral band sPh Splash zone w th Char particle fluidized with an embedded thermocouple (restricted fluidization) w/o th Char particle fluidized without an embedded thermocouple (free fluidization) WD Working distance of the digital camera 71 REFERENCES [1] BP Energy Outlook 2017 – BP Global, 2017 Edition [2] Nejat P, Jomehzadeh F, Mahdi Taheri M, Gohari M, Zaimi M, Majid A. A global review of energy consumption, CO2 emissions and policy in the residential sector (with an overview of the top ten CO2 emitting countries). Renewable and Sustainable Energy Reviews 2015; 43; 843–862 [3] Buhre BJP, Elliott LK, Sheng CD, Gupta RP, Wall TF. Oxy-fuel combustion technology for coal-fired power generation. Progress in Energy and Combustion Science 2005; 31; 283-307 [4] Rubin ES, Mantripragada H, Marks A, Versteeg P, Kitchin J. The outlook for improved carbon capture technology. Progress in Energy and Combustion Science 2012; 38; 630-671 [5] Davidson J. Performance and costs of power plants with capture and storage of CO2. Energy and Combustion Science 2007; 32; 1163-1176 [6] Saastamoinen J, Tourunen A, Pikkarainen T, Hasa H, Miettinen J, Hyppanen T, Myohanen K. Fluidized bed combustion in high concentrations of O2 and CO2. Proceeding of the 19th International Conference on Fluidized Bed Combustion 2006, No 49 [7] Scheffknecht G, Al-Makhadmeh L, Schnell U,Maier J. Oxy-fuel coal combustion-A review of the current state-of-the-art. International Journal of Greenhouse Gas Control 2011; 5; 16-35 [8] Toftegaard MB, Brix J, Jensen PA, Glarborg P, Jensen AD. Oxy-fuel combustion of solid fuels, Progress in Energy and Combustion Science 2010; 36; 581-625 [9] Lupion M, Alvarez I, Otero P, Kuivalainen R, Lantto J, Hotta A, Hack H. 30 MWth CIUDEN Oxy-cfb Boiler - First Experiences. Energy Procedia 2013; 37; 6179-6188 [10] Czakiert T, Zuwala J, Lasek J. Oxy-fuel combustion: The state of the art. Proceedings of 12th International Conference on Fluidized Bed Technology, Kraków (Poland) 2017; 1: 35– 46 (ISBN: 978-83-62079-16-2) REFERENCES 72 [11] Jäntti T, Parkkonen R. Lagisza 460 MWe Supercritical CFB-Experience During First Year After Start of Commercial Operation. Foster Wheeler Energia Oy Finland 2010; Presented at Russia Power at 24-26 March 2010 (Moscow, Russia) [12] Leckner B, Gómez-Barea A. Oxy-fuel Combustion in Circulating Fluidized Bed boilers. Applied Energy 2014; 125; 308-318 [13] Günther C, Weng M, Kather A. Restrictions and Limitations for the Design of a Steam Generator for a Coal-fired Oxyfuel Power Plant with Circulating Fluidised Bed Combustion. Energy Procedia 2013; 37; 1312-1321 [14] Riaza J, Khatami R, Levendis YA, Álvarez L, Gil MV, Pevida C, Rubiera F, Pis JJ. Single particle ignition and combustion of anthracite, semi-anthracite and bituminous coals in air and simulated oxy-fuel conditions. Combustion and Flame 2014; 161; 1096-1108 [15] Singer S, Chen L, Ghoniem AF. The influence of gasification reactions on char consumption under oxy-combustion conditions: Effects of particle trajectory and conversion. Proceedings of the Combustion Institute 2013; 34; 3471-3478 [16] Marek E, Świątkowski B. Experimental studies of single particle combustion in air and different oxy-fuel atmospheres. Applied Thermal Engineering 2014; 66; 35-42 [17] Hecht ES, Shaddix CR, Molina A, Haynes BS. Effect of CO2 gasification reaction on oxycombustion of pulverized coal char. Proceedings of the Combustion Institute 2011; 33; 1699- 1706 [18] Gonzalo-Tirado C, Jiménez S, Ballester J. Gasification of a pulverized sub-bituminous coal in CO2 at atmospheric pressure in an entrained flow reactor. Combustion and Flame 2012; 159; 385-395 [19] Kim D, Choi S, Shaddix CR, Geier M, Effect of CO2 gasification reaction on char particle combustion in oxy-fuel conditions.Fuel 2014; 120; 130-140 [20] Jia L, Tan Y, Wang C, Anthony EJ. Experimental study of oxy-fuel combustion and sulfur capture in a mini-CFBC. Energy Fuels 2007; 21; 3160– 3164 [21] Czakiert T, Bis Z, Muskala W, Nowak W. Fuel conversion from oxy-fuel combustion in a circulating fluidized bed. Fuel Processing Technology 2006; 87; 531–538 [22] Scala F, Chirone R, Fluidized bed combustion of single coal char particles at high CO2 concentration, Chemical Engineering Journal 2010;165:902-906 [23] Scala F, Chirone R. Combusion of single coal char particles under fluidized bed oxyfiring conditions. Proccedings of the 20th Inernational Conference on Fluidized Bed Combustion 2009; 624-629 [24] Scala F, Chirone R. Combusion of single coal char particles under fluidized bed oxyfiring conditions. Industrial & Engineering Chemical research 2010; 49; 11029-11036 [25] Scala F, Chirone R. Combustion of single coal char particle under fluidized bed oxy-firing conditions. 63rd IEA-FBC Meeting, Ponferrada (Spain), 29th to 30thNovember, 2011 73 [26] Roy B, Bhattacharya S. Combustion of single char particles from Victorian brown coal under oxy-fuel fluidized bed conditions, Fuel 2016;165:477-483 [27] Bu C, Liu D, Chen X, Pallarès D, Gómez-Barea A. Ignition behavior of single coal particle in a fluidized bed under O2/CO2 and O2/N2 atmospheres: A combination of visual image and particle temperature. Apply Energy. 2014; 115: 301-8 [28] Bu C, Leckner B, Chen X, Pallarès D, Liu D, Gómez-Barea A, Devolatilization of a single fuel particle in a fluidized bed under oxy-combustion conditions. Part A: Experimental results. Combustion and Flame 2015; 162; 797-808 [29] Bu C, Pallarès D, Chen X, Gómez-Barea A, Liu D, Leckner B, Lu P. Oxy-fuel combustion of a single fuel particle in a fluidized bed: Char combustion characteristics, an experimental study. Chemical Engineering Journal 2016; 287; 649-656 [30] Bu C, Gómez-Barea A, Chen X, Leckner B, Liu D, Pallarès D, Lu P. Effect of CO2 on oxy-fuel combustion of coal-char particles in a fluidized bed: Modeling and comparison with the conventional mode of combustion. Applied Energy 2016; 177: 247-259 [31] Mathekga HI, Oboirien BO, North BC. A review of oxy-fuel combustion in fluidized bed reactors. International Journal of Energy Research 2016; 40: 878-902 [32] Linjewile TM, Hull AS, Agarwal PK. Optical probe measurements of the temperature of burning particles in fluidized bed. Fuel 1994; 73: 1880-1888 [33] Agarwal PK. The residence phase of active particles in fluidized beds of smaller inert particles. Chemical Engineering Science 1987; 42; 2481-2483 [34] Dos Santos FJ, Goldstein Jr L. Experimental aspects of biomass fuels in a bubbling fluidized bed combustor. Chemical Engineering and Processing: Process Intensification 2008; 47; 1541-1549 [35] Stubington JF. Comparison of techniques for measuring the temperature of char particles burning in a fluidised bed. Chemical engineering research & design 1985; 4; 241-249 [36] Manovic V, Komatina M, Oka S. Modeling the temperature in coal char particle during fluidized bed combustion. Fuel 2008; 87; Pages 905-914 [37] Selcuk N, Ozkan M. Simulation of circulating fluidized bed combustors firing indigenous lignite. International Journal of Thermal Sciences 2011; 50; 1109-1115 [38] Ozkan M. Mathematical modeling of circulating fluidized bed combustors, M. Sc. Thesis. Middle East Technical University, Ankara (Turkey) [39] Agarwal PK, Mitchell WJ, La Nauze RD. Mass transfer processes around burning char particles in fluidized beds. International Symposium on Combustion 1989; 22; 279-286 [40] Agarwal PK, Mitchell WJ, La Nauze RD. Transport phenomena in multi-particle systems— III. Active particle mass transfer in fluidized beds of inert particles. Chemical Engineering Science 1988; 43; 2511-2521 REFERENCES 74 [41] Agarwal PK. Transport phenomena in multi-particle systems—IV. Heat transfer to a large freely moving particle in gas fluidized bed of smaller particles. Chemical Engineering Science 1991; 46; 1115-1127 [42] Hayhurst AN, Parmar MS. Does solid carbon burn in oxygen to give the gaseous intermediate CO or produce CO2 directly? Some experiments in a hot bed of sand fluidized by air. Chemical Engineering Science 1998; 53; 427-438 [43] Biggs MJ, Agarwal PK. Mathematical modelling of oscillations in the temperature of freely moving burning carbonaceous particles in bubbling fluidized beds. Fuel 1993; 72; 805-811 [44] Broughton J, Howard JR. Combustion of coal in fluidized beds. In: Howard JR, Elliott D editors. Fluidized beds: Combustion and applications, Applied Science Publishers; 1983. p. 37-77 [45] Chirone R, Salatino P, Scala F. The relevance of attrition to the fate of ashes during fluidized-bed combustion of a biomass. Proceeding of the Combustion Institute 2000; 2279-2286 [46] Basu P. Burning rate of carbon in fluidized beds. Fuel 1977; 56: 390-392 [47] Komatina M, Manovic V, Dakic D. An Experimental Study of Temperature of Burning Coal Particle in Fluidized Bed. Energy Fuels 2006; 20: 114–119 [48] Winter F, Prah ME, Hofbauer H. Temperatures in a fuel particle burning in a fluidized bed: The effect of drying, devolatilization, and char combustion. Combustion and Flame 1997; 108: 302-314 [49] Collier AP, Hayhurst AN, Richardson JL, Scott SA. The heat transfer coefficient between a particle and a bed (packed or fluidised) of much larger particles. Chemical Engineering Science 2004; 59: 4613-4620 [50] Parmar MS, Hayhurst AN. The heat transfer coefficient for a freely moving sphere in a bubbling fluidised bed. Chemical Engineering Science 2002; 57: 3485-3494 [51] Hayhurst AN, Parmar MS. Measurement of the mass transfer coefficient and sherwood number for carbon spheres burning in a bubbling fluidized bed. Combustion and Flame 2002; 130: 361-375 [52] Linjewile TM, Gururajan VS, Agarwal PK. The CO/CO2 product ratio from the combustion of single petroleum coke spheres in an incipiently fluidized bed. Chemical Engineering Science 1995; 50: 1881-1888 [53] Scala F. Fluidized-Bed Combustion of Single Coal Char Particles: An Analysis of the Burning Rate and of the Primary CO/CO2 Ratio. Energy Fuels 2011; 25: 1051-1059 [54] Tomé N, Rangel N, Pincho C. Temperature of Wood Char Particles Burning in a Fluidized Bed Reactor. International Conference on Renewable Energies and Power Quality (ICREPQ’12), Santiago de Compostela (Spain), 28th to 30th March, 2012 75 [55] Prins W, Mass and heat transfer between a fluidized bed and a freely moving submerged sphere, Chapter 7.2 in Coal Science and Technology 22, Ed M Valk, Elsevier, Amsterdam (1995) [56] Hernberg R, Stenberg J. Simultaneous in situ measurement of temperature and size of burning char particles in a fluidized bed furnace by means of fiber optic pyrometry, Combustion and Flame 1993; 95; 191– 205 [57] Heino P, Hernberg R, Stenberg J. Statistical pyrometric sizing of particles in fluidised bed combustion, Combustion and Flame 1997; 108; 315-326 [58] Joutsenoja T, Heino P, Hernberg R, Bonn B. Pyrometric temperature and size measurements of burning coal particles in a fluidized bed combustor. Combustion and Flame 1999; 118; 707–717 [59] Li D. Thermal image analysis using calibrated video imaging. Doctoral Thesis, University of Missouri-Columbia, Colombia, 2006 [60] Lu H, Leong-Teng I, Mackrory A, Werrett L, Scott J, Tree D, Baxter L. Particle surface temperature measurements with multicolor band pyrometry, AIChE Journal 2009; 55; 243– 255 [61] Zukowski W, Baron J, Bulewicz EM, Kowarska B. An optical method of measuring the temperature in a fluidized bed combustor. Combustion and flame 2009; 156; 1445-1452 [62] Brown SL, Lattimer BY. Transient gas-to-particle heat transfer measurements in a spouted bed. Experimental Thermal and Fluid Science, 2013; 44; 883-892 [63] Tsuji T, Miyauchi T, Oh S, Tanaka T. Simultaneous measurement of particle motion and temperature in two-dimensional fluidized bed with heat transfer, KONA Powder and Particle Journal 2010; 28; 167-179 [64] Patil AV, Peters EAJF, Sutkar VS, Deen NG, Kuipers JAM. A study of heat transfer in fluidized beds using an integrated DIA/PIV/IR technique. Chemical Engineering Journal 2015; 259; 90-106 [65] Khatami R, Levendis YA. On the deduction of single coal particle combustion temperature from three-color optical pyrometry. Combustion and Flame 2011; 158; 1822-1836 [66] Grosshandler WL. The effect of soot on pyrometric measurements of coal particle temperature. Combustion and Flame 1984; 55; 59-71 [67] Godoy SM, Lockwood FC. Development of a two-colour infrared pyrometer for coal particle temperature measurements during devolatilisation. Fuel 1997; 77; 995–999 [68] Macêk A, Bulik C, Direct measurement of char-particle temperatures in fluidized bed combustors, In Symposium (International) on Combustion. Elsevier 20.1 (1985) 1223-1230 [69] Salinero J, Goméz-Barea A, Leckner B, Tripiana M, Bu B, Gómez-González E. Improving char temperature measurement by pyrometry with digital camera. Proceedings of 22nd International Conference on Fluidized Bed Conversion Turku (Finland), 2 (2015) 375-383 REFERENCES 76 [70] Gómez-Barea A, Leckner B. Modeling of biomass gasification in fluidized bed. In Progress in Energy and Combustion Science 2010; 36; 444-509 [71] Arthur JR. Reactions between carbon and oxygen. Transaction of the Faraday society 1951;47; 164–78 [72] Förtsch D, Schnell U, Hein KRG, Essenhigh RH. The mass transfer coefficient for the combustion of pulverized carbon particles, Combustion and Flame 2001; 126;1662-1668 [73] Hayhurst AN. The mass transfer coefficient for oxygen reacting with a carbon particle in a fluidized or packed bed. Combustion and Flame 2000; 121; 679-688 [74] Scala F. Calculation of the mass transfer coefficient for the combustion of a carbon particle, Combustion and Flame 2010; 157; 137-142 [75] Sadhukhan AK, Gupta P, Kumar SR.Modeling and experimental studies on combustion characteristics of porous coal char: Volume reaction model. International Journal of Chemical Kinetics 2010; 42; 299–315 [76] Saucedo MA, Butel M, Scott SA, Collings N, Dennis JS. Significance of gasification during oxy-fuel combustion of a lignite char in a fluidised bed using a fast UEGO sensor. Fuel 2015; 144; 423-438 [77] Adánez J, de Diego LF, García-Labiano F,Abad A, Abanades JC. Determination of Biomass Char Combustion Reactivities for FBC Applications by a Combined Method. Industrial & Engineering Chemistry Research 2001; 40; 4317-4323 [78] Evans DD, Emmons HW. Combustion of wood charcoal. Fire Safety Journal 1977; 1; 57-66 [79] Hobbs ML, Radulovic PT, Smoot LD. Modeling fixed-bed coal gasifiers. AIChE Journal 1992; 38; 681–702 [80] Haseli Y, van Oijen JA, de Goey LPH. A detailed one-dimensional model of combustion of a woody biomass particle. Bioresource Technology 2011; 102; 9772-9782 [81] Hong L H, Warren R, Gregory P, Bryan R, Larry L. B. Comprehensive Study of Biomass Particle Combustion. Energy Fuels 2008;22; 2826–2839 [82] Prins W, Siemons R, Van Swaaij WPM, Radovanovic M. Devolatilization and ignition of coal particles in a two-dimensional fluidized bed. Combustion and flame 1989; 75; 57-79 [83] Prins, W, Casteleijn, TP, Draijer W, Van Swaaij WPN. Mass transfer from a freely moving single sphere to the dense phase of a gas fluidized bed of inert particles. Chemical Engineering Science 1985; 40; 481–497 [84] Palchonok G. Heat and mass transfer to a single particle in fluidized bed, Dr Thesis, Chalmers University of Technology, 1998. [85] Cobbinah S, Laguérie C, Gibert H. Simultaneous heat and mass transfer between a fluidized bed of fine particles and immersed coarse porous particles. International Journal of Heat and Mass Transfer 1987; 30; 395-400 77 [86] Scala F. Mass transfer around freely moving active particles in the dense phase of a gas fluidized bed of inert particles. Chemical Engineering Science 2007; 62; 4159-4176 [87] Baskakov AP, Filippovsky NF, Munts VA, Ashichmin AA, Temperature of particles heated in a bed of inert material. Journal Engineering Physics 1987; 52; 574-578 [88] La Nauze RD, Jung K. The kinetics of combustion of petroleum coke particles in a fluidizedbed combustor, International Symposium on Combustion 1982; 19; 1087-1092 [89] Gómez González E. Guía para conocer las cámaras de fotografía digital. ed. Victoria 2012, pp 503 [90] Baskakov AP, Leckner B. Radiative heat transfer in circulating fluidized bed furnace. Powder Technology 1997; 90; 213-218 [91] Planck M. Distribution of energy in the Spectrum, Annals of Physics 1901; 4; 553-563 [92] Incropera FP, De Witt DP. Radiation: Properties and processes. In: Frank P. Incropera and David P. De Witt, Fundamentals of heat and mass. Prentice hall Ed., México, 1999, pp. 633-716. [93] Roscoe JC, Witkowski AR, Harrison D. The temperature of coke particles in a fluidized combustor. Transactions of the Institution of Chemical Engineers 1980; 58; 69-72. [94] Ross IB, Patel MS, Davidson JF. The temperatures of burning carbon particles in fluidized beds. Transactions of the Institution of Chemical Engineers 1981; 59; 83-88. [95] La Nauze RD, Jung K, Kastl J. Mass transfer to large particles in fluidised beds of smaller particles. Chemical Engineering Science 1984; 39; 1623-1633 [96] Davidson JF, Harrison D. Fluidized Particles. Cambridge University Press, Cambridge, 1963 [97] Remiarová B, Markoš J, Žajdlık R, Jelemenský L, Identification of the mechanism of coal char particle combustion by porous structure characterization. Fuel Processing Technology 2004; 85; 303-321 [98] Scala F. A new technique for the measurement of the product CO/CO2 ratio at the surface of char particles burning in a fluidized bed, In Proceedings of the Combustion Institute 2009, 32; 2021-2027 [99] Andrei MA, Sarofim AF, Beér JM. Time-resolved burnout of coal particles in a fluidized bed, Combustion and Flame 1985; 61;17-27 [100] Chen C, Kojima T. Single char particle combustion at moderate temperature: effects of ash. Fuel Processing Technology 1996; 47; 215-232. [101] Massman WJ. A review of the molecular diffusivities of HO, CO, CH, CO, O, SO, NH, NO, NO, and NO in air, O and N near STP. Atmospheric Environment 1998; 32; 1111-1127 [102] Fatehi H, Xue-Song Bai, Structural evolution of biomass char and its effect on the gasification rate, Applied Energy 2017; 185; 998-1006 [103] Klein SA. “Engineering Equation Solver Software (EES)”, Academia Commercial version 10.091-3D (09-05-2017) 79 APPENDIX I. EXPERIMENTAL CHARACTERIZATION OF THE CHAR PARTICLE’S MOVEMENT THROUGH A FLUIDIZED BED Taking into account the description of the movement of a char particle through a fluidized bed made elsewhere [43] and shown in Fig. 2.2, and the relation between the char residence time in each bed phase tj,T and the time spent to complete a cycle T, by θj,T=tj,T T (AI.1) The fraction of time in the splash zone during a period can be expressed by θs=T−tin−bed T (AI.2) where the time during which the particle is inside the bed (emulsion and bubble phases) is tin−bed=hp(1 ud+1 ur) (AI.3) Therefore, the period can be expressed by T= hp 1−θs(1 ud+1 ur) (AI.4) as a function of the upward and downward char particle’s velocities (ur and ud) and the depth reached by a particle in a cycle (hp). Looking at Fig. 2.2, the upward velocity of the char particle through the bed can be expressed by ur=hp m(te+tb) (AI.5) where m is the times that the char is reached by a bubble. The depth reached by the char is expressed by hp=m(tbub−teue) (AI.6) Eqs. AI. 5-6 allow getting the time during which the char particle is inside a bubble as APPENDIX III 86 temperature related to the surface region, whose radiation is measured, is substituted by digital numbers. In this way, the fitting curves store the relation between the calibration parameter and the digital numbers in the temperature range of interest. ln(βi/j)=C2 T(1 λj−1 λi)+5∗ln(λj λi) − ln(DNi DNj) (AIII.17) ln(β k) = ln(Φk εc C1 λk 5 exp(− C2 λk T)+(1−εc) E(∆λk,Tw)) − ln(DNk) (AIII.18)