ANALYSIS AND VISUALIZATION OF FLOW FIELDS USING INFORMATION-THEORETIC TECHNIQUES AND GRAPH-BASED REPRESENTATIONS

Three-dimensional flow visualization plays an essential role in many areas of science and engineering, such as aero- and hydro-dynamical systems which dominate various physical and natural phenomena. For popular methods such as the streamline visualization to be effective, they should capture the underlying flow features while facilitating user observation and understanding of the flow field in a clear manner. My research mainly focuses on the analysis and visualization of flow fields using various techniques, e.g. information-theoretic techniques and graph-based representations. Since the streamline visualization is a popular technique in flow field visualization, how to select good streamlines to capture flow patterns and how to pick good viewpoints to observe flow fields become critical. We treat streamline selection and viewpoint selection as symmetric problems and solve them simultaneously using the dual information channel [81]. To the best of my knowledge, this is the first attempt in flow visualization to combine these two selection problems in a unified approach. This work selects streamline in a view-independent manner and the selected streamlines will not change for all viewpoints. My another work [56] uses an information-theoretic approach to evaluate the importance of each streamline under various sample viewpoints and presents a solution for view-dependent streamline selection that guarantees coherent streamline update when the view changes gradually. When projecting 3D streamlines to 2D images for viewing, occlusion and clutter become inevitable. To address this challenge, we design FlowGraph [57, 58], a novel compound graph representation that organizes field line clusters and spatiotemporal regions hierarchically for occlusion-free and controllable visual exploration. We enable observation and exploration of the relationships among field line clusters, spatiotemporal regions and their interconnection in the transformed space. Most viewpoint selection methods only consider the external viewpoints outside of the flow field. This will not convey a clear observation when the flow field is clutter on the boundary side. Therefore, we propose a new way to explore flow fields by selecting several internal viewpoints around the flow features inside of the flow field and then generating a B-Spline curve path traversing these viewpoints to provide users with closeup views of the flow field for detailed observation of hidden or occluded internal flow features [54]. This work is also extended to deal with unsteady flow fields. Besides flow field visualization, some other topics relevant to visualization also attract my attention. In iGraph [31], we leverage a distributed system along with a tiled display wall to provide users with high-resolution visual analytics of big image and text collections in real time. Developing pedagogical visualization tools forms my other research focus. Since most cryptography algorithms use sophisticated mathematics, it is difficult for beginners to understand both what the algorithm does and how the algorithm does that. Therefore, we develop a set of visualization tools to provide users with an intuitive way to learn and understand these algorithms.

Flow-Centric, Back-In-Time Debugging

Conventional debugging tools present developers with means to explore the run-time context in which an error has occurred. In many cases this is enough to help the developer discover the faulty source code and correct it. However, rather often errors occur due to code that has executed in the past, leaving certain objects in an inconsistent state. The actual run-time error only occurs when these inconsistent objects are used later in the program. So-called back-in-time debuggers help developers step back through earlier states of the program and explore execution contexts not available to conventional debuggers. Nevertheless, even back-in-time debuggers do not help answer the question, ``Where did this object come from?'' The Object-Flow Virtual Machine, which we have proposed in previous work, tracks the flow of objects to answer precisely such questions, but this VM does not provide dedicated debugging support to explore faulty programs. In this paper we present a novel debugger, called Compass, to navigate between conventional run-time stack-oriented control flow views and object flows. Compass enables a developer to effectively navigate from an object contributing to an error back-in-time through all the code that has touched the object. We present the design and implementation of Compass, and we demonstrate how flow-centric, back-in-time debugging can be used to effectively locate the source of hard-to-find bugs.

Efecto de las cubiertas ajardinadas sobre el microclima urbano de verano

El cambio climático y sus efectos requieren con urgencia el desarrollo de estrategias capaces no solo de mitigar pero también permitir la adaptación de los sistemas afectados por este fenómeno a los cambios que están provocando a nivel mundial. Olas de calor más largas y frecuentes, inundaciones, y graves sequías aumentan la vulnerabilidad de la población, especialmente en asentamientos urbanos. Este fenómeno y sus soluciones potenciales han sido ampliamente estudiados en las últimas décadas desde diferentes perspectivas y escalas que analizan desde el fenómeno regional de isla de calor al aumento de la intensidad energética necesaria en los edificios para mantener las condiciones de confort en los escenarios de calentamiento que se predicen. Su comprensión requiere el entendimiento de este fenómeno y un profundo análisis de las estrategias que pueden corregirlo y adaptarse a él. En la búsqueda de soluciones a este problema, las estrategias que incorporan sistemas naturales tales como las cubiertas ajardinadas, las fachadas vegetadas y bosques urbanos, se presentan como opciones de diseño capaces de proporcionan múltiples servicios al ecosistema urbano y de regular y hacer frente a los efectos del cambio climático. Entre los servicios que aportan estos sistemas naturales se incluyen la gestión de agua de tormentas, el control del efecto isla de calor, la mejora de la calidad del aire y del agua, el aumento de la diversidad, y como consecuencia de todo lo anterior, la reducción de la huella ecológica de las ciudades. En la última década, se han desarrollado múltiples estudios para evaluar y cuantificar los servicios al ecosistema proporcionados por las infraestructuras verdes, y específicamente las cubiertas ajardinadas, sin embargo, determinados servicios como la capacidad de la regulación del microclima urbano no ha sido apenas estudiados. La mayor parte de la literatura en este campo la componen estudios relacionados con la capacidad de las cubiertas ajardinadas de reducir el efecto de la isla de calor, en una escala local, o acerca de la reducción de la demanda energética de refrigeración debida a la instalación de cubiertas ajardinadas en la escala de edificio. La escala intermedia entre estos dos ámbitos, la calle, desde su ámbito habitable cercano al suelo hasta el límite superior del cañón urbano que configura, no han sido objeto detallado de estudio por lo que es esta escala el objeto de esta tesis doctoral. Esta investigación tiene como objeto contribuir en este campo y aportar un mayor entendimiento a través de la cuantificación del impacto de las cubiertas ajardinadas sobre la temperatura y humedad en el cañón urbano en la escala de calle y con un especial foco en el nivel peatonal. El primer paso de esta investigación ha sido la definición del objeto de estudio a través del análisis y revisión de trabajos tanto teóricos como empíricos que investigan los efectos de cubiertas ajardinadas en el entorno construido, entendidas como una herramienta para la adaptación y mitigación del impacto del cambio climático en las ciudades. La literatura analizada, revela el gran potencial de los sistemas vegetales como herramientas para el diseño pasivo puesto que no solo son capaces de mejorar las condiciones climáticas y microclimaticas en las ciudades reduciendo su demanda energética, sino también la necesidad de mayor análisis en la escala de calle donde confluyen el clima, las superficies urbanas y materiales y vegetación. Este análisis requiere una metodología donde se integren la respuesta térmica de edificios, las variaciones en los patrones de viento y radiación, y la interacción con la vegetación, por lo que un análisis cuantitativo puede ayudar a definir las estrategias más efectivas para lograr espacios urbanos más habitables. En este contexto, el objetivo principal de esta investigación ha sido la evaluación cuantitativa del impacto de la cubierta ajardinada en el microclima urbano a escala de barrio en condiciones de verano en los climas mediterráneos continentales. Para el logro de este objetivo, se ha seguido un proceso que persigue identificar los modelos y herramientas de cálculo capaces de capturar el efecto de la cubierta ajardinada sobre el microclima, identificar los parámetros que potencian o limitan este efecto, y cuantificar las variaciones que microclima creado en el cañón urbano produce en el consumo de energía de los edificios que rodean éste espacio. La hipótesis principal detrás de esta investigación y donde los objetivos anteriores se basan es el siguiente: "una cubierta ajardinada instalada en edificios de mediana altura favorece el establecimiento de microclimas a nivel peatonal y reduce las temperaturas en el entorno urbano donde se encuentra”. Con el fin de verificar la hipótesis anterior y alcanzar los objetivos propuestos se ha seguido la siguiente metodología: • definición del alcance y limitaciones del análisis • Selección de las herramientas y modelos de análisis • análisis teórico de los parámetros que afectan el efecto de las cubiertas ajardinadas • análisis experimental; • modelización energética • conclusiones y futuras líneas de trabajo Dada la complejidad de los fenómenos que intervienen en la generación de unas determinadas condiciones microclimáticas, se ha limitado el objeto de este estudio a las variables de temperatura y humedad, y sólo se han tenido en cuenta los componentes bióticos y abióticos del sistema, que incluyen la morfología, características superficiales del entorno estudiado, así como los elementos vegetales. Los componentes antrópicos no se han incluido en este análisis. La búsqueda de herramientas adecuadas para cumplir con los objetivos de este análisis ha concluido en la selección de ENVI-met v4 como el software más adecuado para esta investigación por su capacidad para representar los complejos fenómenos que caracterizan el microclima en cañones urbanos, en una escala temporal diaria y con unas escala local de vecindario. Esta herramienta supera el desafío que plantean los requisitos informáticos de un cálculo completo basado en elementos finitos realizados a través de herramientas de dinámica de fluidos computacional (CFD) que requieren una capacidad de cálculo computacional y tiempo privativos y en una escala dimensional y temporal limitada a esta capacidad computacional lo que no responde a los objetivos de esta investigación. ENVI-met 4 se basa es un modelo tridimensional del micro clima diseñado para simular las interacciones superficie-planta-aire en entornos urbanos. Basado en las ecuaciones fundamentales del equilibrio que representan, la conservación de masa, energía y momento. ENVI-met es un software predictivo, y como primer paso ha requerido la definición de las condiciones iniciales de contorno que se utilizan como punto de partida por el software para generar su propio perfil de temperatura y humedad diaria basada en la localización de la construcción, geometría, vegetación y las superficies de características físicas del entorno. La geometría de base utilizada para este primer análisis se ha basado en una estructura típica en cuanto al trazado urbano situada en Madrid que se ha simulado con una cubierta tradicional y una cubierta ajardinada en sus edificios. La estructura urbana seleccionada para este análisis comparativo es una red ortogonal con las calles principales orientadas este-oeste. El edificio típico que compone el vecindario se ha definido como “business as usual” (BAU) y se ha definido con una cubierta de baldosa de hormigón estándar, con un albedo 0.3, paredes con albedo 0.2 (construcción de muro de ladrillo típico) y cerramientos adiabáticos para evitar las posibles interferencias causadas por el intercambio térmico con el ambiente interior del edificio en los resultados del análisis. Para el caso de la cubierta ajardinada, se mantiene la misma geometría y características del edificio con excepción de la cobertura superficial de la azotea. Las baldosas de hormigón se han modificado con una cubierta ajardinada extensiva cubierta con plantas xerófilas, típicas en el clima de Madrid y caracterizado por su índice de densidad foliar, el “leaf area density” (LAD), que es la superficie total de superficie de hojas por unidad de volumen (m2/m3). El análisis se centra en los cañones urbanos entendidos como el espacio de calle comprendido entre los límites geométricos de la calle, verticales y horizontales, y el nivel superior de la cota urbana nivel de cubiertas. Los escenarios analizados se basan en la variación de la los principales parámetros que según la literatura analizada condicionan las variaciones microclimáticas en el ámbito urbano afectado por la vegetación, la velocidad del viento y el LAD de la azotea. Los resultados han sido registrados bajo condiciones de exposición solar diferentes. Las simulaciones fueron realizadas por los patrones de viento típico de verano, que para Madrid se caracterizan por vientos de componente suroeste que van desde 3 a 0 m/s. las simulaciones fueron realizadas para unas condiciones climáticas de referencia de 3, 2, 1 y 0 m/s a nivel superior del cañón urbano, como condición de contorno para el análisis. Los resultados calculados a 1,4 metros por encima del nivel del suelo, en el espacio habitado, mostraron que el efecto de la cubierta ajardinada era menor en condiciones de contorno con velocidades de viento más altas aunque en ningún caso el efecto de la cubierta verde sobre la temperatura del aire superó reducciones de temperatura de aire superiores a 1 º C. La humedad relativa no presentó variaciones significativas al comparar los diferentes escenarios. Las simulaciones realizadas para vientos con velocidad baja, entre 0 y 1 m/s mostraron que por debajo de 0.5 m/s la turbulencia del modelo aumentó drásticamente y se convirtió en el modelo inestable e incapaz de producir resultados fiables. Esto es debido al modelo de turbulencia en el software que no es válido para velocidades de viento bajas, lo que limita la capacidad de ENVI-met 4 para realizar simulaciones en estas condiciones de viento y es una de las principales conclusiones de este análisis en cuanto a la herramienta de simulación. También se comprobó el efecto de las densidades de la densidad de hoja (LAD) de los componentes vegetales en el modelo en la capa de aire inmediatamente superior a la cubierta, a 0,5 m sobre este nivel. Se compararon tres alternativas de densidad de hoja con la cubierta de baldosa de hormigón: el techo verde con LAD 0.3 (hierba típica o sedum), LAD 1.5 (plantas mixtas típicas) y LAD 2.5 (masa del árbol). Los resultados mostraron diferencias de temperatura muy relevante entre las diferentes alternativas de LAD analizadas. Los resultados muestran variaciones de temperatura que oscilan entre 3 y 5 º C al comparar el estándar de la azotea concreta con albedo 0, 3 con el techo con vegetación y vegetación densa, mostrando la importancia del LAD en la cuantificación de los efectos de las cubiertas vegetales en microclima circundante, lo que coincide con los datos reportados en la literatura existente y con los estudios empíricos analizados. Los resultados de los análisis teóricos han llegado a las siguientes conclusiones iniciales relacionadas con la herramienta de simulación y los resultados del modelo: En relación con la herramienta ENVI-met, se han observado limitaciones para el análisis. En primer lugar, la estructura rígida de la geometría, las bases de datos y el tamaño de la cuadrícula, limitan la escala y resolución de los análisis no permitiendo el desarrollo de grandes zonas urbanas. Por otro lado la estructura de ENVI-met permite el desarrollo de este tipo de simulación tan complejo dentro de tiempos razonables de cálculo y requerimientos computacionales convencionales. Otra limitación es el modelo de turbulencia del software, que no modela correctamente velocidades de viento bajas (entre 0 y 1 m/s), por debajo de 0,5 m/s el modelo da errores y no es estable, los resultados a estas velocidades no son fiables porque las turbulencias generadas por el modelo hacen imposible la extracción de patrones claros de viento y temperatura que permitan la comparación entre los escenarios de cubierta de hormigón y ajardinada. Además de las limitaciones anteriores, las bases de datos y parámetros de entrada en la versión pública del software están limitados y la complejidad de generar nuevos sistemas adaptándolos al edificio o modelo urbano que se quiera reproducir no es factible salvo en la versión profesional del software. Aparte de las limitaciones anteriores, los patrones de viento y perfiles de temperatura generados por ENVI-met concuerdan con análisis previos en los que se identificaban patrones de variación de viento y temperaturas en cañones urbanos con patrones de viento, relación de aspecto y dimensiones similares a los analizados en esta investigación. Por lo tanto, el software ha demostrado una buena capacidad para reproducir los patrones de viento en los cañones de la calle y capturar el efecto de enfriamiento producido por la cubierta verde en el cañón. En relación con el modelo, el resultado revela la influencia del viento, la radiación y el LAD en la temperatura del aire en cañones urbanos con relación de aspecto comprendida entre 0,5 y 1. Siendo el efecto de la cubierta verde más notable en cañones urbanos sombreados con relación de aspecto 1 y velocidades de viento en el nivel de “canopy” (por encima de la cubierta) de 1 m/s. En ningún caso las reducciones en la temperatura del aire excedieron 1 º C, y las variaciones en la humedad relativa no excedieron 1% entre los escenarios estudiados. Una vez que se han identificado los parámetros relevantes, que fueron principalmente la velocidad del viento y el LAD, se realizó un análisis experimental para comprobar los resultados obtenidos por el modelo. Para éste propósito se identificó una cubierta ajardinada de grandes dimensiones capaz de representar la escala urbana que es el objeto del estudio. El edificio usado para este fin fue el parking de la terminal 4 del aeropuerto internacional de Madrid. Aunque esto no es un área urbana estándar, la escala y la configuración del espacio alrededor del edificio fueron considerados aceptables para el análisis por su similitud con el contexto urbano objeto de estudio. El edificio tiene 800 x 200 m, y una altura 15 m. Está rodeado de vías de acceso pavimentadas con aceras conformando un cañón urbano limitado por el edificio del parking, la calle y el edificio de la terminal T4. El aparcamiento está cerrado con fachadas que configuran un espacio urbano de tipo cañón, con una relación de aspecto menor que 0,5. Esta geometría presenta patrones de viento y velocidad dentro del cañón que difieren ligeramente de los generados en el estudio teórico y se acercan más a los valores a nivel de canopo sobre la cubierta del edificio, pero que no han afectado a la tendencia general de los resultados obtenidos. El edificio cuenta con la cubierta ajardinada más grande en Europa, 12 Ha cubiertas por con una mezcla de hierbas y sedum y con un valor estimado de LAD de 1,5. Los edificios están rodeados por áreas plantadas en las aceras y árboles de sombra en las fachadas del edificio principal. El efecto de la cubierta ajardinada se evaluó mediante el control de temperaturas y humedad relativa en el cañón en un día típico de verano. La selección del día se hizo teniendo en cuenta las predicciones meteorológicas para que fuesen lo más semejantes a las condiciones óptimas para capturar el efecto de la cubierta vegetal sobre el microclima urbano identificadas en el modelo teórico. El 09 de julio de 2014 fue seleccionado para la campaña de medición porque las predicciones mostraban 1 m/s velocidad del viento y cielos despejados, condiciones muy similares a las condiciones climáticas bajo las que el efecto de la cubierta ajardinada era más notorio en el modelo teórico. Las mediciones se registraron cada hora entre las 9:00 y las 19:00 en 09 de julio de 2014. Temperatura, humedad relativa y velocidad del viento se registraron en 5 niveles diferentes, a 1.5, 4.5, 7.5, 11.5 y 16 m por encima del suelo y a 0,5 m de distancia de la fachada del edificio. Las mediciones fueron tomadas en tres escenarios diferentes, con exposición soleada, exposición la sombra y exposición influenciada por los árboles cercanos y suelo húmedo. Temperatura, humedad relativa y velocidad del viento se registraron con un equipo TESTO 410-2 con una resolución de 0,1 ºC para temperatura, 0,1 m/s en la velocidad del viento y el 0,1% de humedad relativa. Se registraron las temperaturas de la superficie de los edificios circundantes para evaluar su efecto sobre los registros usando una cámara infrarroja FLIR E4, con resolución de temperatura 0,15ºC. Distancia mínima a la superficie de 0,5 m y rango de las mediciones de Tª de - 20 º C y 250 º C. Los perfiles de temperatura extraídos de la medición in situ mostraron la influencia de la exposición solar en las variaciones de temperatura a lo largo del día, así como la influencia del calor irradiado por las superficies que habían sido expuestas a la radiación solar así como la influencia de las áreas de jardín alrededor del edificio. Después de que las medidas fueran tomadas, se llevaron a cabo las siguientes simulaciones para evaluar el impacto de la cubierta ajardinada en el microclima: a. estándar de la azotea: edificio T4 asumiendo un techo de tejas de hormigón con albedo 0.3. b. b. cubierta vegetal : T4 edificio asumiendo una extensa cubierta verde con valor bajo del LAD (0.5)-techo de sedum simple. c. c. cubierta vegetal: T4 edificio asumiendo una extensa cubierta verde con alta joven valor 1.5-mezcla de plantas d. d. cubierta ajardinada más vegetación nivel calle: el edificio T4 con LAD 1.5, incluyendo los árboles existentes a nivel de calle. Este escenario representa las condiciones actuales del edificio medido. El viento de referencia a nivel de cubierta se fijó en 1 m/s, coincidente con el registro de velocidad de viento en ese nivel durante la campaña de medición. Esta velocidad del viento se mantuvo constante durante toda la campaña. Bajo las condiciones anteriores, los resultados de los modelos muestran un efecto moderado de azoteas verdes en el microclima circundante que van desde 1 º a 2 º C, pero una contribución mayor cuando se combina con vegetación a nivel peatonal. En este caso las reducciones de temperatura alcanzan hasta 4 ºC. La humedad relativa sin embargo, no presenta apenas variación entre los escenarios con y sin cubierta ajardinada. Las temperaturas medidas in situ se compararon con resultados del modelo, mostrando una gran similitud en los perfiles definidos en ambos casos. Esto demuestra la buena capacidad de ENVI-met para reproducir el efecto de la cubierta ajardinada sobre el microclima y por tanto para el fin de esta investigación. Las diferencias más grandes se registraron en las áreas cercanas a las zonas superiores de las fachadas que estaban más expuestas a la radiación del sol y también el nivel del suelo, por la influencia de los pavimentos. Estas diferencias se pudieron causar por las características de los cerramientos en el modelo que estaban limitados por los datos disponibles en la base de datos de software, y que se diferencian con los del edificio real. Una observación importante derivada de este estudio es la contribución del suelo húmedo en el efecto de la cubierta ajardinada en la temperatura del aire. En el escenario de la cubierta ajardinada con los arboles existentes a pie de calle, el efecto del suelo húmedo contribuye a aumentar las reducciones de temperatura hasta 4.5ºC, potenciando el efecto combinado de la cubierta ajardinada y la vegetación a pie de calle. Se realizó un análisis final después de extraer el perfil horario de temperaturas en el cañón urbano influenciado por el efecto de las cubiertas ajardinadas y los árboles. Con esos perfiles modificados de temperatura y humedad se desarrolló un modelo energético en el edificio asumiendo un edificio cerrado y climatizado, con uso de oficinas, una temperatura de consigna de acuerdo al RITE de 26 ºC, y con los sistemas por defecto que establece el software para el cálculo de la demanda energética y que responden a ASHRAE 90.1. El software seleccionado para la simulación fue Design Builder, por su capacidad para generar simulaciones horarias y por ser una de las herramientas de simulación energética más reconocidas en el mercado. Los perfiles modificados de temperatura y humedad se insertaron en el año climático tipo y se condujo la simulación horaria para el día definido, el 9 de Julio. Para la simulación se dejaron por defecto los valores de conductancia térmica de los cerramientos y la eficiencia de los equipos de acuerdo a los valores que fija el estándar ASHRAE para la zona climática de Madrid, que es la 4. El resultado mostraba reducciones en el consumo de un día pico de hasta un 14% de reducción en las horas punta. La principal conclusión de éste estudio es la confirmación del potencial de las cubiertas ajardinadas como una estrategia para reducir la temperatura del aire y consumo de energía en los edificios, aunque este efecto puede ser limitado por la influencia de los vientos, la radiación y la especie seleccionada para el ajardinamiento, en especial de su LAD. Así mismo, en combinación con los bosques urbanos su efecto se potencia e incluso más si hay pavimentos húmedos o suelos porosos incluidos en la morfología del cañón urbano, convirtiéndose en una estrategia potencial para adaptar los ecosistemas urbanos el efecto aumento de temperatura derivado del cambio climático. En cuanto a la herramienta, ENVI-met se considera una buena opción para éste tipo de análisis dada su capacidad para reproducir de un modo muy cercano a la realidad el efecto de las cubiertas. Aparte de ser una herramienta validada en estudios anteriores, en el caso experimental se ha comprobado por medio de la comparación de las mediciones con los resultados del modelo. A su vez, los resultados y patrones de vientos generados en los cañones urbanos coinciden con otros estudios similares, concluyendo por tanto que es un software adecuado para el objeto de esta tesis doctoral. Como líneas de investigación futura, sería necesario entender el efecto de la cubierta ajardinada en el microclima urbano en diferentes zonas climáticas, así como un mayor estudio de otras variables que no se han observado en este análisis, como la temperatura media radiante y los indicadores de confort. Así mismo, la evaluación de otros parámetros que afectan el microclima urbano tales como variables geométricas y propiedades superficiales debería ser analizada en profundidad para tener un resultado que cubra todas las variables que afectan el microclima en el cañón urbano. ABSTRACT Climate Change is posing an urgency in the development of strategies able not only to mitigate but also adapt to the effects that this global problem is evidencing around the world. Heat waves, flooding and severe draughts increase the vulnerability of population, and this is especially critical in urban settlements. This has been extensively studied over the past decades, addressed from different perspectives and ranging from the regional heat island analysis to the building scale. Its understanding requires physical and dimensional analysis of this broad phenomenon and a deep analysis of the factors and the strategies which can offset it. In the search of solutions to this problem, green infrastructure elements such as green roofs, walls and urban forests arise as strategies able provide multiple regulating ecosystem services to the urban environment able to cope with climate change effects. This includes storm water management, heat island effect control, and improvement of air and water quality. Over the last decade, multiple studies have been developed to evaluate and quantify the ecosystem services provided by green roofs, however, specific regulating services addressing urban microclimate and their impact on the urban dwellers have not been widely quantified. This research tries to contribute to fill this gap and analyzes the effects of green roofs and urban forests on urban microclimate at pedestrian level, quantifying its potential for regulating ambient temperature in hot season in Mediterranean –continental climates. The study is divided into a sequence of analysis where the critical factors affecting the performance of the green roof system on the microclimate are identified and the effects of the green roof is tested in a real case study. The first step has been the definition of the object of study, through the analysis and review of theoretical and empirical papers that investigate the effects of covers landscaped in the built environment, in the context of its use as a tool for adaptation and mitigation of the impact of climate change on cities and urban development. This literature review, reveals the great potential of the plant systems as a tool for passive design capable of improving the climatic and microclimatic conditions in the cities, as well as its positive impact on the energy performance of buildings, but also the need for further analysis at the street scale where climate, urban surfaces and materials, and vegetation converge. This analysis requires a methodology where the thermal buildings response, the variations in the patterns of wind and the interaction of the vegetation are integrated, so a quantitative analysis can help to define the most effective strategies to achieve liveable urban spaces and collaterally, , the improvement of the surrounding buildings energy performance. In this specific scale research is needed and should be customized to every climate, urban condition and nature based strategy. In this context, the main objective for this research was the quantitative assessment of the Green roof impact on the urban microclimate at a neighbourhood scale in summer conditions in Mediterranean- continental climates. For the achievement of this main objective, the following secondary objectives have been set: • Identify the numerical models and calculation tools able to capture the effect of the roof garden on the microclimate. • Identify the enhancing or limiting parameter affecting this effect. • Quantification of the impact of the microclimate created on the energy consumption of buildings surrounding the street canyon analysed. The main hypothesis behind this research and where the above objectives are funded on is as follows: "An extensive roof installed in medium height buildings favours the establishment of microclimates at the pedestrian level and reduces the temperatures in the urban environment where they are located." For the purpose of verifying the above hypothesis and achieving the proposed objectives the following methodology has been followed: - Definition of hypothesis and objectives - Definition of the scope and limitations - Theoretical analysis of parameters affecting gren roof performance - Experimental analysis; - Energy modelling analyisis - Conclusions and future lines of work The search for suitable tools and models for meeting the objectives of this analysis has led to ENVI-met v4 as the most suitable software for this research. ENVI met is a three-dimensional micro-climate model designed to simulate the surface-plant-air interactions in urban environments. Based in the fundamental equations representing, mass, energy and momentum conservation, the software has the capacity of representing the complex phenomena characterizing the microclimate in urban canyons, overcoming the challenge posed by the computing requirements of a full calculus based on finite elements done via traditional computational fluid dynamics tools. Once the analysis tool has been defined, a first set of analysis has been developed to identify the main parameters affecting the green roof influence on the microclimate. In this analysis, two different scenarios are compared. A neighborhood with standard concrete tile roof and the same configuration substituting the concrete tile by an extensive green roof. Once the scenarios have been modeled, different iterations have been run to identify the influence of different wind patterns, solar exposure and roof vegetation type on the microclimate, since those are the most relevant variables affecting urban microclimates. These analysis have been run to check the conditions under which the effects of green roofs get significance. Since ENVI-met V4 is a predictive software, the first step has been the definition of the initial weather conditions which are then used as starting point by the software, which generates its own daily temperature and humidity profile based on the location of the building, geometry, vegetation and the surfaces physical characteristics. The base geometry used for this first analysis has been based on a typical urban layout structure located in Madrid, an orthogonal net with the main streets oriented East-West to ease the analysis of solar radiation in the different points of the model. This layout represents a typical urban neighborhood, with street canyons keeping an aspect ratio between 0.5 and 1 and high sky view factor to ensure correct sun access to the streets and buildings and work with typical wind flow patterns. Finally, the roof vegetation has been defined in terms of foliage density known as Leaf Area Density (LAD) and defined as the total one-sided leaf area per unit of layer volume. This index is the most relevant vegetation characteristic for the purpose of calculating the effect of vegetation on wind and solar radiation as well as the energy consumed during its metabolic processes. The building as usual (BAU) configuring the urban layout has been defined with standard concrete tile roofs, considering 0.3 albedo. Walls have been set with albedo 0.2 (typical brick wall construction) and adiabatic to avoid interference caused by thermal interchanges with the building indoor environment. For the proposed case, the same geometry and building characteristics have been kept. The only change is the roof surface coverage. The gravel on the roof has been changed with an extensive green roof covered with drought tolerant plants, typical in Madrid climate, and characterized by their LAD. The different scenarios analysed are based in the variation of the wind speed and the LAD of the roof. The results have been recorded under different sun exposure conditions. Simulations were run for the typical summer wind patterns, that for Madrid are characterized by South-west winds ranging from 3 to 0 m/s. Simulations were run for 3, 2, 1 and 0 m/s at urban canopy level. Results taken at 1.4 m above the ground showed that the green roof effect was lower with higher wind speeds and in any case the effect of the green roof on the air temperatures exceeded air temperature reductions higher than 1ºC. Relative humidity presented no variations when comparing the different scenarios. For the analysis at 0m/s, ENVI-met generated error and no results were obtained. Different simulations showed that under 0.5 m/s turbulence increased dramatically and the model became unstable and unable to produce reliable results. This is due to the turbulence model embedded in the software which is not valid for low wind speeds (below 1 m/s). The effect of the different foliage densities was also tested in the model. Three different alternatives were compared against the concrete roof: green roof with LAD 0.3 ( typical grass or sedum), 1.5 (typical mixed plants) and 2.5 (tree mass). The results showed very relevant temperature differences between the different LAD alternatives analyzed. Results show temperature variations ranging between 3 and 5 ºC when comparing the standard concrete roof with albedo 0, 3 with the vegetated roof and vegetated mass, showing the relevance of the LAD on the effects of green roofs on microclimate. This matches the data reported in existing literature and empirical studies and confirms the relevance of the LAD in the roof effect on the surrounding microclimate. The results of the theoretical analysis have reached the following initial conclusions related to both, the simulation tool and the model results: • In relation to the tool ENVI-met, some limitations for the analysis have been observed. In first place, the rigid structure of the geometry, the data bases and the grid size, limit the scale and resolution of the analysis not allowing the development of large urban areas. On the other hand the ENVI-met structure enables the development of this type of complex simulation within reasonable times and computational requirements for the purpose of this analysis. Additionally, the model is unable to run simulations at wind speeds lower than 0.5 m/s, and even at this speed, the results are not reliable because the turbulences generated by the model that made impossible to extract clear temperature differences between the concrete and green roof scenarios. Besides the above limitations, the wind patterns and temperature profiles generated by ENVImet are in agreement with previous analysis identifying wind patterns in urban canyons with similar characteristics and aspect ratio. Therefore the software has shown a good capacity for reproducing the wind effects in the street canyons and seems to capture the cooling effect produced by the green roof. • In relation to the model, the results reveals the influence of wind, radiation and LAD on air temperature in urban canyons with aspect ratio comprised between 0.5 and 1. Being the effect of the green roof more noticeable in shaded urban canyons with aspect ratio 1 and wind speeds of 1 m/s. In no case the reductions in air temperature exceeded 1ºC. Once the relevant parameters have been identified, mainly wind speed and LAD, an experimental analysis was conducted to test the results obtained by the model. For this purpose a large green roof was identified, able to represent the urban scale which is the object of the studio. The building identified for this purpose was the terminal 4, parking building of the international Madrid Airport. Even though this is not a standard urban area, the scale and configuration of the space around the building were deemed as acceptable for the analysis. The building is an 800x200 m, 15 m height parking building, surrounded by access paved paths and the terminal building. The parking is enclosed with facades that configure an urban canyon-like space, although the aspect ratio is lower than 0.5 and the wind patterns might differ from the theoretical model run. The building features the largest green roof in Europe, a 12 Ha extensive green roof populated with a mix of herbs and sedum with a LAD of 1.5. The buildings are surrounded by planted areas at the sidewalk and trees shading the main building facades. Green roof performance was evaluated by monitoring temperatures and relative humidity in the canyon in a typical summer day. The day selection was done taking into account meteorological predictions so the weather conditions on the measurement day were as close as possible as the optimal conditions identified in terms of green roof effects on the urban canyon. July 9th 2014 was selected for the measurement campaign because the predictions showed 1 m/s wind speed and sunny sky, which were very similar to the weather conditions where the effect of the green roof was most noticeable in the theory model. Measurements were registered hourly from 9:00am to 19:00 on July 9th 2014. Temperature, relative humidity and wind speed were recorded at 5 different levels, at 1.5, 4.5, 7.5, 11.5 and 16 m above ground and at 0.5 m distance from the building façade. Measurements were taken in three different scenarios, sunny exposure, shaded exposure, and shaded exposure influenced by nearby trees and moist soil. Temperature, relative humidity and wind speed were registered using a TESTO 410-2 anemometer, with 0.1ºC resolution for temperature, 0.1 m/s resolution for wind speed and 0.1 % for relative humidity. Surface temperatures were registered using an infrared camera FLIR E4, with temperature resolution 0.15ºC. Minimal distance to surface of 0.5 m and Tª measurements range from -20ºC and 250ºC. The temperature profiles measured on the site showed the influence of solar exposure on the temperature variations along the day, as well as the influence of the heat irradiated by the building surfaces which had been exposed to the sun radiation and those influenced by the moist soft areas around the building. After the measurements were taken, the following simulations were conducted to evaluate the impact of the green roof on the microclimate: a. Standard roof: T4 building assuming a concrete tile roof with albedo 0.3. b. Green roof: T4 building assuming an extensive green roof with low LAD value (0.5)-Simple Sedum roof. c. Green roof: T4 building assuming an extensive green roof with high LAD value 1.5- Lucerne and grasses d. Green roof plus street level vegetation: T4 Building, LAD 1.5 (Lucerne), including the existing trees at street level. This scenario represents the current conditions of the building. The urban canopy wind was set as 1 m/s, the wind speed register at that level during the measurement campaign. This wind speed remained constant over the whole campaign. Under the above conditions, the results of the models show a moderate effect of green roofs on the surrounding microclimate ranging from 1ºC to 2ºC, but a larger contribution when combining it with vegetation at pedestrian level, where 4ºC temperature reductions are reached. Relative humidity remained constant. Measured temperatures and relative humidity were compared to model results, showing a close match in the profiles defined in both cases and the good capacity of ENVI met to capture the impact of the green roof in this analysis. The largest differences were registered in the areas close to the top areas of the facades which were more exposed to sun radiation and also near to the soil level. These differences might be caused by differences between the materials properties included in the model (which were limited by the data available in the software database) and those in the real building. An important observation derived from this study is the contribution of moist soil to the green roof effect on air temperatures. In the green roof scenario with surrounding trees, the effect of the moist soil contributes to raise the temperature reductions at 4.5ºC. A final analysis was conducted after extracting the hourly temperature profile in the street canyon influenced by the effect of green roofs and trees. An energy model was run on the building assuming it was a conventional enclosed building. Energy demand reductions were registered in the building reaching up to 14% reductions at the peak hour. The main conclusion of this study is the potential of the green roofs as a strategy for reducing air temperatures and energy consumption in the buildings, although this effect can be limited by the influence of high speed winds. This effect can be enhanced its combination with urban forests and even more if soft moist pavements are included in the urban canyon morphology, becoming a potential strategy for adapting urban ecosystems to the increasing temperature effect derived from climate change.

Variations of local heat transfer coefficient in piped flow of viscous liquids

Measurements were carried out to determine local coefficients of heat transfer in short lengths of horizontal pipe, and in the region of an discontinuity in pipe diameter. Laminar, transitional and turbulent flow regimes were investigated, and mixtures of propylene glycol and water were used in the experiments to give a range of viscous fluids. Theoretical and empirical analyses were implemented to find how the fundamental mechanism of forced convection was modified by the secondary effects of free convection, temperature dependent viscosity, and viscous dissipation. From experiments with the short tube it was possible to determine simple empirical relationships describing the axial distribution of the local 1usselt number and its dependence on the Reynolds and Prandtl numbers. Small corrections were made to account for the secondary effects mentioned above. Two different entrance configurations were investigated to demonstrate how conditions upstream could influence the heat transfer coefficients measured downstream In experiments with a sudden contraction in pipe diameter the distribution of local 1u3se1t number depended on the Prandtl number of the fluid in a complicated way. Graphical data is presented describing this dependence for a range of fluids indicating how the local Nusselt number varied with the diameter-ratio. Ratios up to 3.34:1 were considered. With a sudden divergence in pipe diameter, it was possible to derive the axial distribution of the local Nusse1t number for a range of Reynolds and Prandtl numbers in a similar way to the convergence experiments. Difficulty was encountered in explaining some of the measurements obtained at low Reynolds numbers, and flow visualization techniques wore used to determine the complex flow patterns which could lead to the anomalous results mentioned. Tests were carried out with divergences up to 1:3.34 to find the way in which the local Nusselt number varied with the diameter ratio, and a few experiments were carried out with very large ratios up .to 14.4. A limited amount of theoretical analysis of the 'divergence' system was carried out to substantiate certain explanations of the heat transfer mechanisms postulated.

Flow field analysis of some mixing and conveying screw element regions, within a closely intermeshing, co-rotating twin-screw extruder.

Grafting of antioxidants and other modifiers onto polymers by reactive extrusion, has been performed successfully by the Polymer Processing and Performance Group at Aston University. Traditionally the optimum conditions for the grafting process have been established within a Brabender internal mixer. Transfer of this batch process to a continuous processor, such as an extruder, has, typically, been empirical. To have more confidence in the success of direct transfer of the process requires knowledge of, and comparison between, residence times, mixing intensities, shear rates and flow regimes in the internal mixer and in the continuous processor.The continuous processor chosen for the current work in the closely intermeshing, co-rotating twin-screw extruder (CICo-TSE). CICo-TSEs contain screw elements that convey material with a self-wiping action and are widely used for polymer compounding and blending. Of the different mixing modules contained within the CICo-TSE, the trilobal elements, which impose intensive mixing, and the mixing discs, which impose extensive mixing, are of importance when establishing the intensity of mixing. In this thesis, the flow patterns within the various regions of the single-flighted conveying screw elements and within both the trilobal element and mixing disc zones of a Betol BTS40 CICo-TSE, have been modelled using the computational fluid dynamics package Polyflow. A major obstacle encountered when solving the flow problem within all of these sets of elements, arises from both the complex geometry and the time-dependent flow boundaries as the elements rotate about their fixed axes. Simulation of the time dependent boundaries was overcome by selecting a number of sequential 2D and 3D geometries, used to represent partial mixing cycles. The flow fields were simulated using the ideal rheological properties of polypropylene and characterised in terms of velocity vectors, shear stresses generated and a parameter known as the mixing efficiency. The majority of the large 3D simulations were performed on the Cray J90 supercomputer situated at the Rutherford-Appleton laboratories, with pre- and postprocessing operations achieved via a Silicon Graphics Indy workstation. A mechanical model was constructed consisting of various CICo-TSE elements rotating within a transparent outer barrel. A technique has been developed using coloured viscous clays whereby the flow patterns and mixing characteristics within the CICo-TSE may be visualised. In order to test and verify the simulated predictions, the patterns observed within the mechanical model were compared with the flow patterns predicted by the computational model. The flow patterns within the single-flighted conveying screw elements in particular, showed good agreement between the experimental and simulated results.

Spray drier simulation and air flow pattern studies

The literature pertaining to the key stages of spray drying has been reviewed in the context of the mathematical modelling of drier performance. A critical review is also presented of previous spray drying models. A new mathematical model has been developed for prediction of spray drier performance. This is applicable to slurries of rigid, porous crust-forming materials to predict trajectories and drying profiles for droplets with a distribution of sizes sprayed from a centrifugal pressure nozzle. The model has been validated by comparing model predictions to experimental data from a pilot-scale counter-current drier and from a full-scale co-current drier. For the latter, the computed product moisture content was within 2%, and the computed air exit temperature within 10oC of experimental data. Air flow patterns have been investigated in a 1.2m diameter transparent countercurrent spray tower by flow visualisation. Smoke was introduced into various zones within the tower to trace the direction, and gauge the intensity, of the air flow. By means of a set of variable-angle air inlet nozzles, a variety of air entry configurations was investigated. The existence of a core of high rotational and axial velocity channelling up the axis of the tower was confirmed. The stability of flow within the core was found to be strongly dependent upon the air entry arrangement. A probe was developed for the measurement of air temperature and humidity profiles. This was employed for studying evaporation of pure water drops in a 1.2m diameter pilot-scale counter-current drier. A rapid approach to the exit air properties was detected within a 1m distance from the air entry ports. Measured radial profiles were found to be virtually flat but, from the axial profiles, the existence of plug-flow, well-mixed-flow and some degree of air short-circuiting can be inferred. The model and conclusions should assist in the improved design and optimum operation of industrial spray driers.

Stilling Basin Scour Remediation Using Air Injection and Flat Plate Extension

The South Florida Water Management District (SFWMD) is responsible for managing over 2500 miles of waterways and hundreds of water control structures. Many of these control structures are experiencing erosion, known as scour, of the sediment downstream of the structure. Laboratory experiments were conducted in order to investigate the effectiveness of two-dimensional air diffusers and plate extensions (without air injection) on a 1/30 scale model of one of SFWMD gated spillway structures, the S65E gated spillway. A literature review examining the results of similar studies was conducted. The experimental design for this research was based off of previous work done on the same model. Scour of the riverbed downstream of gated spillway structures has the potential to cause serious damage, as it can expose the foundation of the structure, which can lead to collapse. This type of scour has been studied previously, but it continues to pose a risk to water control structures and needs to be studied further. The hydraulic scour channel used to conduct experiments contains a head tank, flow straighteners, gated spillway, stilling basin, scour chamber, sediment trap, and tailwater tank. Experiments were performed with two types of air diffusers. The first was a hollow, acrylic, triangular end sill with air injection holes on the upstream face, allowing for air injection upstream. The second diffuser was a hollow, acrylic rectangle that extended from the triangular end sill with air injection holes in the top face, allowing for vertical air injection, perpendicular to flow. Detailed flow and bed measurements were taken for six trials for each diffuser ranging from no air injection to 5 rows of 70 holes of 0.04" diameter. It was found that with both diffusers, the maximum amount of air injection reduced scour the most. Detailed velocity measurements were taken for each case and turbulence statistics were analyzed to determine why air injection reduces scour. It was determined that air injection reduces streamwise velocity and turbulence. Another set of experiments was performed using an acrylic extension plate with no air injection to minimize energy costs. Ten different plate lengths were tested. It was found that the location of deepest scour moved further downstream with each plate length. The 32-cm plate is recommended here. Detailed velocity measurements were taken after the cases with the 32-cm plate and no plate had reached equilibrium. This was done to better understand the flow patterns in order to determine what causes the scour reduction with the extension plates. The extension plate reduces the volume of scour, but more importantly translates the deepest point of scour downstream from the structure, lessening the risk of damage.

Automated control of a dragline using machine vision

The research described in this paper is directed toward increasing productivity of draglines through automation. In particular, it focuses on the swing-to-dump, dump, and return-to-dig phases of the dragline operational cycle by developing a swing automation system. In typical operation the dragline boom can be in motion for up to 80% of the total cycle time. This provides considerable scope for improving cycle time through automated or partially automated boom motion control. This paper describes machine vision based sensor technology and control algorithms under development to solve the problem of continuous real time bucket location and control. Incorporation of this capability into existing dragline control systems will then enable true automation of dragline swing and dump operations.

A programming language for software components

Component software has many benefits, most notably increased software re-use; however, the component software process places heavy burdens on programming language technology, which modern object-oriented programming languages do not address. In particular, software components require specifications that are both sufficiently expressive and sufficiently abstract, and, where possible, these specifications should be checked formally by the programming language. This dissertation presents a programming language called Mentok that provides two novel programming language features enabling improved specification of stateful component roles. Negotiable interfaces are interface types extended with protocols, and allow specification of changing method availability, including some patterns of out-calls and re-entrance. Type layers are extensions to module signatures that allow specification of abstract control flow constraints through the interfaces of a component-based application. Development of Mentok's unique language features included creation of MentokC, the Mentok compiler, and formalization of key properties of Mentok in mini-languages called MentokP and MentokL.

Flow modeling in a novel non-perfusion conical bioreactor

We have developed a bioreactor vessel design which has the advantages of simplicity and ease of assembly and disassembly, and with the appropriately determined flow rate, even allows for a scaffold to be suspended freely regardless of its weight. This article reports our experimental and numerical investigations to evaluate the performance of a newly developed non-perfusion conical bioreactor by visualizing the flow through scaffolds with 45° and 90° fiber lay down patterns. The experiments were conducted at the Reynolds numbers (Re) 121, 170, and 218 based on the local velocity and width of scaffolds. The flow fields were captured using short-time exposures of 60 µm particles suspended in the bioreactor and illuminated using a thin laser sheet. The effects of scaffold fiber lay down pattern and Reynolds number were obtained and correspondingly compared to results obtained from a computational fluid dynamics (CFD) software package. The objectives of this article are twofold: to investigate the hypothesis that there may be an insufficient exchange of medium within the interior of the scaffold when using our non-perfusion bioreactor, and second, to compare the flows within and around scaffolds of 45° and 90° fiber lay down patterns. Scaffold porosity was also found to influence flow patterns. It was therefore shown that fluidic transport could be achieved within scaffolds with our bioreactor design, being a non-perfusion vessel. Fluid velocities were generally same of the same or one order lower in magnitude as compared to the inlet flow velocity. Additionally, the 90° fiber lay down pattern scaffold was found to allow for slightly higher fluid velocities within, as compared to the 45° fiber lay down pattern scaffold. This was due to the architecture and pore arrangement of the 90° fiber lay down pattern scaffold, which allows for fluid to flow directly through (channel-like flow).

Perturbation solutions for flow through symmetrical hoppers with inserts and asymmetrical wedge hoppers

Under certain circumstances, an industrial hopper which operates under the "funnel-flow" regime can be converted to the "mass-flow" regime with the addition of a flow-corrective insert. This paper is concerned with calculating granular flow patterns near the outlet of hoppers that incorporate a particular type of insert, the cone-in-cone insert. The flow is considered to be quasi-static, and governed by the Coulomb-Mohr yield condition together with the non-dilatant double-shearing theory. In two dimensions, the hoppers are wedge-shaped, and as such the formulation for the wedge-in-wedge hopper also includes the case of asymmetrical hoppers. A perturbation approach, valid for high angles of internal friction, is used for both two-dimensional and axially symmetric flows, with analytic results possible for both leading order and correction terms. This perturbation scheme is compared with numerical solutions to the governing equations, and is shown to work very well for angles of internal friction in excess of 45 degree.

Designing a workflow system using coloured petri nets

Traditional workflow systems focus on providing support for the control-flow perspective of a business process, with other aspects such as data management and work distribution receiving markedly less attention. A guide to desirable workflow characteristics is provided by the well-known workflow patterns which are derived from a comprehensive survey of contemporary tools and modelling formalisms. In this paper we describe the approach taken to designing the newYAWL workflow system, an offering that aims to provide comprehensive support for the control-flow, data and resource perspectives based on the workflow patterns. The semantics of the newYAWL workflow language are based on Coloured Petri Nets thus facilitating the direct enactment and analysis of processes described in terms of newYAWL language constructs. As part of this discussion, we explain how the operational semantics for each of the language elements are embodied in the newYAWL system and indicate the facilities required to support them in an operational environment. We also review the experiences associated with developing a complete operational design for an offering of this scale using formal techniques.

Pre-crash traffic flow trend analysis on motorways

Crashes on motorway contribute to a significant proportion (40-50%) of non-recurrent motorway congestions. Hence reduce crashes will help address congestion issues (Meyer, 2008). Crash likelihood estimation studies commonly focus on traffic conditions in a Short time window around the time of crash while longer-term pre-crash traffic flow trends are neglected. In this paper we will show, through data mining techniques, that a relationship between pre-crash traffic flow patterns and crash occurrence on motorways exists, and that this knowledge has the potential to improve the accuracy of existing models and opens the path for new development approaches. The data for the analysis was extracted from records collected between 2007 and 2009 on the Shibuya and Shinjuku lines of the Tokyo Metropolitan Expressway in Japan. The dataset includes a total of 824 rear-end and sideswipe crashes that have been matched with traffic flow data of one hour prior to the crash using an incident detection algorithm. Traffic flow trends (traffic speed/occupancy time series) revealed that crashes could be clustered with regards of the dominant traffic flow pattern prior to the crash. Using the k-means clustering method allowed the crashes to be clustered based on their flow trends rather than their distance. Four major trends have been found in the clustering results. Based on these findings, crash likelihood estimation algorithms can be fine-tuned based on the monitored traffic flow conditions with a sliding window of 60 minutes to increase accuracy of the results and minimize false alarms.

Pre-crash and non-crash traffic flow trends analysis on motorways

Crashes that occur on motorways contribute to a significant proportion (40-50%) of non-recurrent motorway congestions. Hence, reducing the frequency of crashes assists in addressing congestion issues (Meyer, 2008). Crash likelihood estimation studies commonly focus on traffic conditions in a short time window around the time of a crash while longer-term pre-crash traffic flow trends are neglected. In this paper we will show, through data mining techniques that a relationship between pre-crash traffic flow patterns and crash occurrence on motorways exists. We will compare them with normal traffic trends and show this knowledge has the potential to improve the accuracy of existing models and opens the path for new development approaches. The data for the analysis was extracted from records collected between 2007 and 2009 on the Shibuya and Shinjuku lines of the Tokyo Metropolitan Expressway in Japan. The dataset includes a total of 824 rear-end and sideswipe crashes that have been matched with crashes corresponding to traffic flow data using an incident detection algorithm. Traffic trends (traffic speed time series) revealed that crashes can be clustered with regards to the dominant traffic patterns prior to the crash. Using the K-Means clustering method with Euclidean distance function allowed the crashes to be clustered. Then, normal situation data was extracted based on the time distribution of crashes and were clustered to compare with the “high risk” clusters. Five major trends have been found in the clustering results for both high risk and normal conditions. The study discovered traffic regimes had differences in the speed trends. Based on these findings, crash likelihood estimation models can be fine-tuned based on the monitored traffic conditions with a sliding window of 30 minutes to increase accuracy of the results and minimize false alarms.