Abstract: Water quality sampling is a key element in tracking water quality monitoring objectives. However, frequencies adapted by different agencies might not be sufficient to provide an accurate indication of water quality status. In this study, data from low- and high-resolution water quality datasets were analyzed to determine the extent to which monitoring objectives could be achieved with different sampling frequencies, with a view to providing recommendations and best practices for water quality monitoring frequency in places with limited resources with which to implement a high-frequency monitoring plan. Water quality data from two watersheds (Maumee River and Raisin River) located in the Western Lake Erie Basin (WLEB) were used since these watersheds have consistent records over substantial periods of time, and the water quality data available have a high resolution (at least daily). The water quality constituents analyzed included suspended solids (SS), total phosphorus (TP), soluble reactive phosphorus (SRP), and nitrate + nitrite (NO2+3). Sources of pollutants for watersheds located in the WLEB include contributions from point sources like discharges from sewage treatment plants and non-point sources such as agricultural and urban storm runoff. Weekly, bi-weekly, monthly, and seasonal datasets were created from the original datasets, following different sampling rules based on the day of the week, week of the month, and month of the year. The resulting datasets were then compared to the original dataset to determine how the sampling frequency would affect the results obtained in a water quality assessment when different monitoring objectives are considered. Results indicated that constituents easily transported by water (such as sediments and nutrients) require more than 50 samples/year to provide a small error (< 10%) with a confidence interval of 95%. Monthly and seasonal sampling were found appropriate to report a stream’s prevailing water quality status and statistical properties. However, these resolutions might not be sufficient to capture longterm trends, in which case bi-weekly samples would be preferable. Limitations of lowresolution sampling frequency could be overcome by including rainfall events and random sampling during specific time windows as part of the monitoring plan.

Abstract: Water quality sampling is a key element in tracking water quality monitoring objectives. However, frequencies adapted by different agencies might not be sufficient to provide an accurate indication of water quality status. In this study, data from low- and high-resolution water quality datasets were analyzed to determine the extent to which monitoring objectives could be achieved with different sampling frequencies, with a view to providing recommendations and best practices for water quality monitoring frequency in places with limited resources with which to implement a high-frequency monitoring plan. Water quality data from two watersheds (Maumee River and Raisin River) located in the Western Lake Erie Basin (WLEB) were used since these watersheds have consistent records over substantial periods of time, and the water quality data available have a high resolution (at least daily). The water quality constituents analyzed included suspended solids (SS), total phosphorus (TP), soluble reactive phosphorus (SRP), and nitrate + nitrite (NO2+3). Sources of pollutants for watersheds located in the WLEB include contributions from point sources like discharges from sewage treatment plants and non-point sources such as agricultural and urban storm runoff. Weekly, bi-weekly, monthly, and seasonal datasets were created from the original datasets, following different sampling rules based on the day of the week, week of the month, and month of the year. The resulting datasets were then compared to the original dataset to determine how the sampling frequency would affect the results obtained in a water quality assessment when different monitoring objectives are considered. Results indicated that constituents easily transported by water (such as sediments and nutrients) require more than 50 samples/year to provide a small error (< 10%) with a confidence interval of 95%. Monthly and seasonal sampling were found appropriate to report a stream’s prevailing water quality status and statistical properties. However, these resolutions might not be sufficient to capture longterm trends, in which case bi-weekly samples would be preferable. Limitations of lowresolution sampling frequency could be overcome by including rainfall events and random sampling during specific time windows as part of the monitoring plan.

Abstract: Resilience has become a fundamental paradigm for communities to deal with disaster planning. Formal methods are used to prioritise and decide about investments for resilience. Strategies and behaviour need to be developed that cannot be based on formal modelling only because the human element needs to be incorporated to build community resilience. Participatory modelling and gaming are methodological approaches that are based on realistic data and address human behaviour. These approaches enable stakeholders to develop, adjust, and learn from interactive models and use this experience to inform their decision?making. In our contribution, we explore which physical and digital elements from serious games can be used to design a participatory approach in community engagement and decision?making. Our ongoing research aims to bring multiple stakeholders together to understand, model, and decide on the trade?offs and tensions between social and infrastructure investments toward community resilience building. Initial observations allow us as researchers to systematically document the benefits and pitfalls of a gamebased approach. We will continue to develop a participatory modelling exercise for resilience planning with university graduate students and resilience experts within academia in Christchurch, New Zealand. Resumen: La resiliencia se ha convertido en un paradigma fundamental para que las comunidades afronten la planificación de las catástrofes. Se utilizan métodos formales para priorizar y decidir sobre las inversiones en resiliencia. Es necesario desarrollar estrategias y comportamientos que no pueden que no pueden basarse únicamente en modelos formales, ya que es necesario incorporar el elemento humano para aumentar la resiliencia de la comunidad. La modelización participativa y los juegos son enfoques metodológicos que se basan en datos realistas y abordan el comportamiento humano. Estos enfoques permiten que las partes interesadas desarrollen, ajusten y aprendan de los modelos interactivos y utilicen esta experiencia para informar su toma de decisiones. En nuestra contribución, exploramos qué elementos físicos y digitales de los juegos seriados pueden utilizarse para diseñar un enfoque participativo en el compromiso y la toma de decisiones de la comunidad. Nuestra investigación en curso pretende reunir a múltiples partes interesadas para comprender, modelar y decidir sobre las compensaciones y tensiones entre las inversiones sociales y las de infraestructura para el desarrollo de la resiliencia de la comunidad. Las observaciones iniciales nos permiten, como documentar sistemáticamente las ventajas y los inconvenientes de un enfoque basado en el juego. Seguiremos desarrollando un ejercicio de modelización participativa para la planificación de la resiliencia con estudiantes universitarios y expertos en resiliencia dentro de universitarios y expertos en resiliencia de Christchurch (Nueva Zelanda).

Abstract: This thesis addresses the numerical optimization of the design and scheduling of a biomethane plant integrated with a reversible solid oxide cell, amongst other technologies, that allow the processing of biogas coming from OFMSW anaerobic digestion as to maximize profitability on a yearly basis under different circumstances. The optimization model is carried out through a mixed integer linear program as to obtain the adequate sizes of the different technologies that compose the integrated energy plant whilst taking advantage of the energy and material stream interactions to reach optimal financial feasibility. The inclusion of a reversible oxide fuel cell allows the system to work in different modes as in Power-to-gas and Gas-to-power, generating interest on to the different services that the system can provide to the gas and electricity grids and most importantly to a future decarbonized energy scenario. Furthermore, inclusion of linear models for an amine scrubbing section for biomethane upgrading, a photovoltaic solar field, gas and heat storage units, a battery energy storage system and a biogas boiler into the MILP is carried out as they are part of the technologies in study for the integrated energy plant. The seasonal variability for gas and electricity prices, part load performances, size effects, weather related data, and technological modelling of components are all part of the challenges to be tackled as to obtain a MILP able to compute accurate, adequate and optimal solutions for the design and scheduling problem whilst subjected to technological, environmental and operational constraints. Resumen: Esta tesis se centra en la optimización numérica del diseño y operación de una planta de biometano integrada con una celda de hidrogeno reversible, entre otras tecnologías, que permiten el procesamiento de biogás proveniente de la digestión anaeróbica de material orgánico con el objetivo de maximizar las ganancias en temporalidad anual ante diferentes circunstancias. El modelo de optimización se lleva a cabo con un algoritmo de programación lineal de entornos mixtos para obtener los tamaños de diseño adecuados de las diferentes tecnologías que componen a la planta integrada de energía mientras se toma ventaja de las interacciones entre cada flujo de energía y/o materia para obtener la mas óptima viabilidad económica. La inclusión de una celda reversible de hidrógeno permite al sistema trabajar en diferentes modos como electricidad a gas o gas a electricidad, generando interés en los distintos servicios que el sistema puede proveer a las redes de gas y electricidad y aun más importante a un futuro escenario mundial de energético descarbonizado. Mas aún, la inclusión de modelos lineales como un sistema de mejora de la calidad del biometano mediante absorción con aminas, un campo de energía solar fotovoltaica, unidades de almacenamiento de gas y calor, un sistema de almacenamiento de energías mediante baterías y una caldera de biogás en el modelo de programación lineal es llevado a cabo pues son parte de las tecnologías en estudio para la planta energética integrada. La variación estacional por los precios de gas y electricidad, desempeño a cargas parciales, efectos de tamaños, data ambiental, y el modelado tecnológico de los componentes son parte de los retos a ser resueltos para poder obtener un MILP capaz de computar precisas, adecuadas y optimas soluciones para el problema de diseño y operación mientras se encuentra sujeto a restricciones tecnológicas, ambientales y operacionales.

Abstract: This project presents three-dimensional conceptual and numerical flow models of a gascondensate field in the foothills region of the Eastern Llanos Basin in Colombia. The Eastern Llanos Basin has been the subject of diverse studies for hydrocarbon exploitation. Over the past decade, this region has been assessed for its geothermal potential in currently exploited oil & gas fields. A gas-condensate field located in the foothills of the basin represents a promising opportunity to produce geothermal energy. The conceptual and numerical models integrate geological, geophysical, geochemical and well data to enhance the understanding of the geothermal system in a sedimentary environment. The Leapfrog Geothermal model shows that the gas-condensate field presents a structurally complex architecture, where hydrocarbons and water accumulate in an relatively permeable reservoir. Well data exhibit a constant geothermal gradient, indicating conductive heat transfer. Based on geochemical data and hydrological analyses, it is identified that meteoric recharge occurs in the Andean orogenic belt and fluid flows from NW to SE. TOUGH2 numerical simulations are developed to generate a natural state model and geothermal production scenarios, which give an estimation of water extraction in the field. The thermal and hydraulic parameters were extracted from published experimental and well data. In order to simplify this first geotermal model, and reduce the number of parameters, it is assumed that the field is completely saturated with water. The natural state model temperature field is concordant with measured well data and simulated mass flow direction confirm expected patterns. Production scenarios demonstrate that it is possible to extract water at more than 90 °C without termal breakthrough during the 30 years simulation. Thermal power calculations yield two optimistic scenarios with more than 30 MWth of heat production, and a conservative scenario with approximately 6 MWth. The conservative scenario provides a better approach to reality with current water production. Thus, this project presents the basis for further studies on geothermal development in actively exploited hydrocarbon fields.

Abstract: In the past, when power systems did not incorporate a considerable amount of distributed generation units, the almost passive behavior of the distribution networks could be reasonably represented by static load models (e.g., the exponential model), because they had relatively little impact on power system stability. Therefore, from a transmission perspective, the distribution systems could be seen as an aggregated load for bulk power system stability studies. However, in modern systems, the impact of distributed generation units can no longer be neglected. The static models cannot represent the dynamics of the active distributed networks, which are now increasingly spread across power systems. However, modeling active distribution networks in detail is not feasible due to information exchange restrictions between transmission and distribution system operators, in addition to the great computational effort that it would take. This thesis proposes an alternative black-box, simulation-based methodology to identify dynamic equivalents of active distribution networks focused on long-term voltage stability studies for bulk power systems. The methodology is based on model identification techniques and Hammerstein-Wiener models. The approach depends only on variables measured at the boundary buses, which prevents the need to have previous knowledge about the system, as is the case with greybox models. Furthermore, to avoid heavy-duty nonlinear model techniques, such as artificial neural networks, which are challenging to implement and use in a rather conservative sector such as power systems planning and operation, the proposed methodology keeps a transparent relationship to linear systems. The results and conclusions are derived from time-domain simulations on the IEEE test system for voltage stability assessment, the Nordic test system. It is concluded that the dynamic equivalents are able to reproduce accurately the nonlinear behavior of the active distribution networks. Resumen: En el pasado, cuando los sistemas de energía no incorporaban una cantidad considerable de unidades de generación distribuida, el comportamiento casi pasivo de las redes de distribución podía representarse razonablemente mediante modelos de carga estática (por ejemplo, el modelo exponencial), porque tenían un impacto relativamente pequeño en la energía. estabilidad del sistema. Por lo tanto, desde una perspectiva de transmisión, los sistemas de distribución podrían verse como una carga agregada para los estudios de estabilidad del sistema de energía a granel. Sin embargo, en los sistemas modernos, el impacto de las unidades de generación distribuida ya no puede pasarse por alto. Los modelos estáticos no pueden representar la dinámica de las redes distribuidas activas, que ahora se distribuyen cada vez más en los sistemas de energía. Sin embargo, modelar las redes de distribución activas en detalle no es factible debido a las restricciones de intercambio de información entre los operadores de los sistemas de transmisión y distribución, además del gran esfuerzo computacional que requeriría. Esta tesis propone una metodología alternativa basada en simulación de caja negra para identificar equivalentes dinámicos de redes de distribución activas enfocadas en estudios de estabilidad de voltaje a largo plazo para sistemas de energía a granel. La metodología se basa en técnicas de identificación de modelos y modelos de Hammerstein-Wiener. El enfoque depende únicamente de las variables medidas en los buses delimitadores, lo que evita la necesidad de tener conocimientos previos sobre el sistema, como es el caso de los modelos de caja gris. Los resultados y las conclusiones se derivan de simulaciones en el dominio del tiempo en el sistema de prueba IEEE para la evaluación de la estabilidad de voltaje, el sistema de prueba nórdico.

Abstract: This Master thesis presents a decentralized DDC strategy for major household appliances in Berlin. It exploits their power flexibility by means of three key indicators and supported by artificial intelligence. It has been examined whether those appliances may be used in a decentralized frequency regulation successfully. The appliances have been divided into two groups according to their operating characteristics- Thermal and Sequential Loads. This work also inquiries about the suitability of the control strategy to overcome power imbalances caused by PV generation. To answer this question, different seasonal and consumption scenarios have been analyzed. The control strategy is implemented using a modular approach. For this purpose, three main stages have been defined and divided into smaller units. For the implementation of these units Pandapower, MATLAB, and its toolbox MATPOWER have been used. The first stage is the power consumption forecast using a Neural Network. A feed-forward Network has been designed to estimate a day-ahead consumption profile for each household. At the second stage the Sequential Loads optimal scheduling takes place. A Genetic Algorithm optimizes the Sequential Loads turn on time based on householder preferences and characteristics. At the last stage the Dynamic Demand Control strategy takes place. On one hand, LSI stands for the characteristics of the appliances and for the user´s comfort. On the other hand, GSI and GCI represent the grid state from a technical and economic point of view. To validate the method a suitable network has been implemented. This network has 1800 households and includes 198 residential rooftop PV systems. Besides, several scenarios have been designed to test and compare the control strategy under different conditions. Influencing factors such as Energy Consumption Patterns and the size of the higher-level distribution grid are been modified. The analysis of the simulations shows that the proposed DDC strategy succeeds in frequency stabilization with flexible loads. The Energy Consumption Patterns play a relevant role to achieve this regulation. ECP d- the most realistic pattern, allows a successful control during the entire year with a significant low X value equal 10. These positive results allow for the conclusion that a decentralized control strategy with no information exchange can be implemented in a city like Berlin. Resumen: Esta tesis de maestría presenta una estrategia de control dinámica de demanda descentralizada para los principales electrodomésticos en la ciudad de Berlín. Esta estrategia explota la flexibilidad energética de los electrodomésticos mediante tres indicadores clave y al apoyarse en inteligencia artificial. Se ha examinado si esos aparatos pueden utilizarse con éxito en la regulación de frecuencia descentralizada. Los aparatos se han dividido en dos grupos según sus características de funcionamiento: cargas térmicas y secuenciales. Este trabajo también indaga sobre la idoneidad de la estrategia de control para superar los desequilibrios energéticos provocados por la generación fotovoltaica. Para responder a esta pregunta se han analizado diferentes escenarios estacionales y de consumo. La estrategia de control se implementa mediante un enfoque modular. Para ello, se han definido tres etapas principales y se han dividido en unidades más pequeñas. Para la implementación de estas unidades se han utilizado Pandapower, MATLAB y su aplicativo MATPOWER. La primera etapa es la pronostico del consumo de energía mediante una red neuronal. Se ha diseñado una red feed-forward para estimar el perfil de consumo diario para cada hogar. En la segunda etapa tiene lugar la programación óptima de cargas secuenciales. Un algoritmo genético optimiza el tiempo de activación de las cargas secuenciales en función de las preferencias y características del hogar. En la última etapa se lleva a cabo la estrategia de control dinámico de la demanda. Por un lado, LSI representa las características de los electrodomésticos y la comodidad del usuario. Por otro lado, GSI y GCI representan el estado de la red desde un punto de vista técnico y económico. Para validar el método se ha implementado una red adecuada. Esta red tiene 1800 hogares e incluye 198 sistemas fotovoltaicos residenciales en tejados. Además, se han diseñado varios escenarios para probar la estrategia de control en diferentes condiciones. Se modifican factores de influencia como los Patrones de Consumo de Energía y el tamaño de la red de distribución de nivel superior. El análisis de las simulaciones muestra que la estrategia DDC propuesta tiene éxito en la estabilización de frecuencia con cargas flexibles. Los Patrones de Consumo de Energía juegan un papel relevante para lograr esta regulación. ECP d- el patrón más realista, permite un control exitoso durante todo el año con un valor X bajo significativo igual a 10. Estos resultados positivos permiten concluir que una estrategia de control descentralizado sin intercambio de información se puede implementar en una ciudad como Berlín.

Abstract: Power systems have been dominated by synchronous generators for more than a century. The power system stability has been classically categorized with respect to the dynamics and interactions of the synchronous generator. The classic categories are rotor-angle, frequency, and voltage stability. However, due to the recent high penetration of inverterbased generators such as solar panels and wind turbines, instabilities have arisen outside the spectrum of the classical power system definitions. Therefore, in 2020, the slowinteraction converter-driven stability has been included in the power system stability classification by the IEEE Power System Dynamic Performance Committee. This instability has been observed in real-life cases showing undamped voltage oscillations, which have led to the partial disconnection of power systems. This work focuses on this recently classified instability whose investigation is in the early stages regarding its cause, detection and prevention. Therefore, this work proposes a methodology for the detection and prediction of this instability. This methodology is based on time-domain simulations, the computation of eigenvalues of the power system and the participation factors of oscillatory modes. The proposed tool has been effective in not only detecting but also anticipating the onset of converter-driven instability in operational scenarios with long-term voltage instability. The elements causing the instability are identified using the participation factors, which facilitate the tuning of inverter-based generators controls. Thereby, it is possible to mitigate the converter-driven instability before its manifestation and thus avoid undesired disconnection of generation units, or in the worst case, a potential system collapse. Resumen: Los sistemas de energía han estado dominados por generadores síncronos durante más de un siglo. La estabilidad del sistema de potencia se ha categorizado clásicamente con respecto a la dinámica y las interacciones del generador síncrono. Las categorías clásicas son el ángulo del rotor, la frecuencia y la estabilidad de voltaje. Sin embargo, debido a la reciente alta penetración de generadores basados en inversores, como paneles solares y turbinas eólicas, han surgido inestabilidades fuera del espectro de las definiciones clásicas de sistemas de energía. Por lo tanto, en 2020, la estabilidad impulsada por el convertidor de interacción lenta ha sido incluida en la clasificación de estabilidad del sistema de energía por el Comité de Desempeño Dinámico del Sistema de Energía de IEEE. Esta inestabilidad se ha observado en casos de la vida real que muestran oscilaciones de tensión no amortiguadas, que han provocado la desconexión parcial de los sistemas eléctricos. Este trabajo se centra en esta inestabilidad recientemente clasificada cuya investigación se encuentra en etapas tempranas en cuanto a su causa, detección y prevención. Por tanto, este trabajo propone una metodología para la detección y predicción de esta inestabilidad. Esta metodología se basa en simulaciones en el dominio del tiempo, el cálculo de los valores propios del sistema eléctrico y los factores de participación de los modos oscilatorios. La herramienta propuesta ha sido eficaz no solo para detectar sino también para anticipar el inicio de la inestabilidad impulsada por el convertidor en escenarios operativos con inestabilidad de voltaje a largo plazo. Los elementos que causan la inestabilidad se identifican utilizando los factores de participación, que facilitan el ajuste de los controles de los generadores basados en inversores. De esta manera, es posible mitigar la inestabilidad impulsada por el convertidor antes de su manifestación y así evitar la desconexión no deseada de las unidades de generación o, en el peor de los casos, un colapso potencial del sistema.

Abstract: The aim of this investigation is to assess the behavior of the low-cost sensors under different levels of temperature and relative humidity, which are the main parameters affecting their performance. Several set of experiments were carried out in both, laboratory and field conditions using four low-cost sensors from the company Alphasense. As a short summary of the experimental plan, the low-cost sensors were firstly tested under six relative humidity and four temperature levels against traditional reference instruments. This with the aim of developing a correction algorithm for these two parameters. Then the implementation of a low-cost dryer was evaluated which would reduce the influence of the mentioned parameters on the low-cost sensor behavior. All of these stated tests were performed in controlled conditions in the laboratory. In the end, a long-term experiment in the field was done. This experiment had a duration of one month, and its aim was to appraise the algorithm calculated from the laboratory when comparing their results against the concentrations measured by a monitoring station in Marienplatz, Stuttgart. The relative humidity levels achieved during the laboratory experiments were from 10% until 85%, increasing in steps of 15%, and the temperature levels were 10°C, 25°C, 35°C and 45°C. All low-cost sensors showed a significant improvement when the developed correction algorithm was applied, which is reflected in the selected evaluation parameters (MAPE, MAE, linearity and offset). The relative humidity levels used for the assessment of the low-cost dryer were 40%, 55%, 70% and 85%, finding similar results obtained from the previous experiments. However, with the advantage that in this case, after the drying process, the relative humidity and temperature levels were stable, which leads to avoid applying a complex correction algorithm. The concentrations measured by the low-cost sensors with the applied correction showed improved results, with absolute errors of 38.18 ppb, 7.64 ppb and 4.27 ppb for CO, NO and NO2, respectively. Further investigation is also suggested, regarding the use of different set of sensor for checking the reproducibility as well as dynamic changes in the influencing parameters.

Abstract: Using low-cost gas sensors for air quality monitoring promises cost e ective and convenient measurement systems. Nevertheless, the results obtained have a questionable quality due to di erent factors that can a ect sensor performance. The most discussed ones are relative humidity and air temperature. This investigation aimed to assess the behavior of B4-series low-cost gas sensors from Alphasense for measuring CO, NO, NO2, and O3 for di erent levels of relative humidity and temperature. These low-cost gas sensors were tested for six relative humidity levels from 10% to 85% with increasing steps of 15% and four temperature levels of 10 C, 25 C, 35 C, and 45 C against reference instruments in the laboratory. The e ect of these parameters on low-cost gas sensors was quantified in laboratory from which a correction algorithm was calculated, which was then applied to the field data. The applied algorithm improved the data quality of the low-cost gas sensors in most of the cases. Additionally, a low-cost dryer was assessed to reduce the influence of these factors on the low-cost gas sensors, which also proved to be suitable to enhance the data quality.