DA - TECNUN - Tesis doctorales
Permanent URI for this collectionhttps://hdl.handle.net/10171/62067
See
151 results
Results
- Diffeomorphic transformations for time series analysis: An efficient approach to nonlinear warping.(Servicio de Publicaciones. Universidad de Navarra, 2023-09-11) Martínez Lopez, I. (Iñigo); Viles-Díez, E. (Elisabeth); García Olaizola, I. (Igor)The proliferation and ubiquity of temporal data across many disciplines has sparked interest for similarity, classification and clustering methods specifically designed to handle time series data. A core issue when dealing with time series is determining their pairwise similarity, i.e., the degree to which a given time series resembles another. Traditional distance measures such as the Euclidean are not well-suited due to the time-dependent nature of the data. Elastic metrics such as dynamic time warping (DTW) offer a promising approach, but are limited by their computational complexity, non-differentiability and sensitivity to noise and outliers. This thesis proposes novel elastic alignment methods that use parametric & diffeomorphic warping transformations as a means of overcoming the shortcomings of DTW-based metrics. The proposed method is differentiable & invertible, well-suited for deep learning architectures, robust to noise and outliers, computationally efficient, and is expressive and flexible enough to capture complex patterns. Furthermore, a closed-form solution was developed for the gradient of these diffeomorphic transformations, which allows an efficient search in the parameter space, leading to better solutions at convergence. Leveraging the benefits of these closed-form diffeomorphic transformations, this thesis proposes a suite of advancements that include: (a) an enhanced temporal transformer network for time series alignment and averaging, (b) a deep-learning based time series classification model to simultaneously align and classify signals with high accuracy, (c) an incremental time series clustering algorithm that is warping-invariant, scalable and can operate under limited computational and time resources, and finally, (d) a normalizing flow model that enhances the flexibility of affine transformations in coupling and autoregressive layers. Taken together, these advancements demonstrate the versatility and potential of closed-form diffeomorphic transformations for a range of time series applications. In summary, this thesis aims to enhance time-series tasks such as alignment, averaging, classification and clustering by leveraging the power of fast, efficient, parametric \& diffeomorphic warping methods.
- Additive Manufacturing & Topology Optimisation of electrical machines.(Servicio de Publicaciones. Universidad de Navarra, 2023-07) Lizarribar Carrillo, B.(Borja); Martínez-Iturralde-Maiza, M. (Miguel); Prieto-Rocandio, B. (Borja)In recent years, additive manufacturing has been gaining ground among traditional manufacturing methods in many different applications. This is mainly due to its ease of prototyping, material savings and above all the highly complex geometries that can be achieved. Despite its many advantages, the use of additive manufacturing in electrical machines has been limited. The lack of maturity of the manufacturing processes and the increase in losses when applying additive approaches to some parts, mainly stators, have led little implementation in the field of electrical machines. The geometric freedom offered by additive manufacturing paves the way for the use of novel optimisation techniques, such as topology optimisation. Traditionally, topology optimisation has not been widely used because the complex geometries obtained by these algorithms are not easily produced by traditional manufacturing methods. However, thanks to additive manufacturing, topology optimisation has started to be used, mainly for mechanical applications and it has demonstrated its ability to reduce weight without compromising mechanical stability in several cases. Nevertheless, topology optimisation for physics other than mechanical has not been thoroughly studied, let alone for the simultaneous optimisation of different physics. In view of the scenario described, this thesis investigates the benefits that additive manufacturing and topology optimisation can bring to the design of electrical machines. For the additive manufacturing of electrical machines, a thorough literature review is carried out on the application of this manufacturing approach to different parts of electrical machines, active and non-active. Two main case studies are presented: in the first, the rotor, the shaft, the stator and the housing of an aerospace actuator are manufactured by L-PBF in FeCoV and the assembled actuator is tested. In the second case, a rotor is manufactured in FeSi by LP- DED. Finally, not considered an additive manufacturing case study itself in this thesis, but two electrical conductor prototypes are manufactured in CuCr1Zr and an additional geometry is manufactured in AlSi10Mg by L-PBF. With regard to topology optimisation of electrical machines, a description of the main topology optimisation methods used is given and their application to electrical machine components is presented. In a similar way to additive manufacturing, two case studies are analysed. First, a novel multiphysics - mechanical and electromagnetic - topology optimisation method is presented, in which the rotor and the stator of a permanent magnet motor are optimised simultaneously. This method is compared with two methods found in the literature, SIMP and on-off. Secondly, another topology optimisation method is described and applied to an electrical conductor model. The proposed algorithm involves a multiphysics - thermal and electromagnetic - hybrid parametric topology optimisation approach. Two reduced length geometries and one full length prototype are built via additive manufacturing. Finally, conclusions regarding additive manufacturing and topology optimisation of electrical machines are drawn from the work presented in this thesis. Future lines of work for additive manufacturing and topology optimisation of electrical machines are also presented.
- Enzyme promiscuity approaches to predict metabolic pathways in the human gut microbiota.(Servicio de Publicaciones. Universidad de Navarra., 2023-07) Balzerani, F. (Francesco); Planes-Pedreño, F.J. (Francisco Javier)Metabolism is the whole set of reactions that take place in an organism. Understanding how humans digest food and absorb nutrients is complex and challenging. The role of the different communities of microorganisms that reside in the human being, called microbiota, has gained growing interest in the last decade. In particular, the human gut microbiota has been linked to several diseases and syndromes, which has largely motivated the study of the interaction between diet, gut microbiota metabolism and health from very different angles. In the area of System Biology, the release of high-quality metabolic reconstructions of the human gut microbiota has opened new avenues to address this question. However, they are still in their infancy, and further developments are required to obtain more comprehensive metabolic models. In this doctoral thesis, the main objective is to develop computational tools to complete these metabolic reconstructions and improve the annotation of important dietary compounds degraded by the human gut microbiota. This work is divided into two parts. The first part focuses on System Biology and enzyme promiscuity with the aim of improving metabolic reconstructions of the human gut microbiota. We present AGREDA v1.1, a new metabolic reconstruction that includes predicted chemical transformations from a popular enzyme promiscuity algorithm called RetroPath RL. The new draft AGREDA v1.1 allows a better representation of phenolic compounds, a relevant class of compounds for both nutrition and health. By means of untargeted metabolomics, we validated the ability of AGREDA v1.1 to predict output microbial compounds in the human gut microbiota. Then, we introduce PROXIMAL2, a novel enzyme promiscuity algorithm that overcomes the limitations of PROXIMAL, a promising previously published algorithm, but it was unable to be applied to our problem of phenolic compound degradation in the human gut microbiota. In particular, PROXIMAL2 overcomes the dependency on KEGG database and extends the scope of application to more complex enzymatic reactions. We obtained complementary results to the ones obtained with RetroPath RL, proposing new tentative degradation pathways for phenolic compounds in the human gut microbiota. The second part of this doctoral thesis is focused on the development of tools to analyze the 16S rRNA gene-sequencing data using metabolic reconstructions of the human gut microbiota. Specifically, we present a novel Python package called q2-metnet, which contextualizes 16S rRNA gene-sequencing data into metabolic networks and provides a quantitative score for reactions and subsystems. By means of different metabolic reconstructions, users can characterize data samples and extract functional features to differentiate between clinical conditions.
- Mechanical and morphological modulation of electrospun polymeric scaffolds for tissue engineering applications(Servicio de Publicaciones. Universidad de Navarra, 2023-06) Bikuna-Izagirre, M. (María); Aldazabal, J. (Javier); Paredes-Puente, J. (Jacobo)Electrospinning technologies herald the arrival of a new era in which previously unthinkable scaffolds for tissue engineering applications will be solved efficiently. However, electrospinning techniques, like solution electrospinning and melt electrowriting are held down by fabrications parameters, technology limitations, and the application perse. The science of scaffolding fabrication seeks to mimic the extracellular matrix of a particular tissue in ways that mitigates the damage or enables its pathophysiological study. Thenceforce, scaffolds have the primordial role of not only supporting the cells, but to replicate as close as possible the native extracellular matrix, taking into consideration the biocompatibility, biodegradability, morphology and mechanical properties. The last two properties are pivotal in the scaffold ́s outcome, as cells communicate with the environment, and behave in response to external signals. In context of scaffolds ́ assembly, electrospinning fabrication parameters should be correctly modulated, to ensure an appropriate cellular environment. In this dissertation we attempt to tackle this concern relying on solution electrospinning and melt electrowriting techniques. As potential tissue engineering applications, the recreation of an artificial human trabecular meshwork and a skeletal muscle platform are developed. The mechanical and morphological requirements of each tissue are evaluated and fabrication parameters adapted. An in vitro human trabecular meshwork scaffold was developed and validated with human trabecular meshwork cells ́ behavioral studies. With the development of a perfusion bioreactor human trabecular meshwork cells react to medicaments inducing measurable pressure changes. Finally, an attempt for skeletal muscle platform was made. This first approach enabled us the optimization of the process for next attempts.
- Bidaiari-trenen esekidura pneumatikoaren karakterizazio eta modelizazio dinamikoa.(Servicio de Publicaciones. Universidad de Navarra, 2023-06-13) Mendia-García, I. (Irati); Gil-Negrete, N. (Nere); Pradera-Mallabiabarrena, A. (Ainara)An essential benefit of using virtual homologation in railway vehicles is the reduction of the high costs associated with on-track tests, which can be reached by analysing the performance of a railway vehicle under various operating conditions and, thus, by enhancing the vehicle’s design so as to optimise its performance under particular circumstances. In fact, as railway vehicle speed increases, the vehicle’s dynamic performance is affected. This demands the development of validated and accepted models that incorporate the influence of all vehicle components, including the wheel-rail contact, bogie frame, suspension elements, carbody, etc. In order to ensure comfortable trips the secondary suspension system aims to reduce and mitigate the vibration transmission. This suspension element is a complex component composed by a pressurized reinforced elastomeric bellows, which lies into a rubberlike emergency spring, connected though a pipeline or an orifice to a reservoir. In this thesis, the viability of the current secondary suspension models into higher frequencies and different working directions is investigated. Where necessary, new models are proposed to extend the frequency range up to 200 Hz (structural-borne vibration transmission frequency range) or incorporate existing non-linearities. Firstly, the available modelling techniques of air spring type pneumatic suspensions are evaluated according to the frequency range they cover, the component number included (bellows, pneumatic system, full secondary system) and the nonlinearities they can account for. FEM models arise as the most suitable modelling technique for the non-linear multidirectional and multiphysic and high frequency range replica of the secondary suspension system. After all, the suspension system is composed by several elements, including mainly a rubber-cord composite bellows, a rubberlike emergency spring and a moving air mass inside the pneumatic system, which results in a highly non-linear suspension element. One aspect that has been covered is the implementation into FEM models of non-linear behaviour of rubberlike elements (the nearly incompressible behaviour, non-linear elasticity, frequency and excitation-amplitude dependencies), the need of experimental characterization and model calibration. Secondly, the singularities of the bellows, the pneumatic system and the full secondary suspension are investigated separately. As far as the air spring is concerned, a bellows’s FEM model developed in ABAQUS is proposed, which is validated with experimental data. It incorporates the uniaxial reinforcements, the coupling between internal pressure and structural deformation, and the polytropic heat exchange definition between the inner air and the environment. Moreover, based on four surface response functions of the axial and transversal static stiffness and first axial and transversal vibration modes, which are function of seven construction parameters, a design tool is suggested. As an interesting outcome, the suspension system shows vibration modes bellow 200 Hz, in the frequency range which structure-borne vibration transmission takes place. Afterwards, the axial dynamic stiffness of the pneumatic suspension, more precisely of a single-lobe air spring connected to a reservoir via a pipeline is investigated, up to 400Hz. After carrying out an exhaustive experimental campaign, an enhanced FEM model is developed which incorporates the resonances due to the air flow between the bellows and the reservoir, the resonances due to the formation of standing waves in the pipeline and the resonances due to structural dynamics of the bellows. Up to date, available modelling techniques disregard the effect of the auxiliar volume (reservoir and pipeline) above 20 Hz. Nevertheless, this research shows, although that in a lesser extent, it modifies the dynamic performance of the suspension system. In addition, structural modes of the air spring can compromise the isolation above 20 Hz. Finally, a non-linear multiphysic FEM model of the full secondary suspension element, which incorporates the emergency spring to the pneumatic system is developed. Static and dynamic results up to 20 Hz of the FEM model are compared with available experimental data and afterwards, the model is extended up to 300 Hz. The dynamic performance of the suspension system in a pure axial, pure transversal or pure roll movements is predicted. In the three directions the model also predicts resonance frequencies below 200 Hz, which might compromise the isolation. As an application, based on the developed FEM model and with the advantage of avoiding any experimental test, the input parameters of the secondary suspension system model of multibody simulations are derived.
- Análisis del funcionamiento de tanques DAF con y sin lamelas a través de la modelización de la hidrodinámica mediante técnicas CFD.(Sevício de publicaciones. Universidad de Navarra, 2023-03) Hlukhov, D. (Dmytro); Sanchez-Larraona, G. (Gorka); Rivas-Nieto, A. (Alejandro)La flotación por aire disuelto (DAF, por sus siglas en inglés) es un proceso de tratamiento de agua que se utiliza para separar las partículas sólidas en suspensión del líquido. El proceso se basa en la adición de microburbujas de aire a la corriente de agua, lo que hace que las partículas sólidas se adhieran a las burbujas, formen los agregados de menor densidad y floten hacia la superficie, donde luego se separan por medios mecánicos. Este proceso se utiliza comúnmente en la industria del tratamiento de aguas residuales para eliminar contaminantes como aceites, grasas, sólidos suspendidos y materia orgánica. También se utiliza en la producción de alimentos y bebidas, minería y la fabricación de papel, entre otros. El DAF es altamente eficiente eliminando materia orgánica y se puede adaptar a una amplia gama de aplicaciones. Sin embargo, requiere de equipos especializados y un control cuidadoso de las condiciones de operación para lograr una separación óptima de sólidos y líquidos. En esta tesis se ha analizado la hidrodinámica de un tanque DAF y para ello, se ha desarrollado un modelo de Dinámica de Fluidos Computacional (CFD) de un DAF de agua y microburbujas, que se ha analizado utilizando enfoques mono-diámetro y multi- diámetro de estas últimas. Se ha introducido un nuevo concepto de Diámetro Crítico, que resultó ser una herramienta de bastante interés para los estudios CFD de los DAF. Se ha demostrado que la utilización de tamaños de burbuja de aire superiores o inferiores al valor del Diámetro Crítico afecta significativamente al contenido de aire, a la estructura del flujo y al límite del manto de agua blanca en el interior de la Zona de Separación (ZS). Los resultados obtenidos mediante el enfoque mono-diámetro estaban en línea con los datos experimentales en las condiciones de trabajo en las que la concentración de aire en la ZS tenía un valor casi constante. Sin embargo, no ha podido predecir el caso de la disminución progresiva de aire por debajo de la mitad de la altura de la ZS. Se requería un efecto combinado de burbujas con diferente velocidad de ascenso para reproducir una curva de perfil de aire suave, como la que se ha medido experimentalmente. En este contexto, el enfoque multi-diámetro ha demostrado ser adecuado para reproducir la estructura estratificada de flujo. Se ha comprobado que la solución obtenida mediante simulación en estado estacionario era suficientemente parecida a la obtenida mediante simulaciones transitorias para un tiempo de flujo suficientemente largo. Además, el modelo multifásico de Mezcla ha resultado ser adecuado para las simulaciones de tanques DAF y ha producido resultados similares a los del modelo de Euler-Euler. Al modelo previamente validado del tanque a escala piloto DAF se le ha añadido el modelo de 20 lamelas planas en la Zona de Separación. Se ha examinado el rendimiento del tanque DAF con lamelas (L-DAF) en dos condiciones de operación que se han identificado previamente, en las que el flujo estaba cortocircuitado o estratificado en ausencia de lamelas. Además, se evaluó la mejora en la eficiencia de eliminación de burbujas obtenida mediante la inclusión de lamelas en cada caso. En segundo lugar, se realizó una investigación en detalle del flujo que se forma en la Zona de Separación como resultado de la inserción del pack de lamelas en esa zona del tanque. El complejo flujo tridimensional observado entre las dos regiones, entre el manto de burbujas y el agua clarificada, está condicionado por la gran diferencia de densidad, que está causada por la presencia de lamelas. El análisis del flujo reveló un mecanismo hasta ahora desconocido en el que el gradiente de densidad desempeña un papel decisivo a la hora de impedir que las burbujas atraviesen las lamelas y acaben escapando con el agua clarificada. Se ha evaluado el impacto de la carga hidráulica en el rendimiento del L-DAF y su eficacia en la eliminación de burbujas. Se ha comprobado que las variaciones más significativas se producen en el flujo entre las lamelas, especialmente en la profundidad del manto de burbujas entre ellas, que aumenta a medida que lo hace la carga hidráulica. Sin embargo, el aumento de la profundidad del manto no es uniforme, sino que es más pronunciado en las primeras lamelas, entre las que pasa más caudal que entre las últimas, debido al mayor gradiente de densidad respecto al agua clarificada que existe en estas. Por último, se ha realizado un análisis paramétrico para evaluar el impacto de la variación de la longitud y de la separación entre lamelas en el comportamiento del tanque. Se ha analizado el efecto de estas modificaciones en la distribución de caudal, la estructura de flujo y la posición media de la capa de burbujas con respecto al suelo. También se ha evaluado el efecto de la reducción de la longitud de las lamelas y el aumento del espaciado interlamelar en el rendimiento de eliminación de burbujas de aire. Se ha encontrado que al incrementar el valor del conocido como “parámetro de la lamela”, que depende estas últimas dimensiones, se logra aumentar el caudal máximo que el tanque puede procesar y que, además coincide con los resultados experimentales de otros autores.
- Interpretable precision medicine for acute myeloid leukemia(Servicio de Publicaciones. Universidad de Navarra, 2023-03) Gimeno-Combarro, M. (Marian); Rubio-Díaz-Cordovés, Á. (Ángel); Carazo-Melo, F.(Fernando)Precision medicine (PM) is a branch of medicine that defines a disease at a higher resolution using genetic and other technologies to enable more specific targeting of its subgroups. Because of its uses in clinical treatment and diagnostics, this field exemplifies the modern era of medicine. PM looks for not just the right drug, but also the right dosage and treatment regimen. PM encounters a variety of challenges, which will be explored in this dissertation. Large-scale sensitivity screens and whole-exome sequencing experiments (WES) have fostered a new wave of targeted treatments based on finding associations between drug sensitivity and response biomarkers. These experiments with the aid of state-of-the-art artificial intelligence (AI) algorithms are opening new therapeutic opportunities for diseases with unmet clinical needs. It has been proved that AI is capable of predicting novel personalized treatments based on complex genotypic and phenotypic patterns in tumors. The scientific community should make an effort to make these algorithms to be interpretable to humans so that the results could be easily approved by the medical regulators. The purpose of this thesis is to apply AI algorithms for precision oncology that are highly accurate, while guaranteeing that the predictions are interpretable by humans. This work is divided in three main sections. The first section comprises a new methodology to increase the predictive power of the discovery of novel treatments in large-scale screenings by exploiting that some biomarkers tend to appear in many treatments. This fact is called hub effect in gene essentiality (HUGE). Content of this section was published in [1]. The second section contains a novel interpretable AI method -called multi-dimensional module optimization (MOM)- that associates drug screening with genetic events and proposes a treatment guideline. Content of this section was published in [2]. Finally, the third section includes a detailed comparison of different recently published algorithms that attempt to overcome the barriers proposed by today's precision medicine. This study also includes two novel algorithms specifically designed to solve the challenges of applicability to clinical practice: Optimal Decision Tree (ODT) and Multinomial Lasso. The characterization of Interpretable Artificial Intelligence as approach with strong potential for use in clinical practice is one of the study's most significant achievements. We presen tunique methods for PM that are highly interpretable, and we summarize the needs that could be considered for constructing interpretable AI. We are confident that this method will transform the way PM is addressed, bridging the gap between AI and clinical practice.
- An enhanced integrity multisensor Fusion for a reliable seamless navigation.(2023-01) Zabalegui-Landa, P. (Paul); Adin-Marcos, I.(Iñigo); De Miguel-Aramburu, G.(Gorka)Since its first applications in the late 20th century, GNSS technology has been deployed by the world’s technologically advanced countries in multiple fields, from fleet monitoring to sport-related topics. This massive deployment has led to new use cases that may not have been expected during the definition of said technology. Different error sources, such as interferences, jamming, signal attenuation due to indoor or urban canyon navigation, and signal-blocking objects may degrade the performance of GNSS-based navigation. Thus, standalone GNSS systems may not fulfil all the requirements a certain scenario might ask for. This has resulted in the research of alternative or supplementary methods to solve the aforementioned issues, such as multisensor navigation. This has become one of the main alternatives to GNSS standalone navigation, as it has been shown in the literature that it can result in an improvement in navigation in terms of availability or continuity, for example. Human-life involvement and high-cost freight transportation, among other factors, have attracted the attention of the users to the definition of a measure of trust that is placed in the correctness of the information supplied by the navigation systems; also called integrity. This concept is employed, among others, to enable the system to detect if it is trustable for navigation, provide warnings, and even act consequently. In this dissertation, we analyze, first, the design of an online multisensory navigation algorithm as a solution to the issues GNSS suffers especially in urban and indoor environments. Moreover, a two-stage integrity-ensuring method is analyzed, being this second algorithm a tailored complementary feature of the proposed navigation one.
- Design and analysis of a novel stator oil-flooded cooling system for aircraft electrical machines with hairpin windings(Servicio de Publicaciones. Universidad de Navarra, 2022-12) Paredes-Puente, J. (Jesús); Elósegui-Simón, I. (Ibón); Prieto-Rocandio, B. (Borja)Growing environmental and economic concerns are driving various sectors to improve energy efficiency. In sectors such as aeronautics, this is achieved through electrification and weight reduction, which necessarily involves replacing conventional pneumatic, mechanical and hydraulic systems with electric drives. In addition, electric machines also offer better efficiency and require less maintenance. Then minimizing the weight of electric machines is therefore essential to reduce fuel consumption and permanent magnet machines offer the highest power density. The use of hairpin windings can help reduce weight, due to their high slot filling factor and short winding ends, and have already demonstrated their potential in sectors such as automotive. Another key factor in improving power density are cooling systems. Among all the liquid cooling systems, stator oil-flooded solutions are the most promising, both for their cooling capacity and for the improved electrical isolation between conductors that make them suitable for aeronautical applications. This dissertation deals with the combination of these two aspects, hairpin winding and flooded stator cooling for aircraft applications, and presents a novel cooling system that has been patented. For this purpose, the proposed solution is described and the advantages of this cooling arrangement are detailed. The equations defining its behavior are presented, its potential is experimentally evaluated by using a motorette and design criteria are provided. Subsequently, the presented concept is applied on an industrialized and field-tested machine for heavy-duty off-road vehicles in substitution of a water-jacket cooling arrangement, ending with a prototype in order to assess gains in power density. Next, the novel cooling arrangement is applied on a conceptual design for an aeronautical application. Finally, the main results of this dissertation are summarized and the main future research lines are outlined.
- Study of grain growth in Nd-Fe-B powders obtained by gas atomization.(Servicio de Publicaciones. Universidad de Navarra, 2022-01) Sarriegui-Estupiñan, G.C. (Gabriela Carolina); Martín-García, J.M. (José Manuel)--