Author: BOTERO POSADA, SIMÓN ANDRÉS
Thesis file: (contact the Doctoral School to confirm you have a valid doctoral degree and to get the link to the thesis)
Programme: DOCTORAL DEGREE IN SUSTAINABILITY
Department: University Research Institute for Sustainability Science and Technology (IS.UPC)
Mode: Normal
Deposit date: 25/07/2022
Reading date: 28/11/2022
Reading time: 18:00
Reading place: ETSECCPB, Sala C1-002, C/ Jordi Girona 1-3, Campus Nord Barcelona
Thesis director: ALVAREZ DEL CASTILLO, JAVIER | RÍOS OSORIO, LEONARDO ALBERTO
Committee:
     PRESIDENT: NICHOLLS ESTRADA, CLARA
     SECRETARI: MORATO FARRERAS, JORDI
     VOCAL NO PRESENCIAL: LAVIÑA GÓMEZ, FRANCISCO JAVIER
Thesis abstract: Agro-food systems in indigenous communities, due to their characteristics of complexity and adaptability, must be studied from a multidimensional perspective. In this class of systems, economic, political, environmental, social and cultural factors intervene, and among these there are multiple systemic interrelationships that affect the processes of production, processing, distribution and consumption of foodThe text includes three chapters, related to the specific objectives of this research in which a methodological research proposal is presented, from which it was possible to build a theoretical model of an agri-food system based on socio-ecological resilience, as an essential characteristic for the Sustainability of the agri-food systems of the Ziparadó community in the Embera Dóbida indigenous reservation of Tanela, located in the municipality of Unguía, department of Chocó, ColombiaThe first chapter corresponds to the documentary stage of the Participatory Evaluative Research -IEP- (acronym in Spanish), where a systematic review of the scientific literature was carried out in accordance with the PRISMA declaration. The selected documents presented topics related to rural communities and their agri-food systems, alternative food networks, farmers' markets, community agriculture, among others. Those texts in which reference was made to industrial-type agri-food systems were excludedIn the second chapter, the diagnostic stage of the (IEP) was developed, which consists of two moments: Exploratory and Deepening, in which, through mixed methods, a characterization of the functional and structural aspects of six agri-food systems of indigenous families was carried outThe third chapter established a theoretical model of a sustainable agri-food system for the indigenous community of Ziparadó, based on socio-ecological resilience and applying the methodology of complex systems.

Author: ORTEGA CARRASCO, CRISTOBAL
Thesis file: (contact the Doctoral School to confirm you have a valid doctoral degree and to get the link to the thesis)
Programme: DOCTORAL DEGREE IN COMPUTER ARCHITECTURE
Department: (DAC)
Mode: Normal
Deposit date: 28/09/2022
Reading date: 28/11/2022
Reading time: 17:00
Reading place: fib_ Videoconferència ( meet.google.com/kvv-tffu-oyo ) -FIB - Sala de Reunions Tillers 3 planta - Campus Nord - Barcelona
Thesis director: MORETÓ PLANAS, MIQUEL | ÁLVAREZ MARTÍ, LLUC
Committee:
     PRESIDENT: SAHUQUILLO BORRÁS, JULIO
     SECRETARI: CAZORLA ALMEIDA, FRANCISCO JAVIER
     VOCAL: VEGA, AUGUSTO JAVIER
Thesis abstract: Every new generation of high performance computing (HPC) processors brings higher complexity to boost performance:the increasing number of cores within the processor itself, cores with the ability to run multiple threads (i.e. Simultaneous Multithreading or SMT), increased available memory bandwidth, and multiple architectural improvements. On top of this, HPC systems are typically composed of multiple sockets with Non-Uniform Memory Access (NUMA) to increase the number of available processors in the system. As a result, the application running on the different processors experience contention in the acceess to shared resources, which might lead to individual application's performance degradation.Modern HPC systems include hardware knobs to adapt architectural parameters to the workload demands with the goal of improving performance, energy, or power consumption. However, a default hardware knob configuration is set by the hardware designers and the responsibility of finding the best hardware knob configuration is left to the user, which is an overwhelming process due to the large search space and the different resource demands of applications. In addition, setting a hardware knob independently of others can result in suboptimal configurations and sometimes to conflicting decisions that jeopardize system power-performance efficiency.Furthermore, the number and heterogeneity of compute devices, even within a single compute node, has been steadily on the rise. Conflicting decisions can appear when incorporating multiple diverse accelerators within a node to provide efficient performance growth through specialization (e.g. graphics processing unit or GPU), since complex heterogeneous systems with multiple discrete accelerators cannot afford to fully power all the devices simultaneously.In a HPC system, the number of discrete devices that can run simultaneously at their highest frequency is limited by the globally-imposed power cap. Current systems incorporate a centralized power management unit that statically controls the distribution of power among the devices within the node. However, such static distribution policies are unaware of the dynamic utilization profile across the devices, which leads to power allocations that end up degrading system throughput performance. The problem is particularly acute in the presence of heterogeneity since type-specific performance-boost capabilities cannot be leveraged via utilization-agnostic static power allocations.The main goal of this Thesis is to leverage runtime system data to intelligently coordinate hardware knobs while taking into account interactions between each other at application level, hardware level, and system level. This Thesis proposes multiple methods to dynamically coordinate multiple hardware knobs to maximize system power-performance efficiency.At application level, we propose an infrastructure for shared memory parallel programming models that configures the different hardware knobs available in the architecture. During execution time, the best hardware configuration is discovered for different regions of the application. This Thesis proposes two mechanisms to discover the best hardware knob configuration: (1) libPRISM: an exploration-based approach useful for reduced search spaces that needs almost no previous knowledge of the system, and (2) MARK: an intelligent approach based on machine learning (ML) to search large design spaces.At hardware level, this Thesis demonstrates that a hardware knob can benefit from fine-grained coordination between its own possible configurations. At system level, this Thesis proposes a hardware/software power distribution mechanism that maximizes the performance of power-constrained heterogeneous systems by leveraging system information to distribute power among all the devices.This thesis demonstrates that fine-grained hardware coordination is needed at different levels to maximize energy-efficiency.

Author: VILLANUEVA RODRÍGUEZ, MANUEL
Thesis file: (contact the Doctoral School to confirm you have a valid doctoral degree and to get the link to the thesis)
Programme: DOCTORAL DEGREE IN CONSTRUCTION ENGINEERING
Department: (DECA)
Mode: Confidentiality
Deposit date: 29/09/2022
Reading date: 29/11/2022
Reading time: 11:00
Reading place: ETSECCPB_ Edifici C1, aula 002. C/ Jordi Girona 1-3. Campus Nord. Barcelona 08034
Thesis director: TURMO CODERQUE, JOSE | RAMOS SCHNEIDER, GONZALO
Committee:
     PRESIDENT: PELLICER ARMIÑANA, EUGENIO
     SECRETARI: FORCADA MATHEU, NURIA
     VOCAL: NEAMTU NEAMTY, LAURENTIU
Thesis abstract: The construction sector has a big impact on our society but compared with other industries, the productivity is historically much lower. The high uncertainty in construction projects could be mentioned as one of the main reasons but there are plenty of motives that make this industry full of complexities and therefore hard to manage. Over the years this situation has forced the industry to discover new systems which could help in having more control and efficiency in construction projects. As a result, during the ¿90s, two main approaches appeared. The first one was a guide created by the Project Management Institute (PMI) called the Project Management Book of Knowledge (PMBOK). They aimed to unify some of the Project Management (PM) lessons learned and they have been updating this guide since then. In the second one, Professor Koskela also challenged the Construction Industry to explore and adopt principles of ¿Lean Production¿ in Construction and introduced the theory of ¿transformation, flow and value¿ named ¿Lean Construction¿ (LC).Although the situation has improved since then, there are still some efforts to be made in order to enhance the effectiveness of construction PM. Thus, the main aim of this research is to demonstrate that the implementation of the Lean Construction methodology in building projects improves project success both from a project management and a customer satisfaction perspective.To achieve it, this investigation proposes both a literature study and a case study. The first one consists of a theoretical framework for identifying how some LCTTs could improve PM performance. For this, a literature review of 61 international papers, from 2010 to 2020 has been quantitatively analyzed.The second part of the research method consists of a retrospective observational study. Using the tools Pattern-matching and Cross-case synthesis, two real case studies developed by Cognita have been compared quantitatively. One of them has been carried out with a Lean approach (ELIS Montevida) and the other one with a traditional approach (BSB Nexus). To compare them considering the most recent project success criteria, this thesis proposes for the first time the concept of the `Lean Triangle¿ which evaluates both the project management performance (iron triangle) and customer satisfaction (lean thinking).As a result, regarding the literature study, Project Managers and Lean practitioners may be able to understand quantitatively the relationship between LC and PM and the benefits beyond this partnership. This will allow them to select the best LCTTs considering the PMKAs that they want to improve. Regarding the case study analysis, quantitative information with statistical relevance has been obtained and has allowed us to demonstrate that the implementation of the Lean Construction methodology in building projects developed by Cognita improves project success both from project management and a customer satisfaction perspective.Finally, because of the aforementioned analysis, firstly theoretical and later applied to real cases, a tailored methodology has been delivered to Cognita through a Lean project management guideline. This will allow Cognita to manage future projects more efficiently throughout a workflow to use during the entire project life cycle. Thanks to this, Cognita will be able to tackle some of the construction problems and improve the project management performance, especially through the iron triangle (cost, time and scope). As a result, they could manage projects more quickly, cheaply and with better customer satisfaction. As a consequence of this investigation, Cognita will be able to develop projects worldwide more successfully with a direct positive impact on their business.

Author: GOMEZ ACERO, DANIEL HERNANDO
Thesis file: (contact the Doctoral School to confirm you have a valid doctoral degree and to get the link to the thesis)
Programme: DOCTORAL DEGREE IN ELECTRICAL ENGINEERING
Department: (DEE)
Mode: Normal
Deposit date: 29/09/2022
Reading date: 01/12/2022
Reading time: 16:00
Reading place: ETSEIB_ Aula Schneider - Dept. Eng. Elèctrica. ETSEIB - Campus Sud
Thesis director: GOMIS BELLMUNT, ORIOL | CHEAH MAÑÉ, MARC
Committee:
     PRESIDENT: SAAD, HANI
     SECRETARI: PRIETO ARAUJO, EDUARDO
     VOCAL: KISH, GREGORY
Thesis abstract: High voltage direct current (HVDC) technologies have been used for bulk power transmission over long distances since the 1950s. These technologies have proven to be the most cost-efficient compared to the high voltage alternate current (HVAC) for some applications such as offshore power transmission, connecting remote loads or generation, and the interconnection of non-synchronized grids.In recent years, the study of HVDC grids has been of interest in some research projects but, its development is still uncertain. The de grid can be planned beforehand or it can use the installed lines. But, from the installed HVDC projects, it can be identified different operating voltages, used technologies, and line topologies. There are two HVDC technologies: the line commutated converter (LCC), and the voltage source converter (\/SC). Four different line topologies are identified: asymmetric monopole, symmetric monopole, bipole, and rigid bipole. Developing a de grid interconnecting isolated lines with different characteristics cannot be possible without an intermediary device: the de-de converter.This thesis studies the de-de converters interconnecting HVDC lines with different characteristics. These converters can be seen as the equivalent of ac transformers in de applications because they are capable to adapt the voltage difference between two de systems. These converters are also capable to adapt the line topology and differenttechnologies. The power electronics required for these de-de converters provide increased control flexibility used to supply additional ancillary services that the classical transformers cannot furnish.Three de-de converter topologies are modeled and simulated for the interconnection between a bipole and a symmetric monopole. The front-to-front modular multi-level converter (F2F-MMC) is chosen as the reference because it represents state­ of-the-art technology. The second converter is the de-de MMC (de-MMC) because of the topological similarityto the MMC. Then, a third converter is proposed and studied as a result of this thesis, the asymmetric de-de converter (ADCC). A set of simulations are performed for multiple operating points and faults scenarios. Then, the converters are compared quantitatively and qualitatively. The results and analysis are used to conclude and bring some perspectives for future works.

Author: JANER ANGELET, MARCEL
Thesis file: (contact the Doctoral School to confirm you have a valid doctoral degree and to get the link to the thesis)
Programme: DOCTORAL DEGREE IN MATERIALS SCIENCE AND ENGINEERING
Department: Department of Mining, Industrial and ICT Engineering (EMIT)
Mode: Normal
Deposit date: 03/10/2022
Reading date: 01/12/2022
Reading time: 12:00
Reading place: ETSEM_Sala d'Actes de l'Escola Politècnica Superior d'Enginyeria de Manresa.
Thesis director: RIERA COLOM, MARIA DOLORES
Committee:
     PRESIDENT: CASELLAS PADRO, DANIEL
     SECRETARI: FERRER REAL, INÉS
     VOCAL: LLORCA ISERN, NURIA
Thesis abstract: Ultrasonic moulding is a promising technology that could be used as a substitute for conventional injection moulding techniques. This relatively new technology has a lower energy consumption, and it could be a sustainable altemative in an industrial environment. In addition, the supply of material processed in ultrasonic moulding is delivered shot-by-shot, which makes this technology very adequate to process short batches of samples without wastage. However, up to now, ultrasonic moulding technology has not been adopted in industrials environrnents due to its lack of robustness and poor repeatability found in its results. In addition, the little knowledge about the influence of the process parameters in the polymer melt is also a handicap for the industrial operator.In this dissertation, applied research and numerical simulation was carried out to improve the ultrasonic moulding technology and to deepen the knowledge ofthe process to promote its industrialization. With this aim, this thesis presents three main areas of work.First, a study of the evolution of ultrasonic moulding machines and configurations was perforrned. This review analysed the experiments published in the literature along with the main conclusions and drawbacks found.On that basis, the developrnent and validation of a new configuration for ultrasonic moulding was done, leading to a great improvement over the performance of the method in terms of repeatability and reduction of impurities in the samples. This new configuration was applied to process polyexymethylene and cyclic olefin polymer, and the results were comparad to the conventional injection moulding. This analysis reveals that the developed rnethod is able to correctly process polymers in a repetitiva way, making ultrasonic moulding a reliable technology for the industry.Finally, research was carried out to study the viscoelastic behaviour of ultrasonic heating of polypropylene cylinders. Results obtained from the numerical simulation of the process were comparad to experimental measurements done with an infrared carnera. The analysis of the results showed an inhomogeneous temperatura distribution along the cylinder and different heating steps can be identified over time. In addition, the comparison between the nurnerical and the experimental results showed that the interaction between the sample and the rnould directly influences the temperatura distribution along the cylinder. Finally, the effect of the main parameters in ultrasonic heating was obtained, both numerically and experimentally, and comparad.As a result, the research perforrned in this dissertation improves the applicability of ultrasonic moulding technology in industrial environments by increasing its repeatability and robustness, and contributing to a better understanding of its main parameters.