Universitat Politècnica de Catalunya
UPCommons. Portal del coneixement obert de la UPCNot a member yet
290727 research outputs found
Sort by
Promoting electrochemical reactions with dual-atom catalysts for high-rate lithium–sulfur batteries
No full textSulfur cathodes offer a promising solution for high-energy-density, cost-effective, and sustainable energy storage. However, their practical application is limited by sluggish and complex multistep sulfur redox reactions (SRRs), involving both electrochemical and chemical processes. Herein, it is demonstrated that accelerating electrochemical processes, particularly Li2S nucleation, over competing chemical pathways is fundamental to minimizing sulfur loss and achieving high-rate performance. To this end, a scalable and cost-effective strategy is presented for synthesizing a series of 3d transition metal–bismuth (TM–Bi) atomic pairs anchored on carbon nitride (CN) and investigate their potential to activate SRRs in lithium-sulfur batteries (LSBs). An initial screening identifies Ni–Bi/CN and Co–Bi/CN as highly effective in improving rate performance. Detailed analysis shows these catalysts promote direct electrochemical transitions and rapid Li2S nucleation over competing chemical reactions, enabling high charge–discharge rates while preventing active material loss and enhancing stability. Electrochemical analysis, density functional theory, and operando spectroscopy reveal that TM-Bi pairing shifts d-band states closer to the Fermi level and modulates HOMO–LUMO levels, promoting lithium polysulfide (LiPS) interaction and facilitating efficient charge transfer. These findings offer valuable insights for designing advanced catalysts for LSBs and broader electrocatalytic applications.IREC and HZB acknowledge financial support from the European Union (Grant agreement No. 101104006 ─ HEALINGBAT ─ HORIZON-CL5- 2022-D2-01). J.Y. and C.H. thank the China Scholarship Council for the financial support. C.Y.Z. and J.Z. acknowledge the support of the Super- computing Center of Lanzhou University, China. I.P.H. acknowledges fund- ing from AGAUR-FI scholarship (2023FI-00268) Joan Oró of the Secre- tariat of Universities of the Generalitat of Catalonia and the European SocialPlus Fund. I.M. acknowledges financial support from the European Union’s Horizon Europe research and innovation programme under the Marie Skłodowska-Curie grant agreement No. 101081337. XAS experi- ments were performed at the beamline-22-CLÆSS (Proposal Number: 2024068488), [50] and operando XRD was performed in beamline BL04- MSPD (Grant Number: 2023097750) at ALBA Synchrotron, Spain. ICN2 acknowledges funding from Generalitat de Catalunya 2021SGR00457. This study is part of the Advanced Materials programme and was supported by MCIN with funding from the European Union NextGenerationEU (PRTR-C17.I1) and by Generalitat de Catalunya (In-CAEM Project). IREC acknowledges funding from Generalitat de Catalunya 2021SGR01581, the European Union NextGenerationEU/PRTR, the 2BoSS project of the ERA-MIN3 program with the Spanish grant number PCI2022- 132985/AEI/10.13039/501100011033, and the SyDECat project from the Spanish MCIN/AEI/FEDER (PID2022-136883OB-C22). J.L. is a Serra Hún- ter Fellow and is grateful to projects MICIN/FEDER PID2024-156765OB- C21 and CEX2023-001300-M (MCIN/AEI 10.13039/501100011033). The authors thank the support from the project AMaDE (PID2023-149158OB- C43), funded by MCIN/ AEI/10.13039/501100011033/ and by “ERDF A way of making Europe”, by the “European Union”. ICN2 is supported by the Severo Ochoa program from Spanish MCIN / AEI (Grant No.: CEX2021-001214-S) and is funded by the CERCA Programme / General- itat de Catalunya. Part of the present work was performed in the frame- work of the Universitat Autònoma de Barcelona Materials Science PhD program. The authors acknowledge the use of instrumentation as well as the technical advice provided by the Joint Electron Microscopy Center at ALBA (JEMCA). ICN2 acknowledges funding from Grant IU16-014206 (METCAM-FIB) funded by the European Union through the European Re- gional Development Fund (ERDF), with the support of the Ministry of Research and Universities, Generalitat de Catalunya. ICN2 is a founding member of e-DREAM.7 - Energia Assequible i No Contaminant13 - Acció per al ClimaPostprint (published version
Un pont pedagògic entre teoria i pràctica. Aprenentatge corporitzat i habilitats humanes en la formació d’enginyeria mecànica en l’era de la IA
No full textLa creixent automatització dels processos de disseny en enginyeria mecànica, impulsada per eines digitals avançades i sistemes basats en intel·ligència artificial (IA), planteja nous reptes pedagògics. En aquest context, resulta necessari identificar quines habilitats humanes continuen essent essencials i com poden ser desenvolupades de manera efectiva dins la formació universitària. Aquest article presenta l’anàlisi del projecte docent de l’assignatura Ampliació d’Expressió Gràfica. Disseny Mecànic (AEGDM), impartida al cinquè quadrimestre del Grau en Enginyeria Mecànica a la UPC. El projecte, desenvolupat al llarg de tot el semestre, actua com un pont pedagògic entre teoria acadèmica i pràctica professional, integrant aprenentatge corporitzat, humanics i eines digitals avançades. L’avaluació, basada en dades quantitatives i qualitatives, mostra que el desenvolupament d’habilitats humanes, com la percepció espacial, el pensament crític, el judici tècnic i la responsabilitat professional, és inseparable de l’assoliment de competències tècniques. Es conclou que aquest tipus de projectes constitueixen un llindar pedagògic necessari abans de la incorporació plena a entorns de disseny automatitzat
A Methodological Review of Marketing Mix Modeling: The Robyn Framework
No full textEl Marketing Mix Modeling (MMM) és cada vegada més rellevant, ja que les empreses cerquen mètodes fiables i compatibles amb la privacitat per quantificar l’impacte incremental de la publicitat a diversos canals. La seva implementació és complexa per l’estadística implicada, l’experiència empresarial necessària i el coneixement especialitzat per modelar dinàmiques temporals com els efectes d’adstock i les funcions de saturació. A més, hi ha pocs recursos transparents i acadèmicament fonamentats sobre la seva aplicació. Aquesta tesi examina la biblioteca de codi obert de Meta, Robyn, establint primer els fonaments teòrics del MMM i les principals hipòtesis economètriques. S’analitza com Robyn operacionalitza aquests principis, incloent adstock, saturació, descomposició temporal amb Prophet i optimització multiobjectiu amb Nevergrad. Finalment, es realitza una implementació empírica amb dades reals, identificant reptes comuns, avaluant robustesa i interpretabilitat, i proposant bones pràctiques. Així, la tesi contribueix a la literatura independent sobre MMM i serveix de referència clara per a estudiants, analistes i organitzacions.El Marketing Mix Modeling (MMM) es cada vez más relevante, ya que las empresas buscan métodos fiables y compatibles con la privacidad para cuantificar el impacto incremental de la publicidad en múltiples canales. Su implementación resulta compleja por la estadística involucrada, la experiencia empresarial requerida y el conocimiento especializado necesario para modelar dinámicas temporales como efectos de adstock y funciones de saturación. Además, existen pocos recursos transparentes y académicamente fundamentados sobre su aplicación. Esta tesis examina la biblioteca de código abierto de Meta, Robyn, estableciendo primero los fundamentos teóricos del MMM y sus principales supuestos econométricos. Se analiza cómo Robyn operacionaliza estos principios, incluyendo adstock, saturación, descomposición temporal con Prophet y optimización multiobjetivo con Nevergrad. Finalmente, se realiza una implementación empírica con datos reales, identificando desafíos comunes, evaluando robustez e interpretabilidad, y proponiendo buenas prácticas. Así, la tesis contribuye a la literatura independiente sobre MMM y sirve de referencia clara para estudiantes, analistas y organizaciones.Marketing Mix Modeling (MMM) has become increasingly important as firms seek reliable, privacy-compliant methods to quantify the incremental impact of advertising across multiple channels. However, implementing MMM remains challenging due to its statistical complexity, the business expertise required, and the specialized knowledge needed to model temporal marketing dynamics, such as adstock effects and saturation functions. Moreover, there is limited availability of transparent, academically grounded resources that guide both the theoretical and practical aspects of implementation. This thesis addresses these challenges by providing a structured and accessible examination of Meta’s open-source MMM library, Robyn. The study first establishes the theoretical foundations of modern MMM, outlining its econometric structure and key modeling assumptions. It then examines how these principles are operationalized in Robyn, with particular attention to its mathematical treatment of adstock and saturation functions, time-series decomposition via Prophet, and multiobjective hyperparameter optimization through Nevergrad. Subsequently, an empirical implementation is conducted using real-world business data, illustrating the library’s behavior under practical conditions. The analysis identifies common implementation challenges, evaluates the robustness and interpretability of the results, and highlights best practices for applying MMM in industry contexts. By combining methodological analysis with empirical evaluation, this thesis contributes to the independent academic literature on MMM, an area currently dominated by the institutions that develop these tools, and serves as a clear reference for students, analysts, and organizations seeking to understand and critically assess MMM methodologies in practice
Intergenerational differences in higher education
No full textIntergenerational differences in higher education present both challenges and opportuni- ties for educators, students, and institutions. As age diversity among students increases, multi-generational classrooms create new dynamics in teaching and learning. Generational cohorts—such as Baby Boomers, Generation X, Millennials, and Generation Z—possess dis- tinct values, communication styles, and technological preferences, which influence their aca- demic expectations and interactions. While older faculty members may adhere to traditional, lecture-based teaching methods, younger students often prefer interactive, technology-driven learning experiences. This digital divide, coupled with differing attitudes toward authority, career aspirations, and feedback mechanisms, requires innovative educational approaches. Higher education institutions have responded to these challenges by implementing adap- tive strategies such as blended learning, peer mentorship programs, co-teaching models, and digital upskilling initiatives for faculty. Additionally, career guidance has evolved to accom- modate both traditional and modern work expectations, integrating entrepreneurship tracks alongside corporate career pathways. Cultural variations further complicate generational dif- ferences, with Asian educational traditions emphasizing hierarchy and respect, while Western models increasingly prioritize open discussion and student-centered learning. Looking for- ward, Generation Alpha is expected to push technological integration even further, which will require ongoing adaptation of educational methods and institutional policies. Addressing intergenerational differences requires fostering inclusive, flexible learning environments that leverage diverse perspectives to provide rich educational experiences. Future research should explore the long-term impact of evolving generational cohorts on higher education structures and methodologies
Enhancing wind turbine diagnostics with SCADA-vibration fusion, contrastive learning, and linear predictive coefcients
No full textWind energy plays a pivotal role in the transition to sustainable power generation. However, maintaining the reliability and efficiency of wind turbine (WT) remains a significant challenge due to complex operational conditions and the high cost associated with unexpected failures. Effective condition monitoring (CM) and predictive maintenance (PM) strategies are critical to mitigate these risks. This study presents a data-driven fault detection framework that fuses supervisory control and data acquisition (SCADA) data with high-frequency vibration signals using deep learning techniques to enhance diagnostic performance. Unlike conventional normal behavior models that rely exclusively on healthy data for training, the proposed framework incorporates limited labeled fault data when available. As only a few types of faults and a few samples are typically available in real-world scenarios, the approach does not assume a complete representation of all possible fault conditions. Instead, it is designed to generalize beyond the specific faults seen during training. This is demonstrated by training the model on healthy conditions and only two known fault types (with labeled data available) and testing it on a third, previously unseen fault type. In particular, Siamese networks with contrastive and reconstruction learning are employed to improve feature representation and anomaly detection. Two distinct methodologies are compared: the first utilizes a binary cross-entropy (BCE) loss function to classify the healthy or faulty status of the WT, while the second uses a triplet loss function for multiclass representation learning. Both methodologies generate low-dimensional representations of the input features, also known as embeddings. The resulting feature embeddings are passed through a k-means clustering algorithm to improve fault separation and identification. Statistical features are extracted from SCADA data to capture key trends or event information, while the linear prediction coefficient (LPC) method,
which models a signal by predicting future values based on its past samples, is applied to the vibration data for better fault characterization. The proposed approach is evaluated using the publicly available ETH Zurich dataset from an Aventa AV-7 turbine. Experimental results indicate that the fusion of SCADA and vibration-based diagnostics, in combination with contrastive and representation learning, substantially improves the predictive accuracy and generalization of fault detection models.This work was partially funded by the Spanish Agencia Estatal de Investigación (AEI), Ministerio de Economía, Industria y Competitividad (MINECO); the Fondo Europeo de Desarrollo Regional (FEDER) through the research projects PID2021-122132OB-C21 and TED2021-129512B-I00; and by the Generalitat de Catalunya through the research project 2021-SGR-01044. Open Access funding was enabled and organized by CRUE-CSUC Gold.Peer ReviewedPostprint (published version
Topological reconstruction of sampled surfaces via Morse theory
No full textIn this work, we study the perception problem for sampled surfaces (possibly with boundary) using tools from computational topology, specifically, how to identify their underlying topology starting from point-cloud samples in space, such as those obtained with 3D scanners. We present a reconstruction algorithm based on a careful topological study of the point sample that allows us to obtain a cellular decomposition of it using a Morse function. No triangulation or local implicit equations are used as intermediate steps, avoiding in this way reconstruction-induced artifices. The algorithm can be run without any prior knowledge of the surface topology, density or regularity of the point-sample. The results consist of a piece-wise decomposition of the given surface as a union of Morse cells (i.e. topological disks), suitable for tasks such as mesh-independent reparametrization or noise-filtering, and a small-rank cellular complex determining the topology of the surface. The algorithm, which we test with several real and synthetic sampled surfaces, can be applied to clouds coming from smooth surfaces with or without boundary, embedded in an ambient space of any dimension.This research work was funded by the European Commission - NextGenerationEU, through Momentum CSIC Programme: Develop Your Digital Talent. Franco Coltraro is staff hired under the Generation D initiative, promoted by Red.es, an organization attached to the Ministry for Digital Transformation and the Civil Service, for the attraction and retention of talent through grants and training contracts, financed by the Recovery, Transformation and Resilience Plan through the European Union’s Next Generation funds. F. Coltraro was also partially supported by the ClothIRI (CSIC 202350E080) project and the RobIRI 2021-SGR-00514 AGAUR project. Jaume Amorós and Maria Alberich-Carramiñana have been partially supported by the projects PID2019-103849GB-I00 and PID2023-146936NB-I00 financed by the Spanish State Agency MICIU/AEI/10.13039/501100011033 and by ERDF/EU, and by the GEOMVAP 2021-SGR-00603 AGAUR project.Peer ReviewedPostprint (published version
Conditional diffusion modeling for probabilistic behind-the-meter PV disaggregation
No full textDistributed behind-the-meter (BTM) PV systems are expanding rapidly, yet their generation is often metered together with demand, resulting in operational challenges for grid operators. This paper introduces a novel generative diffusion-based methodology for probabilistic BTM PV disaggregation that directly learns the distribution of PV generation from low-resolution smart meter data with 30-minute granularity. The proposed conditional diffusion model combines a domain-specific conditioning strategy with a sequential U-Net architecture to generate multiple plausible daily PV profiles, enabling both accurate point estimates and well-calibrated prediction intervals. On a real-world Australian benchmark, the model outperforms traditional deterministic baselines and representative probabilistic alternatives, achieving a mean absolute scaled error (MASE) of 0.954 and a coverage probability of 61.5% for the 50% prediction interval. Sensitivity studies highlight the effect of conditioning signals and training data volume, while cross-dataset experiments confirm generalization to Dutch residential data. By disaggregating BTM PV and quantifying uncertainty, this study demonstrates the potential of generative models to enhance grid visibility and support applications such as load forecasting, planning, and flexibility assessment.This work was part of the I+D+i project OPERA (Operation and Planning tools for Enabling Renewable distribution systems Acceleration based on emerging technologies for sustainable computing) with reference PID2024-160822OB-I00 funded by the Ministerio de Ciencia, Innovación y Universidades, Spain from Spain.Postprint (published version
Última oportunitat per emprendre a l’espai: l’ESA BIC Barcelona tanca convocatòria
No full textEl centre d’incubació d’empreses del Campus del Baix Llobregat de la UPC, busca quatre startups per oferir-los 60.000 € d’incentiu.Les candidatures per optar a les mentories durant l’incubació i l’accés a l’ecosistema espacial europeu es poden presentar fins el 14 de gene
Diseño de un marco metodológico para estimar las emisiones de GEI del sector industrial en el municipio de Rubí
No full textLas actuaciones a escala local representan un papel clave en la lucha contra el cambio climático, ya que permite tener presente el territorio y las particularidades del mismo. Sin embargo, muchos gobiernos municipales carecen de metodologías precisas para inventariar sus emisiones de gases de efecto invernadero (GEI). Esta investigación aborda a esa carencia mediante el diseño de un marco metodológico para la estimación de emisiones industriales a escala municipal, tomando como primera aproximación el sector cementero en Cataluña. Este estudio se basa en la aplicación y adaptación de la metodología propuesta por el Protocolo Global para Inventarios de Emisiones de GEI a Escala Comunitaria (GPC), teniendo en cuenta las particularidades y complejidades de la industria. La metodología se valida en los municipios catalanes que albergan plantas de fabricación de cemento, obteniéndose emisiones de proceso de 604 ktCO₂eq en Sant Vicenç dels Horts, 521ktCO₂eq en Santa Margarida i els Monjos, 331 ktCO₂eq en Alcanar y 280ktCO₂eq en Montcada i Reixac. Se evidencian las dificultades derivadas de la falta de datos públicos y la relevancia de los inventarios locales en el desarrollo de políticas climáticas municipale
