ALM: Atlantic Land Monitor

Scientific Area: Space–Earth Technologies

Funding: UT Austin 100,000 USD

PIs: UT — Jingyi Ann Chen (UT Austin) | PT — Joao Pinelo (Atlantic International Research Centre)

Start and End Dates: 2025-09-01 - 2026-12-31

Summary: The proposed international project, titled "Atlantic Land Monitor (ALM)", aims to create a robust data processing pipeline and prototype of a web-service to continuously monitor and report on land deformation using Interferometric Synthetic Aperture Radar (InSAR) technology. The ALM system will employ the Julia programming language to retrieve, process, analyze, and visualize InSAR data, and to convey information to key stakeholders. Julia's high performance capabilities, combined with its user-friendly syntax, make it an ideal choice for large-scale scientific computing tasks. The ALM web-service will allow users to be notified of land deformation events and interactively explore land deformation patterns, providing valuable insights for various applications such as natural disaster management, infrastructure monitoring, and environmental conservation.

Main Outcomes:

• Develop a NISAR L2 data fetcher.
• Validate our Julia InSAR software package through comparison with existing non-Julia InSAR processing packages.
• Setup software defined object storage and a database to store the interferograms and their metadata.
• Create functionality to process time series data based on a stack of interferograms.
• Create visualization tools.

SusBioA: Sustainable Electrochemical Approach for BioAmine production

Scientific Area: Nanotechnologies

Funding: UT Austin 100,000 USD

PIs: UT — Guihua Yu (UT Austin) | PT — Diana Fernandes (REQUIMTE – Rede de Química e Tecnologia – Associação (REQUIMTE-P)

Start and End Dates: 2025-12-01 - 2026-12-31

Summary: Nitrogen-containing compounds, particularly primary amines, are pivotal building blocks in nature and industry with extensive application in the synthesis of pharmaceuticals, polymer materials and agrochemicals. Still, the majority is industrially produced from fossil fuels using non-environmentally friendly processes. Therefore, the development of effective and greener catalytic processes for the preparation of amines from renewable resources is of utmost importance for a future sustainable society. A potential solution is the application of a new approach based on the emerging concept of Electro-Refinery – Electrochemical Reductive Amination (ERA). The main goals of SusBioA are to: 1. Apply the emerging concept of ERA as a clean, safe, low-cost and scalable catalytic process as an alternative to the conventional CRA using only electrons and protons (avoiding the use of fossil fuel derived molecular H2); 2. Use biomass as a substitute for fossil resources to produce bioamines, and as a substitute of noble metal-based and fossil fuel-based electrocatalysts. In route with this, SusBioA will develop new sustainable and highly efficient electrocatalysts for ERA of two selected biomass-derived platform molecules and establish an optimized ERA procedure with high activity and selectivity for amines production by systematic study of different electrochemical parameters.

Main Outcomes: This multidisciplinary team with complementary knowledge and facilities and with proven experience in the areas will advance the main goals of this project as below:

• Produce N/BCH-based catalysts from WH and MSW with high specific surface area
• Deliver a set of new sustainable and highly efficient SAECs for ERA of biomass-derived platform molecules FUR and HMF. Each SAECs will be specifically designed considering the target bioamine.
• Establish an optimized ERA procedure with high activity and selectivity for amines production by systematic study of different electrochemical parameters (potential bias, electrodes, electrolyte composition and pH, optimal electrocatalyst).

Scalable and Cost-Effective Solar-Powered Water Splitting: Integrating Silicon-Based Photoelectrodes with High-Performance Perovskite Solar Cells

Scientific Area: Nanotechnologies

Funding: UT Austin 100,000 USD

PIs: UT — Edward Yu (UT Austin) | PT — Luisa Andrade (University of Porto)

Start and End Dates: 2025-08-18 - 2026-12-31

Summary: This project will combine highly scalable silicon-based photoanodes developed at UT Austin using a simple, scalable process involving thin-film reactions in the Al/SiO2/Si system and high-speed, low-cost lithographic processes with advanced perovskite-based solar modules developed and fabricated at the University of Porto to create a complete system that performs solar-powered water splitting for low-cost green hydrogen production. UT Austin researchers will exploit the remarkable thin-film reactions that occur in the Al/SiO2/Si material system combined with a revolutionary technique for extremely rapid nanoparticle self-assembly to enable high-speed, highly scalable nanopatterning of catalysts in metal-insulator-silicon photoanodes. University of Porto researchers will use slot-die coating to fabricate perovskite solar mini-modules with areas of 2 cm2 or larger. The photoelectrodes and perovskite solar mini-modules will be combined into a self-contained solar water splitting system to be demonstrated and characterized at UT Austin.

Main Outcomes: This project is expected to yield

• Demonstration of micron-scale patterned silicon-based photoelectrodes at wafer scale and with high efficiency and stability.
• Demonstration of a self-contained solar-powered water splitting system for hydrogen generation that integrates silicon-based photoelectrodes with perovskite solar cells, and
• Validation of the scalability, performance, and potential for cost reduction of the targeted approaches for photoelectrode fabrication.

PATA: Power- and Thermal-aware Management for Energy-Efficient AI Training in HPC Infrastructures

Scientific Area: Advanced Computing

Funding USD: UT Austin 100,000 USD

PIs: UT — John Cazes (TACC, UT Austin) | PT — Ricardo Macedo (INESC TEC)

Start and End Dates: 2025-09-16 - 2026-12-31

Summary: The growing demand for AI applications has led to an increased reliance on HPC infrastructures to train large-scale models. Although these infrastructures provide the computational power for these workloads, the energy consumption and thermal management challenges associated with AI training remain largely neglected. AI training workloads, including large language models (LLMs) and generative AI (GenAI), exhibit distinct computational and memory access patterns that place unprecedented strain on HPC power and cooling systems. Existing HPC energy management techniques designed for general-purpose workloads fail to address the unique characteristics of AI training.

The PATA project proposes a novel power- and thermal-aware management framework tailored explicitly for AI training in HPC infrastructures. PATA aims to integrate workload-specific power-tuning mechanisms with intelligent thermal-aware job scheduling, PATA's power manager component dynamically monitors energy usage across compute resources and adjusts power draw based on the training phase, preventing energy waste without compromising model convergence time and accuracy. Simultaneously, the thermal manager can identify AI training workloads according to their heat profile and strategically place them within the HPC infrastructure to balance thermal distribution and minimize the cooling system's energy demands. By c"

Main Outcomes: PATA will produce three deliverables that are measurable indicators of its success:

• A final report detailing the project's research findings, dissemination activities, and key outcomes;
• A proof-of-concept open-source prototype;
• Two scientific papers: the first, detailing initial findings, will be submitted to a major conference/workshop from the systems and HPC communities (e.g., EuroSys, HPDC, ISC); the second, presenting

The core contributions of the project, will be submitted to a major HPC conference (e.g., Supercomputing, EuroSys, IPDPS).

NanoCatH2: Advanced Nanostructured Catalysts for PFAS Destruction via Catalytic Hydrogenation

Scientific Area: Nanotechnologies

Funding: UT Austin 100,000 USD

PIs: UT — Charles Werth (UT Austin) | PT — Salomé Soares (University of Porto)

Start and End Dates: 2025-08-16 - 2026-12-31

Summary: This project will focus on:

• Synthesis of new nanostructured PFAS defluorination catalysts.
• Evaluation of the relationships between catalytic properties and activity to elucidate mechanism of reaction and improve catalyst design.
• Incorporation of catalysts into electrochemical reactors to assess scale-up possibilities.
• A level 1 technoeconomic assessment to compare the new technology to competing alternatives.

Catalysts will be synthesized in different combinations and sizes, and characterized before and after PFAS activity assessment using x-ray diffraction, x-ray photoelectron spectroscopy, scanning and transmission electron microscopy with energy dispersive spectroscopy. PFAS activity assessment will be performed in small batch and flow through reactors, all as a function of solution conditions, including ionic strength, buffer type and capacity, pH, and PFAS composition. The most promising catalysts will be evaluated using a wider range of PFAS compounds, and assessed for capital and operating costs. These costs will be compared to competing technology costs for new technology assessment.

Main Outcomes: From this project

• Peer-reviewed scientific publication
• 2 conference presentations
• PhD thesis

Workload-Intelligent Data Storage for Next-Generation Advanced Computing Centers

Scientific Area: Advanced Computing

Funding: UT Austin 100,000 USD

PIs: UT — Amit Ruhela (TACC, UT Austin) | PT — João Paulo (INESC TEC and University of Minho)

Start and End Dates: 2025-09-16 - 2026-12-31

Summary: The WISE project aims to revolutionize data management for advanced computing by developing the first generation of storage systems capable of predicting the intrinsic and dynamic I/O patterns of data-intensive workloads and self-tuning their configurations based on such patterns.

By improving the operational efficiency of supercomputers, this project is set to significantly increase the performance and speed at which advanced computing centers support ground-breaking research. However, two main open research questions make this a high-risk, high-reward proposal: i) how can one accurately and timely predict the relevant set of I/O patterns of thousands of heterogeneous and dynamic workloads running at a given supercomputer? and ii) based on the predicted I/O patterns, how can these be used to automatically fine-tune or even develop tailored optimizations that will improve the overall efficiency of supercomputers?

The project will seek to answer these two questions by leveraging the expertise of INESC TEC’s team on storage and AI systems alongside researchers from TACC and MACC with extensive experience in managing HPC infrastructures. The project’s results will be disseminated through top scientific venues and integrated into an open-source proof-of-concept prototype, laying the groundwork for advanced computing centers’ first workload-aware data storage solution.

Main Outcomes: As success indicators, WISE will consider a final report detailing the project’s research and dissemination outputs, along with two proof-of-concept open-source prototypes, including the prediction engine and integrated SDS solution.

Moreover, the project aims to submit at least two research papers to peer-reviewed workshops (e.g., REX-IO, HotStorage, FTXS, or PDSW) and to major HPC conferences (e.g., SC, ATC, HPDC, CCGrid, Cluster, IPDPS).

Finally, the project will include both MSc and PhD students contributing to their training and specialization in the HPC and storage systems fields.

Scalable and Cost-Effective

Scientific Area: Space-Earth Technologies

Funding: UT Austin 100,000 USD

PIs: UT — Fernanda Leite (UT Austin) | PT — Miguel Ângelo Dias Azenha (University of Minho)

Start and End Dates: 2026-01-01 - 2026-12-31

Summary: This project aims to establish an innovative methodology for scalable flood analysis and prediction utilizing advanced airborne data collection techniques such as drone-based photogrammetry and airplane-based LiDAR. By systematically exploring levels of detail (LOD) for hydraulic-relevant features, it seeks to harmonize modeling strategies, propose GeoBIM interoperability frameworks, and validate methodologies through transatlantic case studies in the Gulf of Mexico and Portugal's west coast. The project aims to lay a foundation for scalable, automated, and precise flood modeling applicable globally.

Main Outcomes:

• Create a framework for Levels of Detail (LOD), particularly for hydraulic-relevant infrastructure.
• Establish and propose solutions for GeoBIM interoperability, integrating BIM and GIS for flood modeling.
• Conduct case studies that compare the effects of different LOD and validate our methods.
• Develop and disseminate the latest reality capture datasets of the built environment to support flood modeling studies.

Space Distributed Ocean Monitorization and Exploration

Scientific Area: Space–Earth Technologies

Funding: UT Austin 100,000 USD

PIs: UT — Thinh Doan (Aerospace Engineering and Engineering Mechanics) | PT — Daniel Silvestre (NOVA University Lisbon)

Start and End Dates: 2025-08-16 - 2026-12-31

Summary: Environmental monitoring - such as assessing ocean health - requires coordinated observations from diverse autonomous platforms, including satellites, high-altitude balloons, and aerial or underwater robots. Integrating these heterogeneous assets presents a fundamental multi-vehicle control challenge: maximizing coverage while ensuring safe operation in increasingly congested orbital and operational environments. In this collaborative UT–Portugal Exploratory Project, we will develop a hierarchical control framework for autonomous vehicles operating in uncertain and dynamic environments. At the high level, we will design reinforcement learning–based path planning algorithms to generate action sequences optimized for given desired tasks in uncertain and dynamic environments. These planned actions will be fed into the low-level control layer, where we will implement a unified Quadratic Program–based architecture combining Control Lyapunov Functions for stability with Control Barrier Functions for safety. The integrated planning and control strategy will be validated on autonomous aerial and ground vehicles in UT’s indoor and outdoor test facilities, paving the way for robust, scalable, and safe environmental monitoring missions.

Main Outcomes:
The expected outcomes of this project include a suite of mathematical tools and techniques for enabling autonomous agents to cooperate effectively in uncertain, dynamic environments, with direct applicability to real-world problems. In addition, the project will establish a strong foundation for long-term collaboration between researchers and students at UT and Portugal, fostering knowledge exchange, joint experimentation, and the development of a shared research agenda in safe autonomy. At the end of this project, the research is expected to produce a conference paper and potentially submit a journal paper.