Embedded Files
Ongoing Research Projects
Ongoing Research Projects
Research Projects (co-PI):
1. Long-Range Wildfire Occurrence Forecasting Using Nonlinear Dynamics and Artificial Intelligence, US Forest Services, $50,000.
2. PyroKit: Wildfire Threat, Ignition and Mitigation Toolkit, Department of Energy, $2M.
Past Research Projects
Past Research Projects
SMART SOLUTIONS FOR FIGHTING WILDFIRES: AI-ENHANCED REMOTE SENSING FOR IMPROVED FIRE HAZARD MITIGATION AND RESPONSE
Achieving effective preparedness and timely response to natural disasters necessitates the amalgamation of artificial intelligence with remote sensing to construct scalable solutions for data harnessing and information extraction. Nevertheless, the dearth and diversity of field-extracted ground truth data have impeded numerous such applications. In response, we are developing a scalable semi-supervised framework for wildfire fuel mapping, which is based on propagating the sparse ground truth labels. This framework will empower physics-based computational fire models, enabling a more precise and reliable prediction of wildfire behavior.
This research is funded by Amazon Science Hub and also a part of LEAP-HI, a multi-disciplinary NSF-funded research effort that introduces a data-informed, physics-based computational framework for combatting wildfire hazard. With the existing shortage of real-time fuel biomass maps across the US, we will present a unique contribution to the efforts to restore quality of life, environmental justice, and socio-ecological balance within the affected communities.
MAPPING WILDFIRE VULNERABILITY FOR A MEDITERRANEAN ISLAND (SARDINIA)
Vulnerability assessment is a vital component of wildfire management. This research focused on the development of a framework to measure and map vulnerability levels in one of the largest islands of Italy, where wildfires are a major concern. The framework followed a stepwise approach to evaluate its main components, expressed by exposure, sensitivity and coping capacity. Data on population density, fuel types, protected areas location, roads infrastructure and surveillance activities, among others, were integrated to create composite indices, representing each component and articulated in multiple dimensions. Map was created for Sardinian Island of Italy, with the contribution of local participants from civil protection institutions and forest services. More details of this research can be found in my PhD thesis. This research is part of a large interdisciplinary project S2IGI: An Integrated System for Wildfire Management funded under EU Horizon 2020.
Sardinia, Italy
AN AUTOMATIC MACHINE LEARNING BASED ALGORITHM DEVELOPMENT FOR MAPPING FOREST FIRE FUELS IN EUROPE
Natural vegetation confers various benefits to human society, but also constitutes a potential source of fuel for destructive wildfires. Accordingly, the mapping of wildfire fuel types is essential to prevent these disasters. Hyperspectral imagery, a technology with diverse applications, has emerged as a promising tool for this task. We developed an innovative semi-supervised machine learning approach that automatically discriminates between wildfire fuel types, leveraging hyperspectral imagery captured by PRISMA, a cutting-edge satellite launched by the Italian Space Agency. Our approach generates samples and pseudo-labels using a single spectral signature for each class, applies a fully constrained linear mixing model to unmix mixed pixels, and distinguishes sparse and mountainous vegetation from typical vegetation by analyzing biomass and DEM maps. We demonstrated the effectiveness of this approach by applying it to PRISMA images of the southern part of Sardinia, an island of Italy, and achieving an impressive overall accuracy of 87%. We also tested the stability of our method by repeating the procedure on Bulgaria (having different vegetation and weather conditions), obtained an overall accuracy of around 84% and demonstrated a degree of confidence greater than 95% in terms of repeatability and reproducibility analysis. Our findings indicate that PRISMA imagery holds great potential for wildfire fuel mapping, and our semisupervised learning approach is capable of generating training samples for the machine learning model when no single go-to dataset is available. Furthermore, our procedure can be employed to develop a wildfire fuel map for any part of Europe using LUCAS land cover points from COPERNICUS as input. This research is a part of the project 'Sviluppo di Prodotti Iperspettrali Prototipali Evoluti' funded by Italian Space Agency and worked in collaboration with e-GEOS. Please refer to ML4FM for more details.
THERMALLY STABLE ELECTRO CATALYTIC NICKEL-PHOSPHIDE FILM DEPOSITION ON GRAPHITE FOR HYDROGEN EVOLUTION REACTION ELECTRODE APPLICTION
The deposition of nickel phosphide films is a critical process in the electronic fabrication industry and as a hydrogen generating energy material. Therefore, achieving strong film adhesion on a variety of substrates, including insulators (e.g., Al2O3, SiO2) and conductors (e.g., glassy carbon, graphite, transparent conductors-FTO), is of utmost importance, particularly in the production of printed circuit boards. Nickel phosphide films exhibit desirable properties such as high-temperature thermal stability, anti-corrosion characteristics, large wear resistance, and a competitive over-potential with respect to Platinum metal (for fuel cells). Our study addresses the challenge of depositing an adherent nickel phosphide (Ni-P) film onto a graphite (GR) substrate using an economical electroless (EL) technique. We developed procedure to activate GR for electroless deposition and our method increased the inter-planar distance, resulting in favorable catalytic sites for Ni and P ions during EL deposition. We yielded an adherent film of the Ni3P phase after thermal treatment, known to be thermally stable beyond 1000 ◦C. The Ni-P/GR film exhibits high potential for application in (i) HER (hydrogen evolution reaction) for electro-catalytic energy generation, and (ii) graphite boats that are usable under elevated temperatures. This research is funded by International Advanced Research Center for Powder Metallurgy and New Materials (ARCI) and please refer to Multifunctional NiP deposition for more details on this project.
Left: Electrochemical characterization displaying Tafel curve and LSV indicates the possible role of Ni –P film for HER application; Right: (a) Bare Graphite Boat, (b) Ni –P coated Graphite Boat and (c) Ni –P caoted Graphite Boat (durability test carriedout by heating the NiP-caoted GR boat under steam at 800oC )
Page updated
Report abuse