Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “ENERGY SCIENCES NETWORK”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 1,585 records · Page 88

Artificial Intelligence-Enhanced CMIP6 Climate Projections Across the Conterminous United States

This dataset comprises high-resolution climate projections at 1/24 degree grid (~4km) over the conterminous United States (CONUS) based on ten Global Climate Models (GCMs) that are part of the Coupled Models Intercomparison Project phase 6 (CMIP6). The CMIP6 GCMs are downscaled using two artificial intelligence (AI) techniques, primarily based on the computer vision approach called super-resolution. We train two separate networks: super-resolution convolutional neural network (SRCNN) and super-resolution generative adversarial network (SRGAN). The networks are trained using Daymet observations, originally available at a 1 km resolution. For training purposes, the Daymet data is interpolated to 1/24 degree (~4km), 0.25 degree and 1 degree, which serve as high, intermediate and low-resolution inputs respectively. For each of the SRCNN and SRGAN network, we use a two-step resolution enhancement, the first step generates 4x refinement from 1 degree to 0.25 degree and the second step generates 6x refinement from 0.25 degree to 1/24 degree (~4km). We downscale daily scale precipitation, maximum temperature and minimum temperature for the six CMIP6 GCMs for 1980 to 2019 in the historical period and 2020 to 2059 in the near-term future under the shared socioeconomic pathway 585 and 245 (SSP585 and SSP245) emission scenarios. We also perform double bias-correction with Daymet observations using a quantile mapping approach, first for GCMs prior to making predictions at 1 degree grid and second after making final predictions at ~4km.

13 HYDRO ENERGY↗

Ecobuoys for Scalable Oceanography

An approach to scalable surface-drifting buoys is needed to enable the high spatial and temporal resolution of oceanographic data that the science and meteorological communities are asking for. With the number of active buoys predicted to increase by a factor of 100 or more, the impact on the environment becomes even more important. Here, we present a pathway to a scalable and sustainable generation of buoys. We identify the main criteria to be used when developing such buoys to be low cost, with reliable data and neutral or even positive environmental impact. For each buoy subsystem—hull, electronics, energy generation and storage, sensors, and communication system—cutting-edge technological solutions are presented, many of them from emerging research in marine or other disciplines. We then assess the potential solutions against the design criteria and plot a path toward small, environmentally friendly, low-cost, and low-power buoys.

54 ENVIRONMENTAL SCIENCES↗

Tailorable Adhesives for Multi-material Joining, Facile Repair and Re-assembly

Joining of materials and components is inevitable as it allows versatility in assembly and repair along with reduction in time and cost of manufacturing. This work uses ‘active adhesives’ that inherit all the benefits of bonded joints, such as lightweight, eliminate holes and associated stress-concentrations, and overcomes the shortcomings of disassembly and repair. Thermoplastic adhesives modified by the incorporation of electrically conductive graphene nanoplatelets at a concentration above the percolation point provide a unique synergy of mechanical, thermal and electrical properties. While the choice of the thermoplastic is governed by the desired application the addition of the graphene nanoplatelets allows energy to be deposited primarily in the adhesive. The percolated network of graphene particles in the adhesive at less than 2% can quickly couple to high frequency radiation (microwave, or MW) via non-contact methods and increase the adhesive temperature to above the required processing temperatures. The adhesive melts and flows over the adherends and upon cooling forms a structural adhesive bond. Furthermore, the process can be used to disassemble the adhesive joint if repair or reworking is required. In this work, the concept of aforementioned active adhesives was studied using varying graphene contents in nylon-6 adhesive. Activation and re-assembly was studied on in-plane (lap-shear) joints. Structural behavior of re-assembled joints was found to be similar to thermally bonded joints. Overall, active adhesives such as the ones attempted in this work have a great potential in a wide range of applications wherein in-situ repair, re-assembly and recyclability are essential.

36 MATERIALS SCIENCE↗

Nanoporous Architectures for Multifaceted Clean Energy Applications

Recently, there has been immense interest in the design and synthesis of highly porous organic architectures due to their multifaceted potentials for applications that include gas storage, separation, conductivity, and catalysis. The chemical composition, physical, and textural properties are dictated during synthesis. This allows for the production of materials with enhanced properties relevant to their respective applications. With the exception of microcrystalline covalent-organic frameworks (COFs), these polymeric materials are amorphous with considerably high porosity and well-defined cavities, which render them highly attractive, especially in adsorptive gas storage. Such desirable traits are imparted into organic materials through the use of rigid building blocks that direct the growth of polymer networks without the aid of templating agents. During the award period, we focused our efforts on the synthesis and characterization of porous organic polymers constructed through B-N and C-N bond formation and investigated their performance in gas storage (H 2 , CH 4 , CO 2 ) and selective CO 2 binding over CH 4 and N 2 . We also demonstrated that thermolysis of nitrogen rich organic polymers afford highly porous nitrogen-rich porous carbons that show remarkable CO 2 capture and separation properties.

36 MATERIALS SCIENCE↗

Efficient multi-scale representation of visual objects using a biologically plausible spike-latency code and winner-take-all inhibition

Deep neural networks have surpassed human performance in key visual challenges such as object recognition, but require a large amount of energy, computation, and memory. In contrast, spiking neural networks (SNNs) have the potential to improve both the efficiency and biological plausibility of object recognition systems. Here we present a SNN model that uses spike-latency coding and winner-take-all inhibition (WTA-I) to efficiently represent visual stimuli using multi-scale parallel processing. Mimicking neuronal response properties in early visual cortex, images were preprocessed with three different spatial frequency (SF) channels, before they were fed to a layer of spiking neurons whose synaptic weights were updated using spike-timing-dependent-plasticity. We investigate how the quality of the represented objects changes under different SF bands and WTA-I schemes. We demonstrate that a network of 200 spiking neurons tuned to three SFs can efficiently represent objects with as little as 15 spikes per neuron. Furthermore, studying how core object recognition may be implemented using biologically plausible learning rules in SNNs may not only further our understanding of the brain, but also lead to novel and efficient artificial vision systems.

59 BASIC BIOLOGICAL SCIENCES↗

Machine Learning Downscaling of SoilMERGE in the United States Southern Great Plains

SoilMERGE (SMERGE) is a root-zone soil moisture (RZSM) product that covers the entire continental United States and spans 1978 to 2019. Machine learning techniques, Random Forest (RF), eXtreme Gradient Boosting (XGBoost), and Gradient Boost (GBoost) downscaled SMERGE to spatial resolutions straddling the field scale domain (100 to 3000 m). Study area was northern Oklahoma and southern Kansas. The coarse resolution of SMERGE (0.125 degree) limits this product’s utility. To validate downscaled results in situ data from four sources were used that included: United States Department of Energy Atmospheric Radiation Measurement (ARM) observatory, United States Climate Reference Network (USCRN), Soil Climate Analysis Network (SCAN), and Soil moisture Sensing Controller and oPtimal Estimator (SoilSCAPE). In addition, RZSM retrievals from NASA’s Airborne Microwave Observatory of Subcanopy and Surface (AirMOSS) campaign provided a nearly spatially continuous comparison. Three periods were examined: era 1 (2016 to 2019), era 2 (2012 to 2015), and era 3 (2003 to 2007). During eras 1 and 2, RF outperformed XGBoost and GBoost, whereas during era 3 no model dominated. Performance was better during eras 1 and 2 as opposed to the pre-L band era 3. Improvements across all eras, regions, and models realized from downscaling included an increase in correlation from 0.03 to 0.42 and a decrease in ub RMSE from -0.0005 to -0.0118 m 3 /m 3 . This study demonstrates the feasibility of SMERGE downscaling opening the prospect for the development of a long-term RZSM dataset at a more desirable field-scale resolution with the potential to support diverse hydrometeorological and agricultural applications.

54 ENVIRONMENTAL SCIENCES↗

Technoeconomic Analysis of Microwave-Assisted Dry Reforming Integrated with Chemical Looping for Production of Methanol

Methanol is a key component in producing formaldehyde, acetic acid, and methyl tert -butyl ether (MTBE) and supports a wide array of industries, including plastics, textiles, and automotive. It also plays a growing role in renewable energy solutions. However, the conventional production of methanol involves steam reforming of methane, which is very energy-intensive and produces significant quantities of the greenhouse gas carbon dioxide. In this research, a chemical looping scheme is combined with dry reforming of natural gas in a novel microwave reactor to produce an industrial quantity of methanol. A heat exchanger network is developed to substantially reduce hot and cold utility usage. The effect of the cost of purchasing carbon dioxide from an external source for dry reforming, the capital cost of the microwave reactor, and the cost of electricity on the net present value is analyzed. Technoeconomic comparison with the conventional industrial process that produces methanol via steam reforming of methane indicates that the chemical looping generates a significant positive net present value along with a substantial reduction in carbon dioxide emissions while producing methanol significantly below the U.S. Department of Energy’s goal of $800/ton.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Quantum-enhanced distributed phase sensing with a truncated SU(1,1) interferometer

In recent years, distributed quantum sensing has gained interest for a range of applications requiring networks of sensors, from global-scale clock synchronization to high energy physics. In particular, a network of entangled sensors can improve not only the sensitivity beyond the shot noise limit, but also enable a Heisenberg scaling with the number of sensors. Here, using bright entangled twin beams, we theoretically and experimentally demonstrate the detection of a linear combination of two distributed phases beyond the shot noise limit with a truncated SU(1,1) interferometer. Specifically, we show a quantum noise reduction of 1.7 ± 0.3 dB below what is possible with the corresponding classical configuration. Additionally, we theoretically extend the use of a truncated SU(1,1) interferometer to a multi-phase-distributed sensing scheme that leverages entanglement as a resource to achieve a quantum improvement in the scaling with the number of sensors in the network. Our results pave the way for developing quantum-enhanced sensor networks that can achieve an entanglement-enhanced sensitivity.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

New principles of self‐organization created through the interplay of DNA condensates, microtubules, and motors

Bioinspired design—which holds great promise for a new generation of materials that are robust to defects, scalable under green manufacture, environmentally responsive, and programmably reconfigurable—requires mastery over molecular self-organization. Yet, from its specific mechanisms to most general architectures, the principles governing self-organization remain poorly understood and not even fully enumerated. For living systems, one obvious architectural principle is the modular reuse of a few simple molecular components in myriad combinations to achieve more complex phenomena. For example, the mechanical tasks of a cell are driven by the nonequilibrium dynamics of cytoskeletal filaments and molecular motors—the same filaments and motors, reprogrammed by a variety of modulators, perform tasks ranging from cell movement to division. Similarly, many compartmentalization tasks are performed by liquid-like condensates of simple components, which act as membraneless organelles to localize particular molecules in space and time (e.g. for gene regulation or RNA processing). In a few cases, condensates combine and interact with the cytoskeleton to create still more complex phenomena, e.g. the nucleation of microtubule asters from the centrosome (a protein condensate) to form the mitotic spindle during cell division. Very little is known about the fundamental mechanisms of such filament-plus-condensate phenomena. Despite few examples, the landscape of behaviors that can be achieved through the combination of condensates, filaments, and motors appears vast. However, exploration has been hindered by a lack of systems that have sufficiently programmable and dynamically tunable interactions between component condensates, filaments, and motors. We proposed to combine programmable DNA condensates, filamentous microtubules, and light-controlled motors into self-organizing systems whose principles go beyond those that have been observed in nature. In one limit, our systems will use microtubules and motors to create the molecular analog of a network of roads, which will organize droplets of DNA condensates capable of carrying molecular cargo. DNA condensates coupled to motors will flow from one microtubule aster hub to another, with their direction and timing controlled by DNA circuits. In another limit, microtubules will swim through bulk DNA condensates and exhibit strong interactions with boundaries between different types of condensates. Microtubule swimmers will reflect, get trapped, or refract at boundaries, under a mechanical analog of the classical optical index of refraction. DNA condensates having different mechanical indexes of refraction will be used to construct the analog of optical lenses, so that microtubule swimmers can be manipulated like light—collimated, diffracted, focused, and sorted based on properties analogous to wavelength. These two limits define two new architectures, within which multiple new mechanistic principles for self-organization will be discovered and explored. To explore these architectures, the motor-based coupling between DNA condensates and filaments will be controlled in time and space through the use of opto-proteins that create reversible links between DNA condensates and motors upon illumination. For each principle of interest, patterns of light will create virtual experiments by defining patterns of activity where DNA condensates walk along filaments, or filaments swim through condensates, and patterns of inactivity which will serve either as controls, or as boundary conditions vital to create the desired phenomena. This research serves the goals of Basic Energy Sciences Biomolecular Material Program by elucidating the principles by which the emergent, nonequilibrium behavior of collections of DNA condensates, motors, and microtubules can be programmed by environmental light patterns to create complex motion and materials transport. Because DNA condensates can be readily coupled to virtually any high performance nanomaterial, from carbon nanotubes, to metal nanoparticles, to light harvesting systems, this work provides a path to the construction, self-maintenance and reconfiguration of materials relevant to the Department of Energy.

60 APPLIED LIFE SCIENCES↗

Final Scientific Findings Report- Hydro Research Foundation- Research Awards Program

The Research Awards Program was established in 2014 by the Hydro Research Foundation (Foundation) to support research efforts for the hydropower industry. Using the previously funded Fellowship Program that ran from 2010-2015, the Research Awards Program (RAP or program) captured the lessons learned from the Fellowship program and fine-tuned a nation-wide graduate level research program to support innovation, research and workforce development. The RAP was funded through a $2 million-dollar grant through the Department of Energy’s Water Power Program within the Energy Efficiency and Renewable Energy office (DOE). This award made possible, 33 research projects by 32 students from 21 universities in 11 states. The grant widely engaged the industry by providing the opportunity to work with 25 organizations as well as 63 students and advisors. The funded research focused on a topic list selected by industry and worked to cover a breadth of research needs across many disciplines, basins and types of hydropower issues. The funded work, answered pressing research needs in the fields of the following disciplines; civil engineering, mechanical engineering, environmental engineering, electrical engineering, geography, fisheries, wildlife, water resources, geology, energy science and materials innovation. The Foundation estimates that 116 individuals benefited from this research through either direct mentorship (42), research (32), overseeing the research (31), partnerships (7) and the day-to-day management of the program (4). This investment by DOE showcases how private-public partnerships can leverage limited funds to create a wide range of research products. The project objective was to build on the successful Fellowship Program to develop and execute the RAP on a national level. The program sought to provide competitively awarded hydropower-related graduate-level research awards. The intent of the RAP was to support beneficial, innovative research related to conventional or pumped-storage hydropower. It also sought to stimulate interest among students and universities in this research arena and provide students with a foundation for productive careers within the hydropower industry. The outcome was 33 research findings with 100% of proposed research completed. In addition, 87% of students chose to either join the hydropower industry or continue their hydropower related research at their universities. These research findings advanced the understanding of pressing research topics within the hydropower industry and met the objective of the funding. An ancillary benefit the program identified through this funding was, not only did 63 students and advisors work on hydropower within the academia space, but now there are more professors and students teaching hydropower at the university level. The program saw an addition of 12 new professors, teaching assistants, and graduate researchers continuing to simulate interest in hydropower. Another interesting outcome is that during this grant funding, the program saw alumni of the predecessor program, who are now professors, applying for the RAP with their students. A unique element of this funding is there was significant cost-share provided by host universities, who did not charge facilities and administration overhead to the student’s research awards. In addition, each university provided laboratories, computing capacity and research facilities to support the research which is extraordinarily unique to grant funding at the university level. The total cost share exceeded 30% of the total grant cost. The funding clearly reached many sectors of the industry to provide helpful work and continue building hydropower’s next generation of innovation and workforce. The Foundation has established a broad network of universities interested and engaged in hydropower. The work was supported by the hydropower industry through donations of meeting space, conference registrations, in-kind contributions, mentorship and support of students. The Foundation would recommend finding a vehicle to further advance programs like the graduate level research through competitions, competitively funded awards and student research. It will take many universities to advance hydropower’s most pressing challenges. To let a program like the RAP dwindle away would incapacitate research at universities.

13 HYDRO ENERGY↗

The structure of alkali silicate gel by total scattering methods

The structure of the alkali silicate gel (ASR) collected from the galleries of Furnas Dam in Brazil was determined by a pair distribution function (PDF) analysis of high energy X-ray diffraction data. Since this method is relatively new to concrete structure analysis a detailed introduction on the PDF method is given for glassy SiO{sub 2}. The bulk amorphous structure of the dam material is confirmed as no Bragg peaks are observed in the scattered intensity. The real space results show that the local structure of the amorphous material is similar to kanemite (KHSi{sub 2}O{sub 5}:3H{sub 2}O) however the long range layer structure of the crystal is broken up in the amorphous state, so that ordering only persists of the length scale of a few polyhedra. The silicate layer structure is a much more disordered than predicted by molecular dynamics models. The X-ray results are consistent with the molecular dynamics model of Kirkpatrick et al. (2005) [1] which predicts that most of the water resides in pores within the amorphous network rather than in layers. The total scattering data provide a rigorous basis against which other models may also be tested.

36 MATERIALS SCIENCE↗

Hyperglycemia alters mitochondrial respiration efficiency and mitophagy in human podocytes

Highlights: • Impairment of mitochondrial dynamics underlies the pathogenesis of type 2 diabetes. • Mitochondrial respiration is altered in glomeruli isolated from diabetic STZ-rats. • High glucose induces mitochondrial fission and fragmentation in human podocytes. • Mitochondrial fusion, biogenesis and mitophagy are reduced in HG-treated podocytes. • HG can injure the kidney by alteration of podocytes and glomerular bioenergetics. Podocytes constitute the outer layer of the renal glomerular filtration barrier. Their energy requirements strongly depend on efficient oxidative respiration, which is tightly connected with mitochondrial dynamics. We hypothesized that hyperglycemia modulates energy metabolism in glomeruli and podocytes and contributes to the development of diabetic kidney disease. We found that oxygen consumption rates were severely reduced in glomeruli from diabetic rats and in human podocytes that were cultured in high glucose concentration (30 mM; HG). In these models, all of the mitochondrial respiratory parameters, including basal and maximal respiration, ATP production, and spare respiratory capacity, were significantly decreased. Podocytes that were treated with HG showed a fragmented mitochondrial network, together with a decrease in expression of the mitochondrial fusion markers MFN1, MFN2, and OPA1, and an increase in the activity of the fission marker DRP1. We showed that markers of mitochondrial biogenesis, such as PGC-1α and TFAM, decreased in HG-treated podocytes. Moreover, PINK1/parkin-dependent mitophagy was inhibited in these cells. These results provide evidence that hyperglycemia impairs mitochondrial dynamics and turnover, which may underlie the remarkable deterioration of mitochondrial respiration parameters in glomeruli and podocytes.

60 APPLIED LIFE SCIENCES↗

CuXASNet: Rapid and accurate prediction of copper L-edge x-ray absorption spectra using machine learning

In this work, we have developed CuXASNet, a dense neural network that predicts simulated Cu -edge x-ray absorption spectra (XAS) from atomic structures. Featurization of the Cu local environment is performed using a component of M3GNet, a graph neural network developed for predicting the potential energy surface. CuXASNet is trained on simulated spectra from FEFF9 at the multiple scattering level of theory, and can predict the and edges for Cu sites to quantitative accuracy. To validate our approach, we compare 14 experimental spectra extracted from the literature with the predictions of CuXASNet. The agreement of CuXASNet with experiments is shown by an average mean absolute error of 0.125 and an average Spearman's correlation coefficient of 0.891, which is comparable to FEFF9's values of 0.131 and 0.898 for the same metrics. As such, CuXASNet can rapidly predict a large number of -edge XAS spectra at the same accuracy as FEFF9 simulations. This can be used as a drop-in replacement for multiple scattering codes for fast screening of candidate atomic structure models of a measured system. This model establishes a general framework for Cu XAS prediction, and can be extended to more computationally expensive levels of theory and to other transition metal edges.

36 MATERIALS SCIENCE↗

Co-catabolism of arginine and succinate drives symbiotic nitrogen fixation

Biological nitrogen fixation emerging from the symbiosis between bacteria and crop plants holds promise to increase the sustainability of agriculture. One of the biggest hurdles for the engineering of nitrogen-fixing organisms is an incomplete knowledge of metabolic interactions between microbe and plant. In contrast to the previously assumed supply of only succinate, we describe here the CATCH-N cycle as a novel metabolic pathway that co-catabolizes plant provided arginine and succinate to drive the energy-demanding process of symbiotic nitrogen fixation in endosymbiotic rhizobia. Using systems biology, isotope labeling studies and transposon sequencing in conjunction with biochemical characterization, we uncovered highly redundant network components of the CATCH-N cycle including transaminases that interlink the co-catabolism of arginine and succinate. The CATCH-N cycle uses N2 as an additional sink for reductant and therefore delivers up to 25% higher yields of nitrogen than classical arginine catabolism—two alanines and three ammonium ions are secreted for each input of arginine and succinate. We argue that the CATCH-N cycle has evolved as part of a synergistic interaction to sustain bacterial metabolism in the microoxic and highly acid environment of symbiosomes. Thus, the CATCHN cycle entangles the metabolism of both partners to promote symbiosis. Our results provide a theoretical framework and metabolic blueprint for the rational design of plants and plant-associated organisms with new properties to improve nitrogen fixation.

59 BASIC BIOLOGICAL SCIENCES↗

Integration of physiologically relevant photosynthetic energy flows into whole genome models of light‐driven metabolism

SUMMARY Characterizing photosynthetic productivity is necessary to understand the ecological contributions and biotechnology potential of plants, algae, and cyanobacteria. Light capture efficiency and photophysiology have long been characterized by measurements of chlorophyll fluorescence dynamics. However, these investigations typically do not consider the metabolic network downstream of light harvesting. By contrast, genome‐scale metabolic models capture species‐specific metabolic capabilities but have yet to incorporate the rapid regulation of the light harvesting apparatus. Here, we combine chlorophyll fluorescence parameters defining photosynthetic and non‐photosynthetic yield of absorbed light energy with a metabolic model of the pennate diatom Phaeodactylum tricornutum. This integration increases the model predictive accuracy regarding growth rate, intracellular oxygen production and consumption, and metabolic pathway usage. Through the quantification of excess electron transport, we uncover the sequential activation of non‐radiative energy dissipation processes, cross‐compartment electron shuttling, and non‐photochemical quenching as the rapid photoacclimation strategy in P. tricornutum. Interestingly, the photon absorption thresholds that trigger the transition between these mechanisms were consistent at low and high incident photon fluxes. We use this understanding to explore engineering strategies for rerouting cellular resources and excess light energy towards bioproducts in silico . Overall, we present a methodology for incorporating a common, informative data type into computational models of light‐driven metabolism and show its utilization within the design–build–test–learn cycle for engineering of photosynthetic organisms.

59 BASIC BIOLOGICAL SCIENCES↗

All-polymer organic solar cells with nano-to-micron hierarchical morphology and large light receiving angle

Distributed photovoltaics in living environment harvest the sunlight in different incident angles throughout the day. The development of planer solar cells with large light-receiving angle can reduce the requirements in installation form factor and is therefore urgently required. Here, thin film organic photovoltaics with nano-sized phase separation integrated in micro-sized surface topology is demonstrated as an ideal solution to proposed applications. All-polymer solar cells, by means of a newly developed sequential processing, show large magnitude hierarchical morphology with facilitated exciton-to-carrier conversion. The nano fibrilar donor-acceptor network and micron-scale optical field trapping structure in combination contributes to an efficiency of 19.06% (certified 18.59%), which is the highest value to date for all-polymer solar cells. Furthermore, the micron-sized surface topology also contributes to a large light-receiving angle. A 30% improvement of power gain is achieved for the hierarchical morphology comparing to the flat-morphology devices. These inspiring results show that all-polymer solar cell with hierarchical features are particularly suitable for the commercial applications of distributed photovoltaics due to its low installation requirement.

14 SOLAR ENERGY↗

Smart pixel sensors: towards on-sensor filtering of pixel clusters with deep learning

Highly granular pixel detectors allow for increasingly precise measurements of charged particle tracks. Next-generation detectors require that pixel sizes will be further reduced, leading to unprecedented data rates exceeding those foreseen at the High- Luminosity Large Hadron Collider. Signal processing that handles data incoming at a rate of $\mathcal{O}$(40 MHz) and intelligently reduces the data within the pixelated region of the detector at rate will enhance physics performance at high luminosity and enable physics analyses that are not currently possible. Using the shape of charge clusters deposited in an array of small pixels, the physical properties of the traversing particle can be extracted with locally customized neural networks. In this first demonstration, we present a neural network that can be embedded into the on-sensor readout and filter out hits from low momentum tracks, reducing the detector's data volume by 57.1%–75.7%. The network is designed and simulated as a custom readout integrated circuit with 28 nm CMOS technology and is expected to operate at less than 300 μW with an area of less than 0.2 mm 2 . The temporal development of charge clusters is investigated to demonstrate possible future performance gains, and there is also a discussion of future algorithmic and technological improvements that could enhance efficiency, data reduction, and power per area.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

Distance-Weighted Graph Neural Networks on FPGAs for Real-Time Particle Reconstruction in High Energy Physics

Graph neural networks have been shown to achieve excellent performance for several crucial tasks in particle physics, such as charged particle tracking, jet tagging, and clustering. An important domain for the application of these networks is the FGPA-based first layer of real-time data filtering at the CERN Large Hadron Collider, which has strict latency and resource constraints. We discuss how to design distance-weighted graph networks that can be executed with a latency of less than one μs on an FPGA. To do so, we consider a representative task associated to particle reconstruction and identification in a next-generation calorimeter operating at a particle collider. We use a graph network architecture developed for such purposes, and apply additional simplifications to match the computing constraints of Level-1 trigger systems, including weight quantization. Using the hls4ml library, we convert the compressed models into firmware to be implemented on an FPGA. Performance of the synthesized models is presented both in terms of inference accuracy and resource usage.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗