Search NASA⌕ Search

SEARCH · Search NASA

Results for “deep generative models”

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 1,783 records · Page 99

Joint operations planning for space surveillance missions on the MSX satellite

The Midcourse Space Experiment (MSX) satellite, sponsored by BMDO, is intended to gather broad-band phenomenology data on missiles, plumes, naturally occurring earthlimb backgrounds and deep space backgrounds. In addition the MSX will be used to conduct functional demonstrations of space-based space surveillance. The JHU/Applied Physics Laboratory (APL), located in Laurel, MD, is the integrator and operator of the MSX satellite. APL will conduct all operations related to the MSX and is charged with the detailed operations planning required to implement all of the experiments run on the MSX except the space surveillance experiments. The non-surveillance operations are generally amenable to being defined months ahead of time and being scheduled on a monthly basis. Lincoln Laboratory, Massachusetts Institute of Technology (LL), located in Lexington, MA, is the provider of one of the principle MSX instruments, the Space-Based Visible (SBV) sensor, and the agency charged with implementing the space surveillance demonstrations on the MSX. The planning timelines for the space surveillance demonstrations are fundamentally different from those for the other experiments. They are generally amenable to being scheduled on a monthly basis, but the specific experiment sequence and pointing must be refined shortly before execution. This allocation of responsibilities to different organizations implies the need for a joint mission planning system for conducting space surveillance demonstrations. This paper details the iterative, joint planning system, based on passing responsibility for generating MSX commands for surveillance operations from APL to LL for specific scheduled operations. The joint planning system, including the generation of a budget for spacecraft resources to be used for surveillance events, has been successfully demonstrated during ground testing of the MSX and is being validated for MSX launch within the year. The planning system developed for the MSX forms a model possibly applicable to developing distributed mission planning systems for other multi-use satellites.

Stokes, Grant↗

Quantifying vertical wind shear effects in shallow cumulus clouds over Amazonia

Abstract. This study analyses and quantifies the effects of vertical wind shear (VWS) on the properties of shallow cumulus cloud fields over Central Amazonia. We perform idealised simulations with high resolution (50 m horizontally and 20 m vertically) using the Dutch Atmospheric Large-Eddy Simulation (DALES) model, changing the initial conditions and large-scale forcing of VWS. The resulting cloud field is analysed by applying a cloud tracking algorithm to generate Lagrangian datasets of the life cycle of individual clouds as well as their time-varying core and margin dimensions. The reference run has no wind speed or directional shear and represents a typical day in the local dry season. Numerical experiments with moderate and high wind speed shear are simulated by adding linear increases in the wind speed of 1.2 and 2.4 m s−1 km−1, respectively. Three additional runs are made by adding 90∘ of wind rotation between the surface and the top of the domain (5 km) on top of the three wind speed shear conditions. We find that clouds developing in a sheared environment have horizontal equivalent diameter increased by up to 100 m on average, but the cloud depth is reduced. Our quantification shows that VWS tends to increase the size of the cloud cores but reduces their relative area, volume, and mass fractions compared to the overall cloud dimensions. The addition of 2.4 m s−1 km−1 of VWS decreases the relative core area by about 0.03 (about 10 % of the overall average) and its volume and mass ratios by about 0.05 (10 %–25 % in relative terms). Relevant for the cloud transport properties is that the updraught speed and the liquid water content are lower within the cores, and consequently so is the upward mass flux. All quantifications of mean cloud properties point to the inhibition of convective strength by VWS, therefore hampering the shallow-to-deep transition. However, open questions still remain given that the individually deepest clouds were simulated under high environmental shear, even though they occur in small numbers. This could indicate other indirect effects of VWS that have opposite effects on cloud development if found to be significant in the future.

54 ENVIRONMENTAL SCIENCES↗

Conclusions of a Mini Technical Interchange Meeting on New Cross Risk Integration Projects Managed by the NASA Space Radiation Element

To enable deep space exploration and sustained human presence in space, the NASA Human Research Program’s (HRP) Space Radiation Element (SRE) funds research to characterize and mitigate adverse health outcomes from exposure to space radiation that include risks of carcinogenesis, cardiovascular disease, and central nervous system decrements. Recently, the SRE was tasked with supporting multiple HRP Elements with innovative and enabling projects to inform risk characterization, facilitate mitigation activities, and support crew health and performance. These projects, such as precision health initiative, NASA Omics Archive (NOA) and human sample repositories, are agnostic to any HRP Element, hence the name, Cross-Risk Integration Projects (CRIP). CRIP serves 3 broad purposes: - Services: Generate samples and data and manage the receipt, inventory, archive, and ultimate redistribution of biospecimens created in HRP-funded spaceflight and analog research activities—includes NASA Omics Archive project and various human and animal sample repositories. - Method Development: Identify and evaluate new-to-NASA research or analysis methods or techniques that could fundamentally improve existing or planned research efforts—includes the Translational Radiation Research and Countermeasures Project. - Enabling Capabilities: Demonstrate real-world application by adapting, adopting, and/or developing capabilities to benefit crew health and performance and improve risk mitigation—includes Precision Health Initiative (Pharmacogenomics) and advanced biological systems and engineered tissue microsystems initiative (tissue chips, organ-on-a-chip). To identify new technologies, future work, and solicitations, the SRE organizes themed sessions at annual HRP Investigators’ Workshops (IWS). These technical interchange meetings (TIMs) provide a venue for the scientific community to present ongoing work and engage in open discussion of results, limitations of current approaches, and incorporation of novel experimental strategies, model systems, and other innovative techniques. Here, a summary of the studies presented at the SRE-sponsored mini-TIM at the HRP IWS along with the goals and objectives of the CRIP projects is communicated. The 90-min TIM had 6 speakers who presented impressive novel ideas and work, some of which are funded under CRIP by the Space Radiation Element.

Janapriya Saha↗

Automated defect identification in electroluminescence images of solar modules

Solar photovoltaic (PV) modules are susceptible to manufacturing defects, mishandling problems or extreme weather events that can limit energy production or cause early device failure. Trained professionals use electroluminescence (EL) images to identify defects in modules, however, field surveys or inline image acquisition can generate millions of EL images, which are infeasible to analyze by rote inspection. Here, we develop a rapid automatic computer vision pipeline (~0.5 seconds/module) to analyze EL images and identify defects including cracks, intra-cell defects, oxygen-induced defects, and solder disconnections. Defect identification is achieved with a machine learning model (Random Forest, ResNet models and YOLO) trained on 762 manually-labeled EL images of PV modules. We compare model performance on an imbalanced real-world validation set containing 134 EL images and determine that ResNet18 and YOLO are the optimal models; we next evaluated these models on a dedicated testing set (129 module images) with resulting macro F1 scores of 0.83 (ResNet18) and 0.78 (YOLO). Using a field EL survey of a PV power plant damaged in a vegetation fire, we analyze 18,954 EL images (2.4 million cells) and inspect the spatial distribution of defects on the solar modules. The results find increased frequency of ‘crack’, ‘solder’ and ‘intra-cell’ defects on the edges of the solar module closest to the ground after fire. We also find an abnormal increase of striation rings on cells which were assumed to be caused mainly in fabrication process. Our methods are published as open-source software. It can also be used to identify other kinds of defects or process different types of solar cells with minor modification on models by transfer learning.

14 SOLAR ENERGY↗

ML-based Micro-CT SOFC Microstructure Models (from Kent 2026 Microstructural Augmentation paper)

Overview -------------------------- This repository contains datasets from the manuscript **"Enhanced Generalizability to Deep-Learning Quantification of 3D Microstructural Characteristics through Microstructurally Aware Augmentation of Scarce Data"** (*William F. Kent, Rochan Bajpai, Rachel C. Kurchin, William K. Epting, Harry W. Abernathy, Paul A. Salvador. Submitted 2026*). The methods are also described in the dissertation **Data Intensive Analysis of Solid Oxide Cell Microstructures** (*Doctoral dissertation, Carnegie Mellon University, 2025*). The datasets here are trained convolutional neural network (CNN) models for predicting key microstructural properties of solid oxide cell (SOC) electrodes from low-res, 2-channel 3D images, as well as some helpful code. The parameters for input images are provided in the paper. Sample data is provided in the file `Combined_anode_aug_dual_1k_examples` - that particular data was used to train `anode_all_aug.pth` and will work most accurately with that model. Please familiarize yourself with all caveats on accuracy and applicability, as detailed in the associated paper. Usage -------------------------- The basic usage is as follows, assuming `model_fn` is the path to the .pth file, and `X` is 2-channel input image(s) of the proper dimensions (either one image of shape `[2,12,24,24]`, or a batch of N input images of shape `[N,2,12,24,24]`): from CNN_inferencer import load_model_for_inference model = load_model_for_inference(model_fn) y_predicted = model(X) The model object automatically handles input scaling and output de-scaling based on the way the models were trained - in other words, pass in a 2-channel micro-CT image, and it will output microstructural property values in real units. ## Other model object attributes Note that model has useful attributes other than its forward pass model(X). * `model.output_descaler` - returns the output descaler object. Model does the de-scaling when generating inferences, but you may want to re-use this de-scaler on other values to e.g. compare predictions to ground truth from already-scaled training data. * `model.prop_names` - Gives the property names of the predicted y values, in order. Only exists if there's an output scaler as part of the model object, which there will be in the models provided here. ## Usage with sample data Here is a short script to use with the included sample data. from CNN_inferencer import display_predictions, load_model_for_inference, calculate_mape, parity_plot import h5py import numpy as np model_fn = 'anode_all_aug.pth' data_fn = 'Combined_anode_aug_dual_1k_examples.h5' N_samples = 200 figure_outdir = '.' model = load_model_for_inference(model_fn) with h5py.File(data_fn,'r') as f: XX = f['X'] #These are the 2-channel 3D images yy = f['y'] #These are the ground-truth microstructural properties, but they have been scaled for training - need to de-scale below N = XX.shape[0] #How many images total in the input data file #Run inferences on N_samples random samples from XX. #Run in a batch, much more efficient than one at a time. ii = np.random.choice(N,N_samples,replace=False) ii.sort() y_pred = model(XX[ii]) #Get the original/true (but normalized/scaled) values from the training dataset... #Because they were normalized, they are not in real units yet. So let's also de-scale them using model.output_scaler. y_true = model.output_scaler.transform(yy[ii]) #Let's display actual values for just 5 random ones for i in np.random.choice(N_samples,5,replace=False): display_predictions(y_true[i], y_pred[i], model.prop_names) #Make parity plots for each property (ground truth vs predicted values) #Also label each plot with the mean abs. percent error (MAPE) of the predicted values for i,key in enumerate(model.prop_names): mape = calculate_mape(y_true[:,i], y_pred[:,i]) parity_plot(y_true[:,i], y_pred[:,i], figure_outdir, key, extra_title=f' ({mape:.2f}% MAPE)')

3D microstructure↗

Prognostics for Systems Health Management - Model and Hybrid Based Approaches. Where are We Heading?

To facilitate and solve the prediction problem, awareness of the current state and health of the system is key, since it is necessary to perform condition-based system health predictions. To accurately predict the future state of any system, it is required to possess knowledge of its current health state and future operational conditional. In case of next generation electric aircrafts, computing remaining flying time is safety-critical, since an aircraft that runs out of power (battery charge) while in the air will eventually lose control leading to catastrophe. In order to tackle and solve the prediction problem, it is essential to have awareness of the current health state of the system, especially since it is necessary to perform condition-based predictions. To be able to predict the future state of the system, it is also required to possess knowledge of the current and future operational conditions and flight profiles for accurate estimation of end-of-discharge (EOD) for the batteries. Similar framework can be implemented to other complex systems and subsystems. Our research approach is to develop a system level health monitoring safety indicator which runs estimation and prediction algorithms to estimate remaining useful life predictions at system, subsystem swell as component levels. Given models of the current and future system behavior, a general approach of model-based prognostics is discussed as a solution to the prediction problem and further for decision making. Data driven prognostics approaches have been equally used with good results in the past, where respective approaches have their own challenges to tackle. This limits their applicability to complex real-world domains: (a) high complexity or incompleteness of physics-based models and (b) limited representativeness of the training dataset for data-driven models. With the advent of internet of things for data collection and increased use of ML algorithms, hybrid approaches are the next avenue to reduce the challenges and achieve better results. An hybrid framework for fusing information from physics-based performance models along with deep learning algorithms for prognostics of complex safety critical systems is presented. In this framework, we use physics-based performance models to infer unobservable model parameters related to the system's components health solving a calibration problem.

Prognostics↗

Particle-based high-temperature thermochemical energy storage reactors

Solar and other renewable energy driven gas-solid thermochemical energy storage (TCES) technology is a promising solution for the next generation energy storage systems due to its high operating temperature, efficient energy conversion, ultra-long storage duration, and potential high energy density. Experimental and theoretical studies suggest that the respective gravimetric and volumetric TCES energy storage densities vary from 200 to 3000 kJ kg –1 and 1–3 GJ m –3 . Solar radiation or heat generated from electric furnaces powered by renewable electricity can be stored in the form of chemical energy through endothermic reactions, while the stored chemical energy can be converted to thermal energy via an exothermic reaction when needed. The design of highly effective reactors requires a deep understanding of materials, thermodynamics, chemical kinetics, and transport phenomena. At time of writing, TCES reactors are yet to be deployed at commercially relevant scales, leaving a substantial gap between development efforts and commercial feasibility. Therefore, this review aims to examine the state-of-the-art design and performance of particle-based TCES reactors with different reactive materials. Fundamentals related to TCES reactive materials, reaction conditions, thermodynamics and kinetics, and transport phenomena are reviewed in detail to provide a comprehensive understanding of the reactor design and operation. Five major types of TCES reactors have been comprehensively reviewed and compared, including fixed, moving, rotary, fluidized, and entrained bed reactors. Most reported prototype reactors in the literature operate at lab scale with thermal inputs below 40 kW, and scaled TCES reactors (e.g., at megawatt level) are yet to be demonstrated. The nominal reactor operating temperatures range from 300 to 1500 °C, depending on the selected chemistry, reactive material, and heat sources. To evaluate their designs, the reactors are assessed in aspects of performance, cost, and durability. Discrepancies in performance indicators of energy storage density, extent of reaction, and various energy efficiencies are highlighted. The scale-up of reactors and power block integration, which hold the key to the successful commercialization of TCES systems, are critically analyzed. Furthermore, advanced materials (both reactive materials and ceramic reactor housing materials), effective particle flow control, advanced modeling tools, and novel system design may bring significant improvement to the energy efficiency, storage density and cost competitiveness of particle-based TCES reactors.

25 ENERGY STORAGE↗

Soil Desiccation Treatability Testing at BC Waste Disposal Cribs

During Hanford’s production period, low-level waste products generated from chemical processing of uranium fuel rods were discharged directly to the ground through a system of cribs and trenches located in the 200-BC-1 Operable Unit (OU). The site consists of 6 cribs and 20 trenches that received more than 117,000 m3 of radioactive liquid waste discharged to the soil. These unlined infiltration galleries held volumes of liquid waste while it seeped into the ground, with the understanding that the 100 m (330 ft) thick vadose zone in the area would effectively capture the effluent and prevent groundwater impacts. A conceptual model showing the operation of cribs and trenches is shown in Figure 1. Data show effluent from the 26 cribs and trenches containing about 410 curies of Technetium-99 (Tc-99) is primarily located between 30 m and 70 m (98 ft and 230 ft) depth (Corbin et al., 2005; Ward et al., 2004). Despite no evidence indicating that the contamination has reached the groundwater at BC cribs and trenches, the mobility of Tc-99 had been demonstrated in laboratory tests and was recognized as a threat to groundwater at the site. Using data from numerical models, laboratory analyses, field investigations, and information on historical discharges, the EPA and Ecology identified Tc-99 and U contamination of the vadose zone as a remediation priority. The U.S. DOE was notified by EPA and Ecology regarding risks associated with Tc-99 contamination in a letter requesting development of a strategy for improved methods to understand the nature and extent of vadose zone contamination, specifically Tc-99, and to develop remedial options for addressing such contamination. To develop the appropriate technology for characterizing, remediating, and monitoring the deep vadose zone Tc-99 contamination, the U.S. DOE worked with the EPA and Ecology to create a Treatability Test Plan under a Remedial Investigation/Feasibility Study (RI/FS) for the Hanford 200 Areas. Under this RI/FS, it was determined that a treatability test for soil desiccation should be carried out as it was identified as a promising in-situ treatment technology for mitigating risks posed by Tc-99 contamination to the groundwater table. The BC Cribs and Trenches site was identified as a representative site for Tc-99 and U contamination and selected for the soil desiccation treatability test. In this chapter, we summarize the overlying regulatory framework of RI/FS and treatability tests and illustrate how development and experimentation supported the evaluation of selected remedies. We briefly discuss the RI/FS for the 200 Areas of the Hanford Site and focus on the soil desiccation treatability testing performed at the BC cribs and trenches site under the Deep Vadose Zone Treatability Test Plan for the Hanford Central Plateau (DVZ-TT). The DVZ-TT is one component of the remedial investigation/feasibility study for the Hanford 200 Areas and represents the underlying regulatory framework that drives site operations towards records of decision and site closure.

Mangel, Adam R.↗

Issues in deep space radiation protection

The exposures in deep space are largely from the Galactic Cosmic Rays (GCR) for which there is as yet little biological experience. Mounting evidence indicates that conventional linear energy transfer (LET) defined protection quantities (quality factors) may not be appropriate for GCR ions. The available biological data indicates that aluminum alloy structures may generate inherently unhealthy internal spacecraft environments in the thickness range for space applications. Methods for optimization of spacecraft shielding and the associated role of materials selection are discussed. One material which may prove to be an important radiation protection material is hydrogenated carbon nanofibers. c 2001. Elsevier Science Ltd. All rights reserved.

short duration↗

Modeling hydrogen markets: Energy system model development status and decarbonization scenario results

Hydrogen can be used as an energy carrier and chemical feedstock to reduce greenhouse gas emissions, especially in difficult-to-decarbonize markets such as medium- and heavy-duty vehicles, aviation and maritime, iron and steel, and the production of fuels and chemicals. Significant literature has been accumulated on engineering-based assessments of various hydrogen technologies, and real-world projects are validating technology performance at larger scales and for low-carbon supply chains. While energy system models continue to be updated to track this progress, many are currently limited in their representation of hydrogen, and as a group they tend to generate highly variable results under decarbonization constraints. Here, the present work provides insights into the development status and decarbonization scenario results of 15 energy system models participating in study 37 of the Stanford Energy Modeling Forum (EMF37), focusing on the U.S. energy system. The models and scenario results vary widely in multiple respects: hydrogen technology representation, scope and type of hydrogen end-use markets, relative optimism of hydrogen technology input assumptions, and market uptake results reported for 2050 under various decarbonization assumptions. Most models report hydrogen market uptake increasing with decarbonization constraints, though some models report high carbon prices being required to achieve these increases and some find hydrogen does not compete well when assuming optimistic assumptions for all advanced decarbonization technologies. Across various scenarios, hydrogen market success tends to have an inverse relationship to success with direct air capture (DAC) and carbon capture and storage (CCS) technologies. While most model-scenario combinations predict modest hydrogen uptake by 2050 – <10 million metric tons (MMT) – aggregating the top 10 % of market uptake results across sectors suggests an upper range demand potential of 42–223 MMT. The high degree of variability across both modeling methods and market uptake results suggests that increased harmonization of both input assumptions and subsector competition scope would lead to more consistent results across energy system models. The wide variability in results indicates strongly divergent conclusions on the role of hydrogen in a decarbonized energy future.

08 HYDROGEN↗

SU-E-T-496: A Study of Two Commercial Dose Calculation Algorithms in Low Density Phantom

Purpose: Some lung cancer patients have very low lung density due to comorbidities. We investigate calculation accuracy of Eclipse AAA and Acuros(AXB) using a phantom that simulates this situation. Methods: A 2.5 x 5.0 x 5 cm (long) solid water inhomogeneity positioned 10 cm deep in a Balsa lung phantom (density 0.099 gm/cc) was irradiated with an off-center field such that the central axis was parallel to one side of the inhomogeneity. Radiochromic films were placed at 2.5cm(S1) and 5cm(S2) depths. After CT scanning, Hounsfield Units(HU) were converted to electron(ρe) and mass(ρm) density using in-house(IH) and vendor-supplied(V) calibration curves. IH electron densities were generated using a commercial electron density phantom. The phantom was exposed to 6 MV 3x3 and 20x20 fields. Dose distributions were calculated using the AAA and AXB algorithms. Results: The HU of BW is -910±40 which translates to ρe of 0.088±0.050(IH) and 0.090±0.050(V), and ρm of 0.101±0.045(IH) and 0.103±0.039(V). Both ρe(V) and ρm(V) are higher than ρe(IH) and ρm(IH) respectively by 1.4-5.3% and 0.5-12.3%. The average calculated dose inside the solid water ‘tumor’ are within 3.7% and 2.4% of measurements for both calibrations and field sizes using AAA and AXB. Within 10mm outside the ‘tumor’, AAA on average underestimates by 18.3% and 17.0% respectively for 3x3 using IH and V. AXB underestimates by 5.9%(S1)-6.6%(S2) and 13.1%(S1)-16.0%(S2) respectively using IH and V. For 20x20, AAA and AXB underestimate by 2.8%(S1)-4.4%(S2) and 0.3%(S1)-1.4%(S2) respectively with either calibration. Conclusion: The difference in the HU calibration between V and IH is not of clinical significance in normal field sizes. In the low density region of small fields, the calculations from both algorithms differ significantly from measurements. This may be attributed to the insufficient lateral electron transport modeled by two algorithms resulting in the over-estimation in penumbra expansion.

60 APPLIED LIFE SCIENCES↗

A Warm Garage for a Lunar Rover

Approach: One approach to heating a rover during the lunar night is the so-called thermal wadis concept [1]. This involves heating the regolith with solar concentrators and placing the rover on the heated surface for the night. Since the regolith is heated by a relatively weak heat flux, a high thermal conductivity is required for heating a sufficiently large mass of regolith. However, lunar regolith has a low thermal conductivity. Therefore, the concept involves increasing the conductivity by sintering the regolith, which requires a significant energy input and complex procedures. Here we propose an alternative approach where the low thermal conductivity of regolith is an advantage. Specifically, we propose to use a highly exothermic combustible mixture for heat generation. The mixture pellets are placed in the surface layer of regolith and ignited. The combustion forms condensed products and releases heat, which then slowly spreads to the surrounding regolith. Heat can also be transferred, for example, by heat pipes, into radiant heating surfaces installed on the ground. A greenhouse that transmits sunlight during the day and decreases the radiative heat losses during the night can also be installed. Selection of the Heat-generating Mixture: The reactive mixture should have a high specific energy and generate only condensed products since gases could disturb the regolith layer, carry enthalpy out of the system, and lead to an explosion. There are mixtures, (sometimes called pyrolants) that possess very high specific energies. One example is magnesium-Teflon-Viton mixtures used in flares. However, they produce gases and may cause explosions. Other mixtures that include magnesium cannot be used either because of the high vapor pressure of Mg at temperatures well below the combustion temperature. Recently, mixtures that involve lithium peroxide (Li2O2) have been proposedfor using in space power systems [2]. However, they produce lithium oxide (Li2O), which boils at 2800 K at 1 atm and hence at a lower temperature in vacuum. Fortunately, there exist many mixtures that release a lot of heat and form only condensed products during the combustion. Many such mixtures have been used for self-propagating high-temperature synthesis (SHS) of various materials [3, 4]. For the application discussed here, t itanium/boron (1:2 mole ratio) mixture appears to be particularly promising. The specific energy is 4.0 MJ/kg (1.1 kWh/kg), the adiabatic flame temperature is about 3200 K, and the reaction forms solid titanium diboride (TiB2, melting point: 3500 K). The mixture can be ignited easily with a heated tungsten wire, and it has been used widely as a booster to ignite the main mixture in the SHS process.Estimates: Assuming that specific heat of regolith is 500 J/(kg∙K) [5] and all generated heat is transferred to regolith, 12.5 kg of the Ti/B mixture would increase the temperature of 1000 kg of regolith by 100 K. To evaluate the rate of heat transfer in the regolith, a spherical model was analyzed where the heat released by a 12.5 kg Ti/B core propagates by thermal conduction through a 1000 kg regolith shell with no heat loss from its outer surface. At a bulk density of 1500 kg/m3 [5], the radius of the shell was 54 cm, while the radius of the core was about 11 cm. The calculations were conducted using Thermal Desktop SINDA/FLUINT (Cullimore and Ring Technologies) software at two constant values of bulk thermal conductivityk of the regolith: 0.001 and 0.01 W/(m∙K). The results show that after 14.5 days the core lost 31% of the released heat at the lower k and 77% at the higher k. At a distance of 20 cm from the core surface, the temperature of the regolith increased by only 1 K at the lower k and by 132 K at the higher k. In reality, the regolith near the heat source will be melted, so its thermal conductivity will increase significantly. Nevertheless, the conducted estimates indicate that combustion-based heat generators, placed directly in the regolith, could provide heat during a rather long period such as the lunar night.Conclusion: Heat generators based on gasless combustion of highly energetic reactive mixtures could be installed directly in the surface layer of lunar regolith. Because of the low thermal conductivity of the regolith, such generators would keep thermal energy for days and gradually supply heat to a rover/lander.Acknowledgment: The material presented in this work is based upon the work supported by National Aeronautics and Space Administration (NASA) under Grant #80NSSC20K0293.References: [1] Balasubramaniam R. et al. (2011) J. Thermophys. Heat Trans., 25,130−139. [2] Blair R.G. and Vasu S.S. (2022) Conf. Advanced Power Systems for Deep Space Exploration. [3] Varma A. et al. (1998) Adv. Chem. Eng., 24,79−226. [4] Levashov E.A. et al. (2017) Int. Mater. Rev., 62,203−239. [5] Wood-Robinson R. et al. (2019) J. Geophys. Res. Planets, 124, 1989−2011.

lunar↗

Why it is Unfortunate that Linear Machine Learning “Works” so well in Electromechanical Switching of Ferroelectric Thin Films

Machine learning (ML) is relied on for materials spectroscopy. It is challenging to make ML models fail because statistical correlations can mimic the physics without causality. Here, using a benchmark band-excitation piezoresponse force microscopy polarization spectroscopy (BEPS) dataset the pitfalls of the so-called “better”, “faster”, and “less-biased” ML of electromechanical switching are demonstrated and overcome. Using a toy and real experimental dataset, it is demonstrated how linear nontemporal ML methods result in physically reasonable embedding (eigenvalues) while producing nonsensical eigenvectors and generated spectra, promoting misleading interpretations. A new method of unsupervised multimodal hyperspectral analysis of BEPS is demonstrated using long-short-term memory (LSTM) β-variational autoencoders (β-VAEs) . By including LSTM neurons, the ordinal nature of ferroelectric switching is considered. Further, to improve the interpretability of the latent space, a variational Kullback–Leibler-divergency regularization is imposed . Finally, regularization scheduling of β as a disentanglement metric is leveraged to reduce user bias. Combining these experiment-inspired modifications enables the automated detection of ferroelectric switching mechanisms, including a complex two-step, three-state one. Ultimately, this work provides a robust ML method for the rapid discovery of electromechanical switching mechanisms in ferroelectrics and is applicable to other multimodal hyperspectral materials spectroscopies.

36 MATERIALS SCIENCE↗

Multidimensional Distributional Neural Network Output Demonstrated in Super‐Resolution of Surface Wind Speed

Accurate quantification of uncertainty in neural network predictions remains a central challenge for scientific applications involving high-dimensional, correlated data. While existing methods capture either aleatoric or epistemic uncertainty, few offer closed-form, multidimensional distributions that preserve spatial correlation while remaining computationally tractable. In this work, we present a framework for training neural networks with a multidimensional Gaussian loss, generating a closed-form predictive distribution over outputs informed by non-identically distributed training data. Our approach captures aleatoric uncertainty by iteratively estimating the means and covariance matrices, and is demonstrated on a super-resolution example out-of-training-sample. We leverage a Fourier representation of the covariance matrix to stabilize network training and preserve spatial correlation. We introduce a novel regularization strategy—referred to as information sharing—that interpolates between image-specific and global covariance estimates, enabling convergence of the super-resolution downscaling network trained on image-specific distributional loss functions. This framework allows for efficient sampling, explicit correlation modeling, and extensions to more complex distribution families all without disrupting prediction performance. We demonstrate the method on a surface wind speed downscaling task and discuss its broader applicability to uncertainty-aware prediction in scientific models.

17 WIND ENERGY↗

Building a 40% Energy Saving House in the Mixed-Humid Climate

This report describes a home that uses 40% less energy than the energy-efficient Building America standard - a giant step in the pursuit of affordable near-zero-energy housing through the evolution of five near-zero-energy research houses. This four-bedroom, two-bath, 1232-ft2 house has a Home Energy Rating System (HERS) index of 35 (a HERS rating of 0 is a zero-energy house, a conventional new house would have a HERS rating of 100), which qualifies it for federal energy efficiency and solar incentives. The house is leading to the planned construction of a similar home in Greensburg, Kansas, and 21 staff houses in the Walden Reserve, a 7000-unit "deep green" community in Cookville, Tennessee. Discussions are underway for construction of similar houses in Charleston, South Carolina, Seattle, Washington, Knoxville and Oak Ridge, Tennessee, and upstate New York. This house should lead to a 40% and 50% Gate-3, Mixed-Humid-Climate Joule for the DOE Building America Program. The house is constructed with structurally-insulated-panel walls and roof, raised metal-seam roof with infrared reflective coating, airtight envelope (1.65 air changes per hour at 50 Pascal), supply mechanical ventilation, ducts inside the conditioned space, extensive moisture control package, foundation geothermal space heating and cooling system, ZEHcor wall, solar water heater, and a 2.2 kWp grid-connected photovoltaic (PV) system. The detailed specifications for the envelope and the equipment used in ZEH5 compared to all the houses in this series are shown in Tables 1 and 2. Based on a validated computer simulation of ZEH5 with typical occupancy patterns and energy services for four occupants, energy for this all-electric house is predicted to cost only $0.66/day ($0.86/day counting the hookup charges). By contrast, the benchmark house would require $3.56/day, including hookup charges (these costs are based on a 2006 residential rates of $0.07/kWh and solar buyback at $0.15/kWh). The solar fraction for this home located in Lenoir City, Tennessee, is predicted to be as high as 41%(accounting for both solar PV and the solar water heater). This all-electric home is predicted to use 25 kWh/day based on the one year of measured data used to calibrate a whole-building simulation model. Based on two years of measured data, the roof-mounted 2.2 kWp PV system is predicted to generate 7.5 kWh/day. The 2005 cost to commercially construct ZEH5, including builder profit and overhead, is estimated at about $150,000. This cost - for ZEH5's panelized construction, premanufactured utility wall (ZEHcor), foundation geothermal system, and the addition of the walkout lower level, and considering the falling cost for PV - suggests that the construction cost per ft2 for a ZEH5 two-story will be even more cost-competitive. The 2005 construction cost estimate for a finished-out ZEH5 with 2632 ft2 is $222,000 or $85/ft2. The intention of this report is to help builders and homeowners make the decision to build zero-energy-ready homes. Detailed drawings, specifications, and lessons learned in the construction and analysis of data from about 100 sensors monitoring thermal performance for a one-year period are presented. This information should be specifically useful to those considering structural insulated panel walls and roof, foundation geothermal space heating and cooling, solar water heater and roof-mounted, photovoltaic, grid-tied systems.

14 SOLAR ENERGY↗

Performance and transport in the ARC tokamak

The ARC TM tokamak, a high-field (𝐵 𝑇 = 11.4 T) fusion power plant, under development by Commonwealth Fusion Systems, is studied using a suite of integrated modelling tools to predict its fusion power generation (𝑃𝑓⁡𝑢⁢𝑠), transport and confinement properties. Analysis is based off an ARC operational point scoped first with zero-dimensional (0-D) plasma operational contour (POPCON) modelling to produce 1.13 GW of fusion power. A suite of integrated modelling tools (TRANSP, ASTRA and TORAX) were applied to predict the performance and kinetic profiles of the ARC design point, yielding a range of predicted performance spanning from ∼900 to 1300 MW in rough quantitative agreement with POPCON predictions. The sensitivity of these results to uncertain modelling inputs was probed using scans of pedestal boundary conditions around EPED-predicted values (total pressure and temperature ratios), tungsten concentration and seperatrix density around their nominal assumptions. Pedestal pressure and pedestal top (𝑇 𝑖 /𝑇 𝑒 ) play a large role in 1.5-dimensional performance predictions, able to modify the predicted 𝑃 𝑓⁡𝑢⁢𝑠 by a factor of 2 within reasonable assumptions. High-fidelity core nonlinear gyrokinetic profile predictions, performed using CGYRO (Candy et al. 2016 J. Comput. Phys., vol. 324, pp. 73–93) coupled with the PORTALS (Rodriguez-Fernandez et al. 2024 Nucl. Fusion, vol. 64, 076034; Phys. Plasmas, vol. 31, 2024, 062501) framework, yield substantially lower performance (𝑃 𝑓⁡𝑢⁢𝑠 =677 MW) compared with 0-D and medium-fidelity modelling for nominal assumptions, showing that there is non-negligible uncertainty between models and that future work on SPARC may help resolve discrepancies. Lower overall performance results from significantly reduced volume-averaged densities and temperatures, along with reduced levels of density and temperature peaking. Turbulence and transport are largely dominated by ion temperature gradient across the profile, confirmed by both linear stability and the response of the nonlinear fluxes to changes in gradients, with some impact of kinetic ballooning modes in the deep core. This work represents one of the most complete scoping of potential fusion power plant conditions performed to date. The extensive integrated modelling provides confidence in ARC performance approaching 1 GW, while nonlinear gyrokinetic modelling results in open questions into the physics of density and temperature peaking in fusion-power-plant-relevant operational space. A discussion of results and the role that the SPARC tokamak (Creely et al. 2020 J. Plasma Phys., vol. 86, 865860502) will play in informing ARC design, performance and operation is presented.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Tracking Aerosol Convective Interactions Experiment (TRACER) Science Plan

Convective clouds play an important role in the Earth’s climate system as a driver of large-scale circulations and a primary mechanism for the transport of heat, moisture, aerosols, and momentum throughout the troposphere. Despite their climatic importance, multi-scale models continue to have persistent biases produced by insufficient representation of convective clouds. This is the result of an incomplete understanding of key processes such as convective initiation, updraft and downdraft dynamics, cloud and precipitation microphysics, and aerosol-convection interactions. The Aerosol-Cloud-Precipitation-Climate Initiative, an international research group dedicated to advancing understanding of aerosol impacts on clouds relevant to climate, has identified the Houston, Texas region as an optimal location for targeted studies of aerosol-convection interactions within frequently developing isolated deep convection. Houston lies within a humid subtropical climate regime, where onshore flow and sea-breeze convection interact with a range of aerosol conditions associated with Houston’s urban and industrial emissions. Pilot studies have suggested that convective clouds in this region are potentially significantly impacted by the varying aerosol conditions. To increase our understanding of convective cloud life cycles and aerosol-convection interactions, the Tracking Aerosol Convection Interactions Experiment (TRACER) aims to collect a comprehensive data set focused on the evolution of convective clouds and the environment (including aerosol, cloud, thermodynamics, and lightning) in which the clouds initiate, grow, and decay. A unique component of TRACER is that a large number of individual, isolated convective cells will be tracked and measured in high spatial and temporal resolution for the purposes of: (i) Characterizing and linking convective cloud kinematic and microphysical life cycles, (ii) Quantifying environmental thermodynamic and kinematic controls on convective life cycle properties, (iii) Isolating and quantifying the impacts of aerosol properties on convective cloud kinematic and microphysical evolution. TRACER includes a one-year deployment of the U.S. Department of Energy (DOE) Atmospheric Radiation Measurement (ARM) user facility’s first ARM Mobile Facility (AMF1) and second-generation C-Band Scanning ARM Precipitation Radar (CSAPR2) aimed at collecting statistical data sets of cloud, precipitation, atmospheric state, and aerosol under varying aerosol loading and local circulation (sea breeze and urban heat island) conditions. A four-month intensive operational period (IOP) from June through September will include deployment of an ancillary site with aerosol, atmospheric state, and precipitation measurements at a location rarely impacted by Houston’s anthropogenic emissions, which will be commonplace near the AMF1 site. Convective cell tracking by the CSAPR2 will provide high-resolution polarimetric and velocity observations under a range of environmental (including aerosol loading) conditions. High-frequency radiosonde launches will capture quickly evolving thermodynamic and kinematic conditions near convective cells, a requirement for isolating aerosol effects on clouds. TRACER will additionally leverage a lightning mapping array, surface meteorological networks, and air pollution networks, and proposals will be entered to deploy additional mobile radar and radiosonde assets during the IOP. This unique combination of cloud, precipitation, lightning, aerosol, and atmospheric state measurements associated with tracked convective cells will ultimately improve our understanding of the convective cloud life cycle and its interaction with individual environmental factors such that improved, next-generation cumulus, microphysics, turbulence, and aerosol parameterizations can be designed.

Convective clouds, aerosols, cloud life cycle, atm↗

Neutron Cross Section Measurement In The Protodune-SP Experiment

Understanding the detector response to neutrons will be critical for performing neutrino oscillation analyses in the next-generation Deep Underground Neutrino Experiment (DUNE). The DUNE physics program is centered around measuring the neutrino flavor composition as a function of their energy both at the near and the far detector. Neutrinos in the DUNE beam will have energies ranging between 100~MeV and 10~GeV, which is significant, because individual neutrino energies will not be known beforehand and will have to be reconstructed. Neutrino interactions in DUNE will produce leptons and hadrons -- including protons, pions, and neutrons. Neutrons can transport energy away from their origin and sometimes go undetected. In addition to the primary neutrons produced by the neutrino, subsequent interactions of the charged hadrons can result in secondary neutrons. Neutrons are a source of missing energy and will bias the neutrino energy measurement. Currently, there is also a 20\% energy scale uncertainty and a 40\% uncertainty on the energy resolution for neutrons in DUNE, which must be addressed. ProtoDUNE Single-Phase (ProtoDUNE-SP) is a 770-ton prototype for the DUNE far detector and was designed to both validate the technology that will be employed in DUNE and to measure cross sections for the charged hadrons (pions, protons, and kaons) at the relevant energies for DUNE. The ProtoDUNE-SP experiment, therefore, is in a unique position to characterize the secondary neutron component for DUNE. This is achieved by searching for candidate neutron interactions in ProtoDUNE-SP events and using these to facilitate a measurement of the neutron inelastic cross section as well as an estimate of the neutron energy and number. The cross section measurement presented here is based on neutrons produced in 1~GeV, $\pi^+$ events captured in 2018 by ProtoDUNE-SP in accordance with the production and cross section models in the GEANT4 simulation toolkit, version 4.1 0.6p1. T he best-fit neutron inelastic cross section, in the kinetic energy range of 114 to 314~MeV, is $1.24_{-0.08}^{+0.10}$~(stat.~$\oplus$~syst.)~barns.

Rivera, David Orlando↗