Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Batch”

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 865 records · Page 48
All sourcesNASA NTRSDOE OSTI
DOE OSTI2026

The solubility of ErPO 4 and Er speciation in hydrothermal fluids at varying pH and salinity between 350 and 450 °C

The rare earth elements (REE) are important for the green-energy transition and can be incorporated into the REE phosphates, such as xenotime-(Y), which also hosts heavy REE (Tb– Lu). Xenotime-(Y) is a common accessory mineral in metamorphic rocks and a range of mineral deposits where it controls the mobility of heavy REE, however, the impact of high temperature aqueous fluids on the behavior of heavy REE is largely unknown. Thermodynamic modeling can be utilized as a tool to predict the mobility of REE in hydrothermal aqueous fluids, but must be supported by accurate experimental data. Here, we measured the solubility of endmember synthetic xenotime-structured ErPO4 in NaCl-HCl-NaOH-bearing aqueous solutions at 350 °C and water vapor saturation pressure, at 400 and at 450 °C and 500 bar using batch-type Inconel reactors. Erbium speciation was investigated as a function of pH from 2.8 to 8, where Er chloride species are predominant at acidic conditions (pH <3) and Er hydroxyl complexes are predominant at near- neutral to alkaline conditions (pH >3). At pH 7–9, the measured ErPO 4 solubility (-9.8 to -7.5 log m Er ) is up to 2.5 orders of magnitude lower than thermodynamic predictions (-9.4 to -6.7 log m Er ) using existing thermodynamic databases. At pH 2–3, the predicted ErPO 4 solubility is ~0.5 orders of magnitude higher at 350 °C and ~1 order of magnitude lower at 450 °C compared to experimentally measured Er concentrations. The thermodynamic properties of aqueous Er species were therefore revised in this study. The partial molal Gibbs energy of formation (Δ f G 0 T,P ) for aqueous Er hydroxyl and chloride species are optimized using GEMSFITS and the logarithmic formation constants (logβ n ( Cl,OH) ) were derived at each experimental temperature and pressure. The updated thermodynamic properties for Er hydroxyl species (Er(OH) +2 , Er(OH) 2 + , and Er(OH) 3 0 ) show that their stability shifts to more acidic conditions at and below 400 °C. The Er chloride species (ErCl +2 and ErCl 2 + ) show increased stability compared to Er hydroxyl species at temperatures of 450 °C and 0.01 mol/kg NaCl. The updated thermodynamic properties are implemented into the GEM-Selektor modeling package to investigate the mobility of Er in saline hydrothermal fluids in equilibrium with alkaline rocks. Importantly, the updated properties for Er hydroxyl species result in low Er solubility at rock equilibrated pH conditions due to an expanded hydroxyl predominance zone, but lower aqueous complex stability overall, whereas previous models suggest greater stability for aqueous Er species. Furthermore, ErPO 4 solubility increases with decreasing temperature due to the deprotonation of HCl, which increases the acidity of hydrothermal fluids and the availability of Cl - to complex with the REE. These simulations highlight how fluid-rock reaction and temperature affect the mobility of REE in hydrothermal ore-forming systems.

58 GEOSCIENCES↗
DOE OSTI2026

The solubility and speciation of REE phosphate endmembers (CePO 4 and YPO 4 ) in Cl-rich aqueous fluids from 350 to 450 °C and implications for natural systems

The rare earth elements (REE) are important metals used increasingly in advanced technologies. Within the crust, the elements Ce and Y are commonly more abundant compared to other lanthanides and comprise important end-member constituents of REE-bearing minerals. Specifically, Ce is part of the light (L) REE which have larger ionic radii than the heavy (H) REE, which are grouped together with Y. These differences in ionic radius can lead to important physico-chemical trends within the lanthanide group. Despite a recent increase in experimental and thermodynamic data for the REE at high temperature and pressure, there is still a significant lack of these data at supercritical conditions. In this study we conducted batch-type experiments to measure the solubility of REE phosphates (CePO 4 and YPO 4 ) at varying starting pH (1.5–10), and salinity (0.01–1.4 mol/kg NaCl) at 350 and P sat , and from 400 to 450 °C at 700 bar. Results show that the solubility of Ce (33–0.14 ppb) is generally higher than Y (13–0.13 ppb) and that Ce complexes more strongly with both chloride and hydroxyl ligands compared to Y. The solubilities of both REE phosphates are highly pH-dependent and, to a lesser extent, depend on salinity at the studied conditions. The solubility data from this study were implemented into the GEMSFITS program to optimize the thermodynamic properties of Ce and Y hydroxyl and chloride species. The updated standard partial molal Gibbs energies of formation (Δ f G 0 T,P ) are used within the experimental temperature and pressure range to accurately predict the CePO 4 and YPO 4 solubility and Ce and Y speciation behavior. Based on the updated thermodynamic properties we also provide formation constants (log β n Cl,OH ) for Ce and Y hydroxyl and chloride species. Updated thermodynamic properties are applied to model REE-apatite dissolution and REE mobility based on the Pea Ridge iron oxide apatite deposit in Missouri, USA. The apatite dissolution model replicates natural observations including the replacement of monazite and xenotime after apatite and is an example of the utility of the new thermodynamic constants applied to supercritical crustal fluids. Furthermore, the findings of this study advance the predictive capabilities of geochemical models, our understanding of the behavior of individual REE, and permit modeling the overarching fractionation trends between LREE and HREE in supercritical crustal fluids.

58 GEOSCIENCES↗
DOE OSTI2021

Temperature-responsive smart tracers for field-measurement of inter-well thermal evolution: Heterogeneous kinetics and field demonstration

Temperature-responsive smart tracers enable advanced warning of a “premature thermal breakthrough” and may therefore improve reservoir management and reduce financial risk. A successful calculation of inter-well fluid temperatures requires that temperature-dependent kinetics are known, which typically result from homogeneous batch reactor experiments. However, recent meso-scale field experiments at the Altona Field Laboratory involving fluid flow in a discretely-fractured reservoir suggest that silica-fluid interfaces may accelerate hydrolysis kinetics. Here, the Arrhenius parameters of phenyl acetate hydrolysis are investigated under heterogeneous reaction conditions in packed-bed column experiments. The breakthrough curve of the reaction product is compared to an inert reference tracer (carbon-cored nanoparticles). Temperatures experienced during field testing ranging from 10 to 40 °C were studied using a phosphate buffer solution to maintain a near-neutral pH, as found in the reservoir. The empirically-determined pre-exponential factor and activation energy values are subsequently used in a first-order kinetic model to improve calculations of effective reservoir temperatures for these meso-scale field tests. Here, the results suggest that hydrolysis reactions can directly monitor inter-well reservoir temperatures, but only if the influence of solid-fluid interactions are carefully accounted for.

15 GEOTHERMAL ENERGY↗
DOE OSTI2022

Thermal and solubility effects on fault leakage during geologic carbon storage

Geologic carbon storage (GCS) is a promising method for reducing anthropogenic CO 2 emissions to the atmosphere. To safely deploy GCS in the field, it is necessary to assess risks and the effect of uncertainty on safe storage. The effect of uncertainty can be quantified using batches of simulations, but the high computational costs of high-resolution simulations necessitate use of reduced-order models (ROMs). Previous work involves ROMs for quantifying the risk of different potential leakage paths from storage reservoirs to shallow formations. However, previous studies on development of fault-leakage ROMs have limited numbers of uncertain parameters and do not explicitly examine impacts of CO 2 solubility and thermal stresses on fault reactivation, which can generate high-permeability pathways and compromise CO 2 storage. Here, we analyze an ensemble of simulations considering CO 2 leakage from a storage reservoir to a shallow aquifer through a fault while varying a number of uncertain parameters related to thermo-hydro-mechanical properties and CO 2 injection. We show the effects of solubility on: free-phase CO 2 -leakage rates, brine-leakage rates, and poroelastic fault destabilization. We find that CO 2 solubility is more important for estimating free-phase CO 2 -leakage rates compared to brine-leakage rates or poroelastic fault destabilization. We also find that thermal stresses and overpressures have different spatial distributions within the fault, indicating that the spatial variability of overpressures due to variation in flow parameters does not necessarily make the spatial variability of thermal stresses negligible. We suggest the use of the CO 2 phase-change path as a variable in future fault-leakage ROMs.

03 NATURAL GAS↗
DOE OSTI2024

Particle-wall heat transfer in narrow-channel bubbling fluidized beds for thermal energy storage

Robust metal oxide particles can provide low-cost and stable thermal energy storage (TES) to temperatures up to 1200 °C or higher. The transfer of heat into and out of the particles in cost-effective high-temperature particle heat exchangers remains a principal challenge to implementing particle-based TES. The present work expands on prior studies of particle-wall heat transfer in narrow-channel fluidized beds operated in the bubbling fluidization regime. Batch-mode experiments with various oxide particles over a range of temperatures and airflow rates indicate that particle-wall heat transfer increases with higher bed temperature and decreasing particle size. Measured particle-wall heat transfer coefficients in a 12 mm deep channel are fit to a Nusselt number correlation proportional to a non-monotonic function of excess fluidization velocity. Particle-wall heat transfer coefficients rise rapidly with excess fluidization velocities until reaching a maximum at intermediate air velocities due to a trade-off between enhanced transverse particle mixing and decreasing particle volume fraction with increased fluidization velocity. The heat transfer coefficient reaches a maximum of ≈ 400% of values without fluidization for oxide particles ranging in diameter d p from 159 to 408 μm. For the smallest particles tested, composed of olivine sand, particle-wall heat transfer coefficients peak above 1100 Wm –2 K –1 at 450 °C. 100-h tests at 500 °C and near-optimal heat transfer fluidization velocities indicated minimal wall wear or oxide scaling. Furthermore, high particle-wall heat transfer coefficients suggest narrow channel fluidized beds as a potential pathway for reducing the required surface area and cost of high-temperature particle heat exchangers, as needed in large-scale thermal energy storage applications.

14 SOLAR ENERGY↗
DOE OSTI2021

Characterization of a selective, zero power sensor for distributed sensing of hydrogen in energy applications

The use of hydrogen as a clean and renewable energy source is increasing rapidly for both vehicle and stationary applications. There are safety concerns for the locations in which hydrogen is made, used, and transported (i.e., pipelines and tanker trucks). Sensors are needed to comply with safety regulations and to enable a smooth and safe rollout of hydrogen as an alternative energy. However, hydrogen sensors do not yet exist that have the combined features of small size and low power for easy deployment coupled with high-volume manufacturability and low cost. This is necessary to accommodate the emerging fixed and mobile markets while retaining critical metrological metrics, including measurement range, detection limits, selectivity, fast response, stability, and long lifetime. An amperometric gas sensor for hydrogen (AGS) has been developed using an innovative manufacturing method. The sensor was designed using scalable fabrication strategies based on “Printed Electronics” (PE) methodology which are compatible with large-scale production. Prototype sensors were batch fabricated with multiple individual elements on a substrate compatible in size with a standard 8-inch wafer to enable high-volume, low-cost manufacturing, thereby leveraging PE and semiconductor fabrication infrastructure. This novel AGS was instrumented with control circuitry and evaluated for hydrogen detection in the range of 0–5000 ppmv H2 in air. Specific performance evaluations included assessment of the sensor measurement range, repeatability, and selectivity.

08 HYDROGEN↗
DOE OSTI2026

Hydrogen production from full-strength corn stover fermentation effluent in single-chamber replaceable-cathode microbial electrolysis cells

Lignocellulosic residual biomass generated by the agricultural sector is an abundant feedstock for biohydrogen production via dark fermentation. However, this process is intrinsically inefficient, converting only ~30% of the reductant energy into H2 and leaving substantial amounts of reduced byproducts. These byproducts, mostly found in the fermentation effluents, can be further valorized in microbial electrolysis cells (MECs) to enhance the overall H2 recovery. However, current MEC configurations are typically dual- or single-chamber systems, yet both suffer from key inefficiencies. Dual-chamber systems rely on proton exchange membranes that are costly and prone to rapid biofouling, whereas single-chamber, membraneless systems are limited by reduced productivity due to H2 recycling and methanogenic consumption In this study, three single-chamber, 50-mL replaceable-cathode microbial electrolysis cells (RC-MECs) were 3D-printed and equipped with a physical separator to isolate anode and cathode compartments and limit H2 migration. Full-strength milled corn stover (MCS) fermentation effluent (COD of ~23.8 g-COD/L) was treated in fed-batch mode over two operational periods spanning 21 (Run 1) and 80 (Run 2) days. The RC-MECs exhibited comparable performance in both runs: after biofilm maturation, current densities exceeded 100 A/m²_cathode, COD removal reached up to 43%. Notably, extended RC-MECs operation led to a substantial methanogenic activity with the CH4 fraction in the cathode gas increasing to as high as 80% of the total biogas. Additions of a methanogenesis inhibitor 2-bromoethanesulfonate (2-BES) produced transient increases in hydrogen yields (11.51 and 5.12 L-H2/L_reactor/day in Runs 1 and 2, respectively); however, sustained 2-BES addition in subsequent cycles reduced total biogas production, decreased COD removal, and led to volatile fatty acid accumulation. Overall, single-chamber MECs can treat high-strength dark fermentation effluents while improving H2 recovery, but methanogenesis remains a key bottleneck, and complete long-term inhibition may be operationally unsustainable.

Hydrogen Production↗
DOE OSTI2020

Preparation of 241 Am/ 243 Am gravimetric mixtures and development of Am isotopic and assay measurement techniques using thermal ionization mass spectrometry

Three gravimetric mixtures with 241 Am/ 243 Am isotope ratios at nominally 1:1, 20:1, and 200:1 were prepared for calibration of the thermal ionization mass spectrometer instruments used for americium isotopic and assay measurements by isotope dilution mass spectrometry. The total evaporation analytical technique was developed for high-accuracy and precision measurements of Am isotopic ratios. The technique was also applied to Am assay measurements using isotope dilution mass spectrometry. The Am isotope ratio and assay measurement techniques were utilized to characterize batches of americium oxide separated from plutonium materials in storage at Los Alamos National Laboratory in support of commercial use of 241 AmO 2 .

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2023

An integrated data management and informatics framework for continuous drug product manufacturing processes: A case study on two pilot plants

The pharmaceutical industry continuously looks for ways to improve its development and manufacturing efficiency. In recent years, such efforts have been driven by the transition from batch to continuous manufacturing and digitalization in process development. To facilitate this transition, integrated data management and informatics tools need to be developed and implemented within the framework of Industry 4.0 technology. Here, in this regard, the work aims to guide the data integration development of continuous pharmaceutical manufacturing processes under the Industry 4.0 framework, improving digital maturity and enabling the development of digital twins. This paper demonstrates two instances where a data integration framework has been successfully employed in academic continuous pharmaceutical manufacturing pilot plants. Details of the integration structure and information flows are comprehensively showcased. Approaches to mitigate concerns in incorporating complex data streams, including integrating multiple process analytical technology tools and legacy equipment, connecting cloud data and simulation models, and safeguarding cyber-physical security, are discussed. Critical challenges and opportunities for practical considerations are highlighted.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2021

Ionic liquid-water mixtures enhance pretreatment and anaerobic digestion of agave bagasse

Agave bagasse (AB), a byproduct of tequila production, is a waste generated in large quantities and an emerging alternative for biofuels production. Nevertheless, the use of lignocellulosic biomass for this purpose requires a pretreatment that maximizes its degradation and utilization. Although there are several pretreatment types, using ionic liquids (IL) has become attractive due to its ability to dissolve the lignocellulosic biomass under mild conditions of time and temperature. As a drawback are the high cost of IL's and their high viscosity; thus mixing them with water is a way to overcome these hindrances. This work aimed to evaluate the effect of pretreatment with three diluted ILs on the A. tequilana bagasse structure and sugars production, as well as the potential of the enzymatic hydrolysate to produce methane in batch mode. The evaluated ILs were: 1-ethyl-3-methylimidazolium acetate [Emim][OAc], choline lysinate [Ch][Lys] and ethanolamine acetate [EOA][OAc]. The results showed that the IL with the best performance was [Ch][Lys], which not only offered the highest yields of sugar production (0.57 ± 0.03 g total sugars / g bagasse) but also, it is possible to use it at 30 % in a mix with water obtaining similar yields as with pure IL. Enzymatic hydrolysis of IL pretreated AB achieved 50.7 kg sugars and 49.3 kg of sugars for 90 %-[Ch][Lys] and 30 %-[Ch][Lys], respectively per 100 kg of untreated biomass. The enzymatic hydrolysate from the 30 %-[Ch][Lys] pretreated AB was able to achieve 0.28 L CH4/g CODfed, demonstrating that both sugar and methane yields are maintained when pure or 30 %-[Ch][Lys] are used for pretreatment, representing a step forward towards process feasibility.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2020

The effect of moisture on hydrocarbon-based solvent liquefaction of pine, cellulose and lignin

In this report we examined the effect of moisture on the solvent liquefaction of loblolly pine, cellulose, and lignin in tetralin at 280 °C and pressures ranging from 15 to 70 bar. Solvent liquefaction experiments were conducted in a quasi-batch reactor capable of continuous pressure control and external vapor condensation. Moisture was varied by the controlled addition of de-ionized water. Control over the system pressure subsequently impacted the removal of water vapor from the reactor. Liquid yield decreased by 25, 21, and 35 wt%, for pine, cellulose, and lignin, respectively, when moisture content was increased from 1 to 50 wt% at 42 bar. Humins were observed in the solid residue from liquefaction of wet cellulose. Wet lignin yielded a substantial amount of solid residue compared to dry lignin with a corresponding decrease in total phenolic monomer production. It was concluded that the ionic dissociation of water was an important factor in loss of liquid yield in the presence of water. Although water was less than 20 wt% of the solvent loading in these experiments, it strongly influenced the liquefaction of biomass and biomass components.

09 BIOMASS FUELS↗
DOE OSTI2026

scPlantAnnotate: an accurate and robust transformer-based model for plant cell type annotation

Accurate cell type annotation remains a major bottleneck in plant single-cell RNA sequencing (scRNA-seq), where existing tools are often adapted from animal studies and perform sub-optimally on plant data. The lack of plant-specific computational frameworks limits the construction of plant cell atlases and downstream biological discovery. We develop and evaluate scPlantAnnotate, a Transformer-based reference annotation framework tailored for plant scRNA-seq data, and benchmark it against state-of-the-art deep learning and conventional methods across multiple plant species. Species-specific scPlantAnnotate models were trained using curated datasets from Arabidopsis thaliana, Zea mays, Oryza sativa, and Glycine max. We compared scPlantAnnotate with leading baselines under both standard random-split evaluation and a more stringent leave-one-dataset-out setting, which tests robustness to completely unseen datasets and tissue types. scPlantAnnotate consistently outperforms existing approaches across all four species under random-split evaluation. In the leave-one-dataset-out setting for A. thaliana, where performance drops markedly for all methods due to strong batch effects and dataset heterogeneity, scPlantAnnotate nonetheless achieves the highest Accuracy, Macro-F1, Balanced Accuracy, and Macro-AUROC on average and ranks first on most held-out datasets. These results demonstrate improved robustness to dataset shifts, a critical yet underexplored challenge in plant scRNA-seq analysis. A freely accessible web server enables users to annotate their own datasets using pretrained models. scPlantAnnotate provides a plant-specific, Transformer-based framework for single-cell annotation that delivers state-of-the-art performance and enhanced robustness to unseen datasets. By addressing limitations of existing tools and enabling scalable reference-based annotation, scPlantAnnotate supports the development of comprehensive plant cell atlases and facilitates broader use of single-cell genomics in plant biology.

Bioinformatics↗
DOE OSTI2026

Understanding the effect of organic coatings on 4-nitrostyrene hydrogenation selectivity over Pt/TiO 2 catalysts

The hydrogenation of 4-nitrostyrene was investigated over Pt/TiO 2 catalysts to understand the effects of different catalyst coatings on product selectivity and reaction rates. This reaction is a much-investigated selectivity probe due to difficulty in controlling reactivity between the nitro and vinyl groups of 4-nitrostyrene, with C=C reduction generally being more favorable. 4-Nitrostyrene hydrogenation reactions were carried out in a batch reactor using uncoated, organothiol coated, phosphonic acid (PA) coated, and ion-conducting polymer (ionomer) coated catalysts. Thiols form dense coatings on Pt surfaces, whereas PAs preferentially accumulate on TiO 2 ; ionomer binding preference is unknown for this type of catalyst. The coating densities of the thiols and PAs were varied to observe the effects of loading on catalytic activity and reaction selectivity relative to the uncoated catalyst. Changes in selectivity were observed for catalysts with all coatings, with thiol coatings shifting selectivity towards the nitro reduction product, 4-vinylaniline (4-VA), and both ionomer and PA coatings shifting towards 4-ethylnitrobenzene, the olefin hydrogenation product. It was observed that selectivity toward 4-VA increased with thiol coating density, where 1,2-benzenedithiol (BDT) coated catalysts exhibited a selectivity of >99%. These changes were attributed to the identity of the adsorption sites blocked by the different coatings: blocking of metal sites via thiols decreased the rate of C=C hydrogenation, whereas blocking of TiO 2 sites by PAs decreased the rate of nitro group hydrogenation. The changes observed for ionomer and PA coatings suggested that ionomers, like PAs, bind preferentially to the metal oxide support rather than the active metal. This work provides new support to the proposal that nitro group hydrogenation preferentially occurs on TiO 2 sites after hydrogen spillover and underscores how selection of ligand functionality can be used to tune catalyst selectivity.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2021

Unseeded, spontaneous nucleation of spherulitic magnesium calcite

Most of the sedimentary carbonates deposited in the marine environments are composed of calcium carbonate minerals with varying amounts of incorporated Mg 2+ . However, understanding how interactions of impurities with carbonate and their incorporation affect sediments behavior remains a challenge. Here, a new insight is obtained by monitoring solution composition, morphology, and electrokinetic potential of carbonate particles formed in a spontaneous unseeded batch precipitation experiment using electrochemical and scanning electron microscopy methods. The solid composition and growth rate are extracted from changes in the bulk composition and fitted to chemical affinity rate law, revealing that the precipitation pathway consists of second-order dissolution and first-order precipitation. The molecular dynamics simulations show that the lattice strain induced by randomly substituting Ca 2+ by Mg 2+ stabilizes spherical nanoparticles and reduces their surface area and volume. Combining kinetics and thermodynamics insight, we conclude that variation in the carbonate bulk and interfacial energies, along with the solution supersaturation, lead to the dissolution-precipitation transformation pathway from Mg-rich to Mg-poor carbonate phase that preserves spherulitic morphology. Furthermore, our findings are relevant for long-standing questions of how impurities influence diagenesis of carbonate sediments and spherulitic carbonate particles' origin.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2026

Chemical-Free Lithium Separation from High-Salinity Brines Using Model-Informed and Machine Learning-Optimized Multi-Column Zwitterionic Chromatography

Direct Lithium Extraction (DLE) technologies often struggle to produce high-purity lithium salts from high-salinity brines, as current approaches require chemical-based elution, regeneration, and precipitation steps, resulting in significant environmental footprints. A novel salt fractionation approach using carboxybetaine resin, known as zwitterionic chromatography (ZIC), has demonstrated that lithium ions can be separated from divalent cations under high-salinity conditions using only water as eluent, with no regeneration required. To enable continuous and scalable deployment of this approach, we developed a chemical-free Multi-column Zwitterionic Chromatography (MZC) process and its theoretical and process models. To predict and optimize this nontraditional separation system, we introduced a novel anti-Langmuir isotherm, and the isotherm parameters were estimated through a machine learning-driven optimization based on artificial neural network ensembles with numerical feasibility assessment. Using machine learning-driven optimization, the MZC process achieved 98.0% lithium recovery, 99.5 % Li/(Li + Mg + Ca) purity, a 31.3% productivity increase, and a 33% reduction in water use compared to batch operation. The proposed MZC process enables lithium separation at $0.6-1.2 kg-1 Li, with costs dominated by resin manufacturing, while offering lower separation costs and carbon footprint compared with conventional carbonation. Overall, these findings position the MZC process as an effective polishing step within scalable and sustainable lithium production pipelines.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2021

A physics-informed and hierarchically regularized data-driven model for predicting fluid flow through porous media

This paper presents a new deep learning data-driven model for predicting structure dependent pore-fluid velocity fields in rock. The model is based on a Convolutional Auto-Encoder (CAE) artificial neural network capable of learning from image data generated by direct numerical simulations of fluid flow through pore-structures, such as by Lattice Boltzmann or molecular dynamics methods. The main novelty of the model in comparison to previous CAE-based data-driven approaches consists of three parts. The first is a methodology for decomposing the full-domain of the porous media into sub-regions, or “sub-domains”, in order to reduce the overall size of the CAE, batch process the sub-domains in parallel, and enable the CAE to learn local and generalizable nonlinear mappings of pore-fluid velocities. The second consists of embedding the finite difference solutions of the incompressible Navier-Stokes and continuity equations into convolutional layers prior to the CAE in order to provide the CAE with knowledge of fluid dynamics physics (PhyFlow). The third main novelty is that the training of the CAE is regularized with a hierarchical loss function that encourages the learning of fluid flow patterns (in a way similar to ranked modes in principal component analysis), ranking from most to least important. This is shown to increase the stability in learning, reduce over-fitting, and promote interpretability of the CAE neural network layers (HierCAE). The comprehensive new data-driven model, which we call the PhyFlow-HierCAE model, is shown to exhibit improved accuracy and generalizability of flow field predictions over conventional CAE models, attributable to the embedded physical knowledge and the hierarchical regularization, as well as realize orders of magnitude speed-ups in computation times as a surrogate for the direct numerical simulations. Examples of training and forward predictions on unseen pore-structures are provided and evaluated for data from Lattice Boltzmann and molecular dynamics simulations of pore-fluid flow. The model is shown to be a fast and accurate emulator (or “surrogate”) for predicting effective permeability of unseen pore-structures based on learning from relatively small direct numerical simulation datasets.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
DOE OSTI2022

Learning functional priors and posteriors from data and physics

In this work, we develop a new Bayesian framework based on deep neural networks to be able to extrapolate in space-time using historical data and to quantify uncertainties arising from both noisy and gappy data in physical problems. Specifically, the proposed approach has two stages: (1) prior learning and (2) posterior estimation. At the first stage, we employ the physics-informed Generative Adversarial Networks (PI-GAN) to learn a functional prior either from a prescribed function distribution, e.g., Gaussian process, or from historical data and physics. At the second stage, we employ the Hamiltonian Monte Carlo (HMC) method to estimate the posterior in the latent space of PI-GANs. In addition, we use two different approaches to encode the physics: (1) automatic differentiation, used in the physicsinformed neural networks (PINNs) for scenarios with explicitly known partial differential equations (PDEs), and (2) operator regression using the deep operator network (DeepONet) for PDE-agnostic scenarios. We then test the proposed method for (1) meta-learning for one-dimensional regression, and forward/inverse PDE problems (combined with PINNs); (2) PDE-agnostic physical problems (combined with DeepONet), e.g., fractional diffusion as well as saturated stochastic (100-dimensional) flows in heterogeneous porous media; and (3) spatial-temporal regression problems, i.e., inference of a marine riser displacement field using experimental data from the Norwegian Deepwater Programme (NDP). The results demonstrate that the proposed approach can provide accurate predictions as well as uncertainty quantification given very limited scattered and noisy data, since historical data could be available to provide informative priors. In summary, the proposed method is capable of learning flexible functional priors, e.g., both Gaussian and non-Gaussian process, and can be readily extended to big data problems by enabling mini-batch training using stochastic HMC or normalizing flows since the latent space is generally characterized as low dimensional.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
DOE OSTI2022

Low temperature upcycling of polyethylene to gasoline range chemicals: Hydrogen transfer and heat compensation to endothermic pyrolysis reaction over zeolites

Selective production of gasoline ranged chemicals were of great interest in the field of plastic upcycling. Here this work reports exothermic hydrogen transfer reaction which was observed when HY zeolite was adopted to catalytic pyrolyze the polyethylene to gasoline ranged chemicals, while it was not observed when H-ZSM5 was used as catalyst. Both zeolite catalysts could significantly bring down the reaction temperature from 500 to 300–350 °C. TG-FITR analysis revealed that products obtained with HY zeolites were dominated with saturated hydrocarbons as compared to H-ZSM5 where the combination of alkanes, alkenes, and aromatics was obtained. In addition to the formation of saturated compounds, the TG-DSC analysis confirmed that, over the HY zeolites, the endothermic hydrogen transfer reaction of olefins occurred along with the pyrolysis of LDPE visibly when the zeolite ratio in the zeolite/polyethylene mixture exceeded 50% by mass. The TG results also indicated that the polyethylene undergone nearly 100% converted with very few residues. Further product analysis from batch pyrolysis by GC-MS confirmed that gasoline ranged iso-alkanes were produced over HY zeolite. This study highlighted that coupling exothermic hydrogen transfer reactions with endothermic pyrolysis could be an energy-efficient way of producing gasoline ranged chemicals. Chemical upcycling of plastics is tunable by using different zeolite combinations for the formation of desired products.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗