Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “batch sampling”

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 37 records · Page 2

DEFENSE WASTE PROCESSING FACILITY ANALYTICAL METHOD VERIFICATION FOR THE SLUDGE BATCH 5 QUALIFICATION SAMPLE

For each sludge batch that is processed in the Defense Waste Processing Facility (DWPF), the Savannah River National Laboratory (SRNL) performs confirmation of the applicability of the digestion method to be used by the DWPF lab for elemental analysis of Sludge Receipt and Adjustment Tank (SRAT) receipt samples and SRAT product process control samples. DWPF SRAT samples are typically dissolved using a room temperature HF-HNO3 acid dissolution (i.e., DWPF Cold Chem Method, see Procedure SW4-15.201) and then analyzed by inductively coupled plasma - atomic emission spectroscopy (ICP-AES). This report contains the results and comparison of data generated from performing the Aqua Regia (AR), Sodium Peroxide/Hydroxide Fusion (PF) and DWPF Cold Chem (CC) method digestion of Sludge Batch 5 (SB5) SRAT Receipt and SB5 SRAT Product samples. The SB5 SRAT Receipt and SB5 SRAT Product samples were prepared in the SRNL Shielded Cells, and the SRAT Receipt material is representative of the sludge that constitutes the SB5 Batch composition. This is the sludge in Tank 51 that is to be transferred into Tank 40, which will contain the heel of Sludge Batch 4 (SB4), to form the SB5 Blend composition. The results for any one particular element should not be used in any way to identify the form or speciation of a particular element in the sludge or used to estimate ratios of compounds in the sludge. A statistical comparison of the data validates the use of the DWPF CC method for SB5 Batch composition. However, the difficulty that was encountered in using the CC method for SB4 brings into question the adequacy of CC for the SB5 Blend. Also, it should be noted that visible solids remained in the final diluted solutions of all samples digested by this method at SRNL (8 samples total), which is typical for the DWPF CC method but not seen in the other methods. Recommendations to the DWPF for application to SB5 based on studies to date: (1) A dissolution study should be performed on the WAPS sample by SRNL which consists of the final composition of the sludge (the SB5 Blend); (2) Given the heel of SB4 in Tank 40, the DWPF lab should monitor the aluminum concentration in the first 10 SRAT Receipt batches of SB5 using both CC and sodium peroxide/hydroxide fusion to evaluate the adequacy of aluminum recovery by the CC method for this sludge batch; and (3) SRNL and the DWPF lab should investigate if comparisons between the elemental concentrations of the SME product glass (adjusted for frit addition) obtained by the mixed acid and peroxide fusion digestion and the SRAT Receipt and SRAT Product elemental concentrations obtained via the DWPF CC method provide insight into the adequacy of the CC method for analysis of the SRAT Product. The DWPF lab would need to calcine the SRAT product at 1050 C for the best comparison. If a consistent difference in elemental concentrations is revealed, another type of digestion (i.e. sodium peroxide/hydroxide fusion) should be used to determine the concentration of the element in question. Particular emphasis should be placed on monitoring the aluminum concentration in SB5.

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗

VERIFICATION OF THE DEFENSE WASTE PROCESSING FACILITY'S (DWPF) PROCESS DIGESTION METHOD FOR THE SLUDGE BATCH 7A QUALIFICATION SAMPLE

For each sludge batch that is processed in the Defense Waste Processing Facility (DWPF), the Savannah River National Laboratory (SRNL) performs confirmation of the applicability of the digestion method to be used by the DWPF lab for elemental analysis of Sludge Receipt and Adjustment Tank (SRAT) receipt samples and SRAT product process control samples. DWPF SRAT samples are typically dissolved using a room temperature HF-HNO{sub 3} acid dissolution (i.e., DWPF Cold Chem Method, see DWPF Procedure SW4-15.201) and then analyzed by inductively coupled plasma - atomic emission spectroscopy (ICP-AES). This report contains the results and comparison of data generated from performing the Aqua Regia (AR), Sodium peroxide/Hydroxide Fusion (PF) and DWPF Cold Chem (CC) method digestions of Sludge Batch 7a (SB7a) SRAT Receipt and SB7a SRAT Product samples. The SB7a SRAT Receipt and SB7a SRAT Product samples were prepared in the SRNL Shielded Cells, and the SRAT Receipt material is representative of the sludge that constituates the SB7a Batch or qualification composition. This is the sludge in Tank 51 that is to be transferred into Tank 40, which will contain the heel of Sludge Batch 6 (SB6), to form the Sb7a Blend composition.

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗

VERIFICATION OF THE DEFENSE WASTE PROCESSING FACILITY PROCESS DIGESTION METHOD FOR THE SLUDGE BATCH 6 QUALIFICATION SAMPLE

For each sludge batch that is processed in the Defense Waste Processing Facility (DWPF), the Savannah River National Laboratory (SRNL) confirms applicability of the digestion method to be used by the DWPF lab for elemental analysis of Sludge Receipt and Adjustment Tank (SRAT) receipt samples and SRAT product process control samples.1 DWPF SRAT samples are typically dissolved using a room temperature HF-HNO3 acid dissolution (i.e., DWPF Cold Chem (CC) Method, see DWPF Procedure SW4-15.201) and then analyzed by inductively coupled plasma - atomic emission spectroscopy (ICPAES). In addition to the CC method confirmation, the DWPF lab's mercury (Hg) digestion method was also evaluated for applicability to SB6 (see DWPF procedure 'Mercury System Operating Manual', Manual: SW4-15.204. Section 6.1, Revision 5, Effective date: 12-04-03). This report contains the results and comparison of data generated from performing the Aqua Regia (AR), Sodium Peroxide/Hydroxide Fusion (PF) and DWPF Cold Chem (CC) method digestion of Sludge Batch 6 (SB6) SRAT Receipt and SB6 SRAT Product samples. For validation of the DWPF lab's Hg method, only SRAT receipt material was used and compared to AR digestion results. The SB6 SRAT Receipt and SB6 SRAT Product samples were prepared in the SRNL Shielded Cells, and the SRAT Receipt material is representative of the sludge that constitutes the SB6 Batch or qualification composition. This is the sludge in Tank 51 that is to be transferred into Tank 40, which will contain the heel of Sludge Batch 5 (SB5), to form the SB6 Blend composition. In addition to the 16 elements currently measured by the DWPF, this report includes Hg and thorium (Th) data (Th comprising {approx}2.5 - 3 Wt% of the total solids in SRAT Receipt and SRAT Product, respectively) and provides specific details of ICP-AES analysis of Th. Thorium was found to interfere with the U 367.007 nm emission line, and an inter-element correction (IEC) had to be applied to U data, which is also discussed. The results for any one particular element should not be used in any way to identify the form or speciation of a particular element without support from XRD analysis or used to estimate ratios of compounds in the sludge.

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗

Investigation of a Wedge Adhesion Test for Edge Seals

Many photovoltaic (PV) technologies have been found to be sensitive to moisture that diffuses into a PV package. Even with the use of impermeable frontsheets and backsheets, moisture can penetrate from the edges of a module. To limit this moisture ingress pathway from occurring, manufacturers often use a low permeability polyisobutylene (PIB) based edge seal filled with desiccant to further restrict moisture ingress. Moisture ingress studies have shown that these materials are capable of blocking moisture for the 25-year life of a module; but to do so, they must remain well-adhered and free of cracks. This work focuses on adapting the Boeing Wedge test for use with edge seals laminated using glass substrates as part of a strategy to assess the long-term durability of edge seals. The advantage of this method is that it duplicates the residual stresses and strains that a glass/glass module may have when the lamination process results in some residual glass bending that puts the perimeter in tension. Additionally, this method allows one to simultaneously expose the material to thermal stress, humidity, mechanical stress, and ultraviolet radiation. The disadvantage of this method generally is that we are limited by the fracture toughness of the glass substrates that the edge seal is adhered to. However, the low toughness of typical uncrosslinked or sparsely crosslinked PIB makes them suitable for this technique. We present data obtained during the development of the wedge test for use with PV edge seal materials. This includes development of the measuring techniques and evaluation of the test method with relevant materials. We find consistent data within a given experiment, along with the theoretical independence of fracture toughness measurements with wedge thickness. This indicates that the test methodology is reproducible. However, even though individual experimental sets are consistent, the reproducibility between experimental sets is poor. We believe this may be due to inconsistencies in sample history, sample batch, or small changes in sample preparation/assembly from one month to the next. Because the fracture strength of typical edge seal materials is so low, they cannot be relied upon for mechanical strength. A small stress or strain on the edge seal is capable of promoting delamination or tearing causing the edge seal to fail. Because of this, edge seals are very dependent on the processing and construction parameters in the full size PV module such that any long term evaluation of their durability must be conducted on full size modules to be accurate.

14 SOLAR ENERGY↗

Accelerated Dimension-Independent Adaptive Metropolis

This work describes improvements from algorithmic and architectural means to black-box Bayesian inference over high-dimensional parameter spaces. The well-known adaptive Metropolis (AM) algorithm [33] is extended herein to scale asymptotically uniformly with respect to the underlying parameter dimension for Gaussian targets, by respecting the variance of the target. The resulting algorithm, referred to as the dimension-independent adaptive Metropolis (DIAM) algorithm, also shows improved performance with respect to adaptive Metropolis on non-Gaussian targets. This algorithm is further improved, and the possibility of probing high-dimensional (with dimension d 1000) targets is enabled, via GPU-accelerated numerical libraries and periodically synchronized concurrent chains (justi ed a posteriori). Asymptotically in dimension, this GPU implementation exhibits a factor of four improvement versus a competitive CPU-based Intel MKL parallel version alone. Strong scaling to concurrent chains is exhibited, through a combination of longer time per sample batch (weak scaling) and yet fewer necessary samples to convergence. The algorithm performance is illustrated on several Gaussian and non-Gaussian target examples, in which the dimension may be in excess of one thousand.

97 MATHEMATICS AND COMPUTING↗

Enhanced H 2 production from subsurface through organic acids stimulation Report

Geologic hydrogen (H 2 ) generation through serpentinization of ultramafic rocks presents a promising alternative to conventional hydrogen production, utilizing Earth's abundant magnesium- and iron-rich minerals. This study explores the use of short-chain organic acids (specifically malic acid, MA; and citric acid, CA) to enhance hydrogen yields by stimulating mineral dissolution and iron oxidation from ultramafic rock samples. Batch experiments were conducted with crushed peridotite samples from the Tamarack Complex (Minnesota) and Del Puerto Ophiolite (California). Experiments employed 0.05 M and 0.1 M solutions of MA or CA, under 500 psi (35 bar) nitrogen pressure, at temperatures up to 150 °C. Real-time gas monitoring, aqueous chemistry and solid-phase analyses were performed to track reaction progress and hydrogen production.

08 HYDROGEN↗

Enhanced H 2 production from subsurface through organic acids stimulation

Geologic hydrogen (H₂) generation through serpentinization of ultramafic rocks presents a promising alternative to conventional hydrogen production, utilizing Earth's abundant magnesium- and iron-rich minerals. This study explores the use of short-chain organic acids (specifically malic acid, MA; and citric acid, CA) to enhance hydrogen yields by stimulating mineral dissolution and iron oxidation from ultramafic rock samples. Batch experiments were conducted with crushed peridotite samples from the Tamarack Complex (Minnesota) and Del Puerto Ophiolite (California). Experiments employed 0.05 M and 0.1 M solutions of MA or CA, under 500 psi (35 bar) nitrogen pressure, at temperatures up to 150 °C. Real-time gas monitoring, aqueous chemistry and solid-phase analyses were performed to track reaction progress and hydrogen production.

08 HYDROGEN↗

Clean quartz matters for cosmogenic nuclide analyses: An exploration of the importance of sample purity using the CRONUS-N reference material

We report reference materials are key for assessing inter-laboratory variability and measurement quality, and for placing analytical uncertainty bounds on sample analyses. Here, we investigate four years of data resulting from repeated processing of the CRONUS-N reference material for cosmogenic 10 Be and 26 Al analyses. At University of Vermont, we prepared a CRONUS-N aliquot with most of our sample batches from 2013 to 2017; these reference material samples were then distributed to four different accelerator mass spectrometry facilities, yielding 73 10 Be analyses and 58 26 Al analyses. We determine CRONUS-N 10 Be concentrations of (2.26 ± 0.14) x 10 5 atoms g -1 (n = 73, mean, 1 SD) and 26 Al concentrations of (1.00 ± 0.08) x 10 6 atoms g -1 (n = 58, mean, 1 SD). We find a reproducibility of 6.3% for 10 Be and 7.7% for 26 Al (relative standard deviations). We also document highly variable 27 Al and Mg concentrations and a 10 Be dispersion twice as large as the mean AMS analytic uncertainty. Analyses of the CRONUS-N material with and without density separation demonstrate that non-quartz minerals are present in the material and have a large impact on measured concentrations of 27 Al, 10 Be, and impurities; these non-quartz minerals represent only a very small portion of the total mass (0.6–0.8%) but have a disproportionally large effect on the resulting data. Our results highlight the importance of completely removing all non-quartz mineral phases from samples prior to Be/Al extraction for the determination of in situ cosmogenic 10 Be and 26 Al concentrations.

58 GEOSCIENCES↗

Analyses and Comparison of Bulk and Coil Surface Samples from the DWPF Slurry Mix Evaporator

Sludge samples from the DWPF Slurry Mix Evaporator (SME) heating coil frame and coil surface were characterized to identify differences that might help identify heat transfer fouling materials. The SME steam coils have seen increased fouling leading to lower boil-up rates. Samples of the sludge were taken from the coil frame somewhat distant from the coil (bulk tank material) and from the coil surface (coil surface sample). The results of the analysis indicate the composition of the two SME samples are very similar with the exception that the coil surface sample shows ~5-10X higher mercury concentration than the bulk tank sample. Elemental analyses and x-ray diffraction results did not indicate notable differences between the two samples. The ICP-MS and Cs-137 data indicate no significant differences in the radionuclide composition of the two SME samples. Semi-volatile organic analysis revealed numerous organic molecules, these likely result from antifoaming additives. The compositions of the two SME samples also match well with the analyzed composition of the SME batch with the exception of significantly higher silicon, lithium, and boron content in the batch sample indicating the coil samples are deficient in frit relative to the SME batch composition.

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗

Summary of Results from Batch 1 Qualification Samples for Tank Closure Cesium Removal 1A (TCCR 1A)

Savannah River Remediation (SRR) is currently operating the Tank Closure Cesium Removal 1A (TCCR 1A) process to remove 137 Cs from tank waste supernate using an ion exchange process. As part of that process, Savannah River National Laboratory (SRNL) receives and analyzes samples in support of the qualification of each waste supernate batch to be processed. SRNL recently received supernate samples retrieved from Tank 10H as well as in-tank batch contact samples for characterization in support of qualifying Batch 1 for processing through the TCCR 1A unit.

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗

Conversion kinetics of container glass batch melting

Abstract Understanding the batch‐to‐glass conversion process is fundamental to optimizing the performance of glass‐melting furnaces and ensuring that furnace modeling can correctly predict the observed outcome when batch materials or furnace conditions change. To investigate the kinetics of silica dissolution, gas evolution, and primary foam formation and collapse, we performed X‐ray diffraction, thermal gravimetry, feed expansion tests, and evolved gas analysis of batch samples heated at several constant heating rates. We found that gas evolving reactions, foaming, and silica dissolution depend on the thermal history of the batch in a similar manner: the kinetic parameters of each process were linear functions of the square root of the heating rate. This kinetic similarity reflects the stronger‐than‐expected interdependence of these processes. On the basis of our results, we suggest that changes in furnace operating conditions, such as firing or boosting, influence the melting rate less than what one would expect without consideration of batch conversion kinetics.

Ueda, Nanako↗

Investigating Using Oscillating Jets for Fluid Mixing

Scaled mixing experiments were conducted to evaluate maintaining mobilized particles in a uniform suspension (the condition of concentration uniformity) using jet pumps to mix the suspension. The study experimentally evaluated uniformity in a 1/12-scale experiment varying the Reynolds number, Froude number, and gravitational settling parameter space. The initial matrix specified only tests at 100% U0D0 and 25% U0D0. After initial tests were conducted with small diameter, low viscosity simulant this matrix was revised to allow evaluation of a broader range of U0D0. The revised matrix included a full factorial test between 100% and 50% U0D0 and two half-factorial tests at 75% and 25% U0D0. Adding points at 75% U0D0 and 50% U0D0 allowed evaluation of curvature. Eliminating points at 25% U0D0 decreased the testing time by several weeks. Test conditions were achieved by varying the simulant viscosity, the mean particle size (dp), and the jet nozzle exit velocity (U0). Concentration measurements at sampling locations throughout the tank were used to assess the degree of uniformity achieved during each test. Concentration data was obtained using a real time ultrasonic attenuation probe and discrete batch samples. The undissolved solids concentration at these locations was analyzed to determine whether the tank contents were uniform (< ±10% variation about mean) or nonuniform (> ±10% variation about mean) in concentration. Concentration inhomogeneity was modeled as a function of dimensionless groups.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Results of Hg Speciation Testing on DWPF Batch 735 RCT and OGCT Samples

Eurofins Frontier Global Sciences, Inc. in Seattle, WA on behalf of the Savannah River Remediation (SRR) Mercury Task Team tasked the Savannah River National Laboratory (SRNL) with preparing and shipping samples for Hg speciation. The fifth shipment of samples was designated to include Defense Waste Processing Facility (DWPF) samples from Batch 735 including two Recycle Collection Tank (RCT) samples and one Off Gas Condensate Tank (OGCT) sample; both are described fully in Table 1. This work was requested by DWPF in a Technical Assistance Request (TAR).

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗

Experimental Evaluation of Dual-Opposed Jet Mixer Pump Performance for Slurry Mixing

Million-gallon double-shell tanks at Hanford are used to store transuranic, high-level, and low-level radioactive wastes. These wastes consist of a large volume of salt-laden solution covering a smaller volume of settled sludge primarily containing metal hydroxides. These wastes will be retrieved and processed into immobile waste forms suitable for permanent disposal. Retrieval is an important step in implementing these disposal scenarios. The retrieval concept evaluated is to use submerged dual-nozzle jet mixer pumps with horizontally oriented nozzles located near the tank floor that produce horizontal jets of fluid to mobilize the settled solids. The mixer pumps are oscillated through 180 about a vertical axis so the high velocity fluid jets sweep across the floor of the tank. After the solids are mobilized, the pumps will continue to operate at a reduced flow rate producing lower velocity jets sufficient to maintain the particles in a uniform suspension (concentration uniformity). Several types of waste and tank configurations exist at Hanford. The jet mixer pump systems and operating conditions required to mobilize sludge and maintain slurry uniformity will be a function of the waste type and tank configuration. The focus of this work was to conduct a 1/12-scale experiment to develop an analytical model to relate slurry uniformity to tank and mixer pump configurations, operating conditions, and sludge properties. This experimental study evaluated concentration uniformity in a 1/12-scale experiment varying the Reynolds number (Re), Froude number (Fr), and gravitational settling parameter (Gs) space. Simulant physical properties were chosen to obtain the required Re and Gs where Re and Gs were varied by adjusting the kinematic viscosity and mean particle diameter, respectively. Test conditions were achieved by scaling the jet nozzle exit velocity in a 75-in. diameter tank using a mock-up of a centrally located dual-opposed jet mixer pump located just above the tank floor. Concentration measurements at sampling locations throughout the tank were used to assess the degree of uniformity achieved during each test. Concentration data was obtained using a real time in-situ ultrasonic attenuation probe and post-test analysis of discrete batch samples. The undissolved solids concentration at these locations was analyzed to determine whether the tank contents were uniform (≤ ±10% variation about mean) or nonuniform (> ±10% variation about mean) in concentration. Concentration inhomogeneity was modeled as a function of dimensionless parameters. The parameters that best describe the maximum solids volume fraction that can be suspended were found to be 1) the Fr based on nozzle average discharge velocity and tank contents level and 2) the dimensionless particle size based on nozzle diameter. The dependence on the jet Re does not appear to be statistically significant.

correlation↗

Thermal Analysis of Waste Glass Batches: Effect of Batch Makeup on Gas-Evolving Reactions

Batches made with a variety of precursors were subjected to thermo-gravimetric analysis. The baseline modifications included all-nitrate batch with sucrose addition, all-carbonate batch, and batches with different sources of alumina. All batches were formulated for a single glass composition (a vitrified simulated high-alumina high-level waste). Batch samples were heated from the ambient temperature to 1200°C at constant heating rates ranging from 1 K/min to 50 K/min. Major gas evolving reactions began at temperatures just above 100°C and were virtually complete by 650°C. Activation energies for major reactions were obtained with the Kissinger’s method. A rough model for the overall kinetics of the batch-conversion was developed to be eventually applied to a mathematical model of the cold cap.

Pierce, David A.↗

Effects of particle size and briquetting of soda-lime-silicate glass batch on viscosity during batch-to-melt conversion

To assess the effects of grain size and briquetting on glass-batch melting, we measured the viscosity (?b) of an industrial glass batch during its conversion to molten simulated architectural glass within the temperature interval of 1050 °C to 1300 °C and at a shear rate of 0.02 s-1. On the basis of the mass fraction of undissolved sand (xs), which was determined using X-ray diffraction analysis on quenched batch samples, and the gas-phase fraction (p), which was obtained from volumetric measurements, we established the relationship ln(?b/?m) = a0 + bsxs + app, where ?m is the transition melt viscosity, a0 is the melt homogeneity coefficient, bs is the coefficient for undissolved sand, and ap is the coefficient for bubbles. The obtained values, bs = 8.756 and ap = 0.658, indicate that undissolved silica sand strongly influences batch viscosity. Consequently, compared with coarse-grained batches, fine raw materials and briquetting resulted in lower batch viscosity during conversion. Owing to their lower viscosities and higher heat conductivities, batches from finer raw materials or batches shaped into briquettes are expected to enhance the melting rate of batch blankets in industrial glass-melting furnaces.

Doi, Yoji↗

Single-Molecule Fluorescence Microscopy Reveals Energy Transfer Active versus Inactive Nanocrystal/Dye Conjugate Pairs

Defect-mediated energy transfer (EnT) is a radiative process that occurs between donor defect states in the forbidden bandgap of semiconductor nanocrystals (NCs) and dye molecules bound to their surfaces. The EnT efficiency depends on the number of dye molecules attached to each NC, the donor–acceptor distance, and the dipole orientation factor between the donor and acceptor, all of which vary across all individual NCs in a sample. While ensemble-level fluorescence spectroscopy measurements have provided average values for donor–acceptor distances, dye-to-NC ratios, and EnT rate constants, questions remain about the impact of donor/acceptor heterogeneity on observed EnT efficiencies. Notably, ensemble-level measurements cannot distinguish between bare NCs and EnT-active versus inactive NC/dye pairs in the same sample batch, limiting the ability to design systems with 100% EnT efficiency. To address this, we studied defect-mediated EnT between AlexaFluor 555 dye acceptors chemically bound to ZnO NC donors at the level of single molecules and single NCs. Interestingly, 20% of bound NC/dye pairs are EnT-inactive, likely contributing to residual defect photoluminescence (PL) observed in ensemble-level measurements and reducing overall EnT efficiency. Single particle-level ZnO defect PL and acceptor fluorescence trajectories exhibited distinct microfluctuations, which are absent in bare ZnO NCs. We hypothesized that our observations can be explained with a competitive dye fluorescence quenching pathway, possibly due to charge transfer between the excited state dye and the ZnO NC. Numerical simulations of single-molecule PL traces for this scenario produced microfluctuations consistent with the experimental results. These findings highlight the impact of sample heterogeneity on EnT processes and provide insights for designing light-harvesting systems with optimized EnT efficiency.

25 ENERGY STORAGE↗

Stochastic projective splitting

Here, we present a new, stochastic variant of the projective splitting (PS) family of algorithms for inclusion problems involving the sum of any finite number of maximal monotone operators. This new variant uses a stochastic oracle to evaluate one of the operators, which is assumed to be Lipschitz continuous, and (deterministic) resolvents to process the remaining operators. Our proposal is the first version of PS with such stochastic capabilities. We envision the primary application being machine learning (ML) problems, with the method’s stochastic features facilitating “mini-batch” sampling of datasets. Since it uses a monotone operator formulation, the method can handle not only Lipschitz-smooth loss minimization, but also min–max and noncooperative game formulations, with better convergence properties than the gradient descent-ascent methods commonly applied in such settings. The proposed method can handle any number of constraints and nonsmooth regularizers via projection and proximal operators. We prove almost-sure convergence of the iterates to a solution and a convergence rate result for the expected residual, and close with numerical experiments on a distributionally robust sparse logistic regression problem.

97 MATHEMATICS AND COMPUTING↗