Search NASA⌕ Search

SEARCH · Search NASA

Results for “Stochastic optimization”

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 1,315 records · Page 73

Preserving nonlinear constraints in variational flow filtering data assimilation

Data assimilation aims to estimate the states of a dynamical system by optimally combining sparse and noisy observations of the physical system with uncertain forecasts produced by a computational model. The states of many dynamical systems of interest obey nonlinear physical constraints, and the corresponding dynamics is confined to a certain sub-manifold of the state space. Standard data assimilation techniques applied to such systems yield posterior states lying outside the manifold, violating the physical constraints. This work focuses on particle flow filters which use stochastic differential equations to evolve state samples from a prior distribution to samples from an observation-informed posterior distribution. The variational Fokker-Planck (VFP)—a generic particle flow filtering framework—is extended to incorporate non-linear, equality state constraints in the analysis. To this end, two algorithmic approaches that modify the VFP stochastic differential equation are discussed: (i) VFPSTAB, to inexactly preserve constraints with the addition of a stabilizing drift term, and (ii) VFPDAE, to exactly preserve constraints by treating the VFP dynamics as a stochastic differential-algebraic equation (SDAE). Additionally, an implicit-explicit time integrator is developed to evolve the VFPDAE dynamics. The strength of the proposed approach for constraint preservation in data assimilation is demonstrated on three test problems: the double pendulum, Korteweg-de-Vries, and the incompressible Navier-Stokes equations.

97 MATHEMATICS AND COMPUTING↗

A data-driven sensor placement approach for detecting voltage violations in distribution systems

Stochastic fluctuations in power injections from distributed energy resources (DERs) combined with load variability can cause constraint violations (e.g., exceeded voltage limits) in electric distribution systems. To monitor grid operations, sensors are placed to measure important quantities such as the voltage magnitudes. Here, in this paper, we consider a sensor placement problem which seeks to identify locations for installing sensors that can capture all possible violations of voltage magnitude limits. We formulate a bilevel optimization problem that minimizes the number of sensors and avoids false sensor alarms in the upper level while ensuring detection of any voltage violations in the lower level. This problem is challenging due to the nonlinearity of the power flow equations and the presence of binary variables. Accordingly, we employ recently developed conservative linear approximations of the power flow equations that overestimate or underestimate the voltage magnitudes. By replacing the nonlinear power flow equations with conservative linear approximations, we can ensure that the resulting sensor locations and thresholds are sufficient to identify any constraint violations. Additionally, we apply various problem reformulations to significantly improve computational tractability while simultaneously ensuring an appropriate placement of sensors. Lastly, we improve the quality of the results via an approximate gradient descent method that adjusts the sensor thresholds. We demonstrate the effectiveness of our proposed method for several test cases, including a system with multiple switching configurations.

24 POWER TRANSMISSION AND DISTRIBUTION↗

Strong-field QED limitations on TeV-class plasma wakefield accelerators

We demonstrate that quantum and classical radiation effects can become non-negligible for TeV-class beams propagating through plasma channels typical of staged plasma accelerators. Although the quantum nonlinearity parameter χ e remains small under currently envisioned experimental conditions, the cumulative influence of radiation over long acceleration distances can lead to significant modifications to the beam’s energy spread, emittance, and polarization. Our analytic models, validated by particle-in-cell simulations, highlight that for standard Gaussian beams, the orbit-induced energy spread dominates over quantum stochastic effects but can be mitigated by tailoring the beam profile, for example, through ring-shaped transverse distributions. In regimes where the radiation reaction approaches the accelerating force, the emittance may be cooled, forming distinctive ring-shaped phase-space structures. Finally, we analyze the influence of the radiation effect on spin transport inside the wakefield. These findings underscore the importance of considering both classical and quantum radiation dynamics in the design and optimization of future high-intensity plasma accelerators.

accelerator↗

Scalable deep text comprehension for Cancer surveillance on high-performance computing

Background: Deep Learning (DL) has advanced the state-of-the-art capabilities in bioinformatics applications which has resulted in trends of increasingly sophisticated and computationally demanding models trained by larger and larger data sets. This vastly increased computational demand challenges the feasibility of conducting cutting-edge research. One solution is to distribute the vast computational workload across multiple computing cluster nodes with data parallelism algorithms. In this study, we used a High-Performance Computing environment and implemented the Downpour Stochastic Gradient Descent algorithm for data parallelism to train a Convolutional Neural Network (CNN) for the natural language processing task of information extraction from a massive dataset of cancer pathology reports. We evaluated the scalability improvements using data parallelism training and the Titan supercomputer at Oak Ridge Leadership Computing Facility. To evaluate scalability, we used different numbers of worker nodes and performed a set of experiments comparing the effects of different training batch sizes and optimizer functions.Results: We found that Adadelta would consistently converge at a lower validation loss, though requiring over twice as many training epochs as the fastest converging optimizer, RMSProp. The Adam optimizer consistently achieved a close 2nd place minimum validation loss significantly faster; using a batch size of 16 and 32 allowed the network to converge in only 4.5 training epochs.Conclusions: We demonstrated that the networked training process is scalable across multiple compute nodes communicating with message passing interface while achieving higher classification accuracy compared to a traditional machine learning algorithm.

Qiu, John X.↗

Physics-based modeling and data analytics [Slides]

This presentation contains a summary of ongoing work within the physics-based modeling and data analytics work package within the Nuclear Materials Discovery and Qualification initiative (NMDQi). Topics include work on MOOSE-based crystal plasticity, molecular dynamics modeling of recombination in metals and alloys, the MOOSE Stochastic Tools Module, and machine learning and atomistic modeling to predict thermo-kinetic properties of nuclear structural materials.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗

Spectral proper orthogonal decomposition and its relationship to dynamic mode decomposition and resolvent analysis

We consider the frequency domain form of proper orthogonal decomposition (POD), called spectral proper orthogonal decomposition (SPOD). Spectral POD is derived from a space–time POD problem for statistically stationary flows and leads to modes that each oscillate at a single frequency. This form of POD goes back to the original work of Lumley ( Stochastic Tools in Turbulence , Academic Press, 1970), but has been overshadowed by a space-only form of POD since the 1990s. We clarify the relationship between these two forms of POD and show that SPOD modes represent structures that evolve coherently in space and time, while space-only POD modes in general do not. We also establish a relationship between SPOD and dynamic mode decomposition (DMD); we show that SPOD modes are in fact optimally averaged DMD modes obtained from an ensemble DMD problem for stationary flows. Accordingly, SPOD modes represent structures that are dynamic in the same sense as DMD modes but also optimally account for the statistical variability of turbulent flows. Finally, we establish a connection between SPOD and resolvent analysis. The key observation is that the resolvent-mode expansion coefficients must be regarded as statistical quantities to ensure convergent approximations of the flow statistics. When the expansion coefficients are uncorrelated, we show that SPOD and resolvent modes are identical. Our theoretical results and the overall utility of SPOD are demonstrated using two example problems: the complex Ginzburg–Landau equation and a turbulent jet.

Mechanics↗

Thermodynamics of order and randomness in dopant distributions inferred from atomically resolved imaging

Abstract Exploration of structure-property relationships as a function of dopant concentration is commonly based on mean field theories for solid solutions. However, such theories that work well for semiconductors tend to fail in materials with strong correlations, either in electronic behavior or chemical segregation. In these cases, the details of atomic arrangements are generally not explored and analyzed. The knowledge of the generative physics and chemistry of the material can obviate this problem, since defect configuration libraries as stochastic representation of atomic level structures can be generated, or parameters of mesoscopic thermodynamic models can be derived. To obtain such information for improved predictions, we use data from atomically resolved microscopic images that visualize complex structural correlations within the system and translate them into statistical mechanical models of structure formation. Given the significant uncertainties about the microscopic aspects of the material’s processing history along with the limited number of available images, we combine model optimization techniques with the principles of statistical hypothesis testing. We demonstrate the approach on data from a series of atomically-resolved scanning transmission electron microscopy images of Mo x Re 1- x S 2 at varying ratios of Mo/Re stoichiometries, for which we propose an effective interaction model that is then used to generate atomic configurations and make testable predictions at a range of concentrations and formation temperatures.

25 ENERGY STORAGE↗

Toroidal modeling of runaway electron loss due to 3-D fields in DIII-D and COMPASS

The 3-D field induced relativistic runaway electron (RE) loss has been simulated for DIII-D, COMPASS and ITER plasmas, utilizing the MARS-F code incorporated with the recently developed and updated RE orbit module (REORBIT). Modeling shows effectively 100% loss of a post-disruption, high-current runaway beam in DIII-D, due to the 1 kG level of magnetic field perturbation produced by a fast growing n = 1 resistive kink instability. The RE loss is shown to be independent of the particle energy or the initial location of particles in the configuration space. Applied resonant magnetic perturbation (RMP) fields from in-vessel coils are not effective for RE beam mitigation in DIII-D, but do produce finite (>10%) RE loss in COMPASS, consistent with experimental observations in above two devices. The major reasons for this difference in RE control by RMP between these two devices are (i) the coil proximity to the RE beam and (ii) the effective coil current scaling versus the machine size and the toroidal magnetic field. About 10% RE loss fraction is also predicted for an ITER 15 MA scenario with pre-disruption plasma, highlighting the role of the plasma response. Up to 30% loss is computed, however, by artificially scaling the equilibrium pressure to zero. This is due to the more resistive plasma response and stronger resulting field line stochasticity. Distributions of the lost REs to the limiting surface show poloidally peaked profile near the high-field-side in both DIII-D and COMPASS, covering about 100o poloidal angle. Higher perturbation field level and/or higher particle energy also result in REs being lost to the low-field-side of the limiting surface of these two devices, increasing the effective wetted area. Finally, for the modeled ITER plasmas, the applied RMP field, with optimal poloidal spectrum that maximizes the edge localized mode control, leads to REs being lost to the lower divertor region of the limiting surface within a narrow poloidal band.

3-D field↗

Stochastic Games for Continuous-Time Jump Processes Under Finite-Horizon Payoff Criterion

In this paper we study two-person nonzero-sum games for continuous-time jump processes with the randomized history-dependent strategies under the finite-horizon payoff criterion. The state space is countable, and the transition rates and payoff functions are allowed to be unbounded from above and from below. Under the suitable conditions, we introduce a new topology for the set of all randomized Markov multi-strategies and establish its compactness and metrizability. Then by constructing the approximating sequences of the transition rates and payoff functions, we show that the optimal value function for each player is a unique solution to the corresponding optimality equation and obtain the existence of a randomized Markov Nash equilibrium. Furthermore, we illustrate the applications of our main results with a controlled birth and death system.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Statistical behaviour of adaptive multilevel splitting algorithms in simple models

Adaptive multilevel splitting algorithms have been introduced rather recently for estimating tail distributions in a fast and efficient way. In particular, they can be used for computing the so-called reactive trajectories corresponding to direct transitions from one metastable state to another. The algorithm is based on successive selection–mutation steps performed on the system in a controlled way. It has two intrinsic parameters, the number of particles/trajectories and the reaction coordinate used for discriminating good or bad trajectories. We investigate first the convergence in law of the algorithm as a function of the timestep for several simple stochastic models. Second, we consider the average duration of reactive trajectories for which no theoretical predictions exist. The most important aspect of this work concerns some systems with two degrees of freedom. They are studied in detail as a function of the reaction coordinate in the asymptotic regime where the number of trajectories goes to infinity. We show that during phase transitions, the statistics of the algorithm deviate significatively from known theoretical results when using non-optimal reaction coordinates. In this case, the variance of the algorithm is peaking at the transition and the convergence of the algorithm can be much slower than the usual expected central limit behaviour. The duration of trajectories is affected as well. Moreover, reactive trajectories do not correspond to the most probable ones. Such behaviour disappears when using the optimal reaction coordinate called committor as predicted by the theory. We finally investigate a three-state Markov chain which reproduces this phenomenon and show logarithmic convergence of the trajectory durations.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

High-energy ion beams generated with high efficiency using laser-driven 3D microstructures

Laser-driven ion acceleration in plasma is being proposed as a source of ion beams with a high peak current that can be useful in many fields of science and medicine. Using this method, high proton energies have been achieved by increasing the laser power and by using ultrathin (≤ 200 nm) foils. However, this approach is limited by survivability of the nanotargets to laser prepulses and by difficulty in controlling the plasma acceleration properties. Here, we introduce a new target platform using two-photon polymerization, 3D laser-printed “clone” microstructures with average densities lower than solid that are relatively insensitive to the laser prepulse. Two types of microstructured targets consisting of either a multilayered log-pile or a stochastic arrangement of one micron diameter wires are used. Both demonstrate a higher energy and higher yield proton acceleration compared to thin solid-density foil targets by the robust target normal sheath acceleration (TNSA) mechanism. We find that when such 10–20 μm thick structures are irradiated with a petawatt laser, protons with energies up to 110 MeV and a laser-to-proton conversion efficiency of ~ 10% are obtained. Our work suggests that such microstructures optimized for 60–200 MeV compact proton accelerators are promising for future radiotherapy and other applications.

Physics - Plasma physics↗

On Social Optima of Non-Cooperative Mean Field Games

This paper studies the social optima in noncooperative mean-field games for a large population of agents with heterogeneous stochastic dynamic systems. Each agent seeks to maximize an individual utility functional, and utility functionals of different agents are coupled through a mean field term that depends on the mean of the population states/controls. The paper has the following contributions. First, we derive a set of control strategies for the agents that possess *-Nash equilibrium property, and converge to the mean-field Nash equilibrium as the population size goes to infinity. Second, we study the social optimal in the mean field game. We derive the conditions, termed the socially optimal conditions, under which the *-Nash equilibrium of the mean field game maximizes the social welfare. Third, a primal-dual algorithm is proposed to compute the *-Nash equilibrium of the mean field game. Since the *-Nash equilibrium of the mean field game is socially optimal, we can compute the equilibrium by solving the social welfare maximization problem, which can be addressed by a decentralized primal-dual algorithm. Numerical simulations are presented to demonstrate the effectiveness of the proposed approach.

Li, Sen↗

The RISMC Approach to Perform Advanced PRA Analyses

The existing fleet of nuclear power plants is in the process of extending its lifetime and increasing the power generated from these plants via power uprates. In order to evaluate the impact of these two factors on the safety of the plant, the RISMC Pathway aims to develop simulation-based tools and methods to assess risks for existing nuclear power plants in order to optimize safety. This pathway, by developing new methods, is extending the state-of-the-practice methods that have been traditionally based on logic structures such as Event-Trees and Fault-Trees. These static types of models mimic system response in an inductive and deductive way respectively, yet are restrictive in the ways they can represent spatial and temporal constructs. RISMC analyses are performed by using a combination of thermal-hydraulic codes and a stochastic analysis tool currently under development at the Idaho National Laboratory, i.e. RAVEN. This paper presents a case study in order to show the capabilities of the RISMC methodology to assess impact of power uprate of a boiling water reactor system during a station blackout accident scenario. We employ the system simulator code, RELAP5-3D, coupled with RAVEN which perform the stochastic analysis. Our analysis is in fact performed by: 1) sampling values of a set of parameters from the uncertainty space of interest, 2) simulating the system behavior for that specific set of parameter values and 3) analyzing the set of simulation runs. Results obtained give a detailed investigation of the issues associated with a plant power uprate including the effects of station blackout accident scenarios. We are able to quantify how the timing of specific events was impacted by a higher nominal reactor core power. Such safety insights can provide useful information to the decision makers to perform risk informed margins management.

97 MATHEMATICS AND COMPUTING↗

Universal energy-speed-accuracy trade-offs in driven nonequilibrium systems

The connection between measure theoretic optimal transport and dissipative nonequilibrium dynamics provides a language for quantifying nonequilibrium control costs, leading to a collection of thermodynamic speed limits, which rely on the assumption that the target probability distribution is perfectly realized. This is almost never the case in experiments or numerical simulations, so here we address the situation in which the external controller is imperfect. We obtain a lower bound for the dissipated work in generic nonequilibrium control problems that (1) is asymptotically tight and (2) matches the thermodynamic speed limit in the case of optimal driving. Along with analytically solvable examples, we refine this imperfect driving notion to systems in which the controlled degrees of freedom are slow relative to the nonequilibrium relaxation rate, and identify independent energy contributions from fast and slow degrees of freedom. Furthermore, we develop a strategy for optimizing minimally dissipative protocols based on optimal transport flow matching, a generative machine learning technique. Furthermore, this latter approach ensures the scalability of both the theoretical and computational framework we put forth. Crucially, we demonstrate that we can compute the terms in our bound numerically using efficient algorithms from the computational optimal transport literature and that the protocols we learn saturate the bound.

59 BASIC BIOLOGICAL SCIENCES↗

Understanding the Impacts of Biomass Blending on the Uncertainty of Hydrolyzed Sugar Yield from a Stochastic Perspective

Feedstock price and availability are key challenges for biorefinery development and biomass blending has been suggested as a route to overcome these limitations. However, the impacts of feedstock blending on the variability of yields of hydrolyzed sugar remained unclear. This study quantifies the uncertainties in the sugar yields from the blends of corn stover, switchgrass, and grass clippings by considering both variable and model uncertainties. The results indicate that feedstock blending reduces the uncertainties in sugar yields and delivers feedstock of more uniform quality. A further investigation on xxx demonstrates that the reduction in overall uncertainty is achieved by reduced compositional uncertainties of blended feedstock. Based on these results, the optimization of blending ratios was performed with respect to three objectives: (1) probability of meeting the sugar yield target; (2) expected sugar yields; and (3) sugar yield per unit feedstock expense, while satisfying constraints of feedstock availability and cost. In conclusion, the optimization method developed in this study, is readily applied to other combinations of feedstocks, biofuels production processes, and constraints.

09 BIOMASS FUELS↗

Improved Sampling Algorithms in the Risk-Informed Safety Margin Characterization Toolkit

The RISMC approach is developing advanced set of methodologies and algorithms in order to perform Probabilistic Risk Analyses (PRAs). In contrast to classical PRA methods, which are based on Event-Tree and Fault-Tree methods, the RISMC approach largely employs system simulator codes applied to stochastic analysis tools. The basic idea is to randomly perturb (by employing sampling algorithms) timing and sequencing of events and internal parameters of the system codes (i.e., uncertain parameters) in order to estimate stochastic parameters such as core damage probability. This approach applied to complex systems such as nuclear power plants requires to perform a series of computationally expensive simulation runs given a large set of uncertain parameters. These types of analysis are affected by two issues. Firstly, the space of the possible solutions (a.k.a., the issue space or the response surface) can be sampled only very sparsely, and this precludes the ability to fully analyze the impact of uncertainties on the system dynamics. Secondly, large amounts of data are generated and tools to generate knowledge from such data sets are not yet available. This report focuses on the first issue and in particular employs novel methods that optimize the information generated by the sampling process by sampling unexplored and risk-significant regions of the issue space: adaptive (smart) sampling algorithms. They infer system response from surrogate models constructed from existing samples and predict the most relevant location of the next sample. It is therefore possible to understand features of the issue space with a small number of carefully selected samples. In this report, we will present how it is possible to perform adaptive sampling using the RISMC toolkit and highlight the advantages compared to more classical sampling approaches such Monte-Carlo. We will employ RAVEN to perform such statistical analyses using both analytical cases but also another RISMC code: RELAP-7.

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗

HYPPO: A Surrogate-Based Multi-Level Parallelism Tool for Hyperparameter Optimization

We present a new software, HYPPO, that enables the automatic tuning of hyperparameters of various deep learning (DL) models. Unlike other hyperparameter optimization (HPO) methods, HYPPO uses adaptive surrogate models and directly accounts for uncertainty in model predictions to find accurate and reliable models that make robust predictions. Using asynchronous nested parallelism, we are able to significantly alleviate the computational burden of training complex architectures and quantifying the uncertainty. HYPPO is implemented in Python and can be used with both TensorFlow and PyTorch libraries. We demonstrate various software features on time-series prediction and image classification problems as well as a scientific application in computed tomography image reconstruction. Finally, we show that (1) we can reduce by an order of magnitude the number of evaluations necessary to find the most optimal region in the hyperparameter space and (2) we can reduce by two orders of magnitude the throughput for such HPO process to complete.

adaptation models↗

Collisional simulations of the modulator section in coherent electron cooling

The first section of any coherent electron cooling (CeC) system is the modulator, where the density of the electron beam is modulated by the copropagating ion beam. This density modulation is a result of Coulomb collisions between the individual particles of the two beams. The pairwise, stochastic part of the interactions impacts the overall performance of the CeC process. We present the first simulations of the density modulations of the electron beams from a collisional picture of the dynamics, considering the proof-of-principle CeC experiments at Brookhaven National Laboratory. These simulations were performed using PHAD, which is the first efficient, large-scale collisional numerical method in beam physics that we have previously developed and benchmarked. Realistic beam distributions and external fields have been optimized to provide strong modulation signals necessary for variations of coherent electron cooling systems. Cooling performance limits and potential collisionless simulation pitfalls are pointed out. Published by the American Physical Society 2024

43 PARTICLE ACCELERATORS↗